In looking at so many of the direct digital screen captures posted lately, I have been very surprised at how visible a difference there is between VC1 encodes and encodes in the other codecs, Mpeg2 and AVC.

Since these captures are not "photos" of screens, they are able to capture the full detail of the actual encoding on the disc.

What I have noticed very much, is that VC1 is able to give the highest fine detail of what is in the material to be encoded, without "exaggerating" the "film noise", "grain noise" or "video noise" (depending on what people choose to call it, and whether it was film-sourced, or video-sourced).

Some folks have chosen to say that if there is not a lot of "noise" or "grain" in a HD DVD or Bluray release, then this means that the codec is "filtering" the master.

I disagree with that entirely. Visible film grain, or video noise, are merely "artifacts" of the medium used to capture a scene, and are NOT part of the scene itself. Therefore why should it be part of our HD and BD releases, unless the director intended it to say something?

It is uncanny how VC1 is able to reduce the "annoyance factor" of the film grain, while still keeping more detail from the original picture.

If VC1 was indeed "filtering" the image, as some folks maintain, then VC1 would not be preserving more detail than the other codecs. Therefore, I feel that there is something "a bit different" going on here.

What could that be?

Well, I for one think that it is possible that Microsoft's VC1 encoder may be doing a little Temporal Analysis, or Temporal Interpolation during the encoding process.

Temporal Interpolation would look at adjacent frames for additional details and place them in the frame being encoded. Also, and random noise (or film grain) which does not seem to be associated with true picture detail, would be ignored or "marginalised". This would give encodes that may appear to have less grain, but yet have more picture detail.

Let's look at a few captures from Madshi's recent Thread, which took several shots from various different encodes of Serenity.

Of the various posters who commented, most people stated they felt the VC1 encode was best, with the "second place" going to the German AVC encode from "Premiere". I've added the USA Mpeg2 version also.

I chose to take some close-ups from these two versions to highlight what I am saying here, and will also post some shots from other movies which are captured on this forum by Xylon.

Let's start with Serenity, then. The Mpeg Images are to the left, then AVC in Middle, and the HD DVD images are to the right:

http://www.CampaignHD.com/images/Ship1_MPG.png http://www.CampaignHD.com/images/ship1_DE.png http://www.CampaignHD.com/images/ship1_HD.png

http://www.CampaignHD.com/images/Ship2_MPG.png http://www.CampaignHD.com/images/ship2_DE.png http://www.CampaignHD.com/images/ship2_HD.png

http://www.CampaignHD.com/images/Ship3_MPG.png http://www.CampaignHD.com/images/ship3_DE.png http://www.CampaignHD.com/images/ship3_HD.png

And from the scene of the mountain palace.

NOTE that the bitrate of the Mpeg source used here is 11.44 mbps, the German AVC is 17.89 mbps, while the VC1 source on the HD DVD disc is 15.77 mbps. All fairly comparable.

http://www.CampaignHD.com/images/Palace1_MPG.png http://www.CampaignHD.com/images/Palace1_DE.png http://www.CampaignHD.com/images/Palace1_HD.png

http://www.CampaignHD.com/images/Palace3_MPG.png http://www.CampaignHD.com/images/Palace3_DE.png http://www.CampaignHD.com/images/Palace3_HD.png

http://www.CampaignHD.com/images/Palace4_MPG.png http://www.CampaignHD.com/images/Palace4_DE.png http://www.CampaignHD.com/images/Palace4_HD.png

Now, there may be some who say they can't see the differences.

So, I have also blown up these images by 200%, using the "Lanczos" sizing option. Lanczos is the best sizing option, as it takes whatever detail is present in a picture and uses it to interpolate a better end result.

Therefore, the picture with more detail to begin with will show more detail after being resized with Lanczos.

Again, Mpeg is first, AVC is the second picture, and VC1 is third:

http://www.CampaignHD.com/images/Ship1_MPG_2xL.png http://www.CampaignHD.com/images/ship1_DE_2xL.png http://www.CampaignHD.com/images/ship1_HD_2xL.png

http://www.CampaignHD.com/images/Ship2_MPG_2xL.png http://www.CampaignHD.com/images/ship2_DE_2xL.png http://www.CampaignHD.com/images/ship2_HD_2xL.png

http://www.CampaignHD.com/images/Ship3_MPG_2xL.png http://www.CampaignHD.com/images/ship3_DE_2xL.png http://www.CampaignHD.com/images/ship3_HD_2xL.png

Notice how much sharper the HD DVD versions look after resizing with Lanczos!

And the same applies to these Palace shots (again, Mpeg to left, AVC next, then VC1):

http://www.CampaignHD.com/images/Palace1_MPG_2xL.png http://www.CampaignHD.com/images/Palace1_DE_2xL.png http://www.CampaignHD.com/images/Palace1_HD_2xL.png

http://www.CampaignHD.com/images/Palace3_MPG_2xL.png http://www.CampaignHD.com/images/Palace3_DE_2xL.png http://www.CampaignHD.com/images/Palace3_HD_2xL.png

http://www.CampaignHD.com/images/Palace4_MPG_2xL.png http://www.CampaignHD.com/images/Palace4_DE_2xL.png http://www.CampaignHD.com/images/Palace4_HD_2xL.png

Ok - here is Mission Impossible, Mpeg on the left (from Bluray) VC1 on right (from HD DVD).

Although the "appearance" of more grain noise in the Mpeg version may make one believe that it has more detail, a closer look will show that the VC1 version actually has the same or even more detail, despite the lesser intrusion of this noise.

Look carefully at the lower third of the images on the left hand side of the pictures for a perfect example, and look at the diagonal lines of the stone mouldings.

NOTE - the bitrates are equivalent, since these both came from the BD and HD DVD discs...

http://www.CampaignHD.com/images/MI3a_MPG.png http://www.CampaignHD.com/images/MI3a_VC1.png

http://www.CampaignHD.com/images/MI3a_MPG_2xL.png http://www.CampaignHD.com/images/MI3a_VC1_2xL.png

Here are some shots from King Kong.

First is Mpeg2 (11 mbps), Then AVC (17.mbps), then VC1 (19 mbps)...

While the Mpeg is rather typical for broadcast, the AVC and VC1 rates are virtually the same and are comparable.

Notice that there is far less grain "noise" in the VC1 version, yet far more detail. Note also the better contrast in the VC1 version:

http://www.CampaignHD.com/images/KK1a_MPG.png http://www.CampaignHD.com/images/KK1a_AVC.png http://www.CampaignHD.com/images/KK1a_VC1.png

http://www.CampaignHD.com/images/KK1a_MPG_2xL.png http://www.CampaignHD.com/images/KK1a_AVC_2xL.png http://www.CampaignHD.com/images/KK1a_VC1_2xL.png

More from King Kong - same video rates as listed above.

Note same attributes as listed above, more detail yet lower "noise" in the VC1 version, as well as better contrast in VC1 version.

http://www.CampaignHD.com/images/KK2a_MPG.png http://www.CampaignHD.com/images/KK2a_AVC.png http://www.CampaignHD.com/images/KK2a_VC1.png

http://www.CampaignHD.com/images/KK2a_MPG_2xL.png http://www.CampaignHD.com/images/KK2a_AVC_2xL.png http://www.CampaignHD.com/images/KK2a_VC1_2xL.png

Here are some more shots from King Kong.

First is Mpeg2 (11 mbps), Then AVC (17.mbps), then VC1 (19 mbps)... While the Mpeg is rather typical for broadcast, the AVC and VC1 rates are virtually the same and are comparable.

Notice again that there is less grain "noise" in the VC1 version, yet more detail and better contrast:

http://www.CampaignHD.com/images/KK1c_MPG.png http://www.CampaignHD.com/images/KK1c_AVC.png http://www.CampaignHD.com/images/KK1c_VC1.png

http://www.CampaignHD.com/images/KK1c_MPG_2xL.png http://www.CampaignHD.com/images/KK1c_AVC_2xL.png http://www.CampaignHD.com/images/KK1c_VC1_2xL.png

Here are a few from Aeon Flux - Mpeg on Bluray first, then VC1 on HD DVD second.

Notice the better skin/lip colors on the VC1 version, as well as the better detail on the hairs:

http://www.CampaignHD.com/images/Aeon1b_MPG.png http://www.CampaignHD.com/images/Aeon1b_VC1.png

http://www.CampaignHD.com/images/Aeon1c_MPG.png http://www.CampaignHD.com/images/Aeon1c_VC1.png

http://www.CampaignHD.com/images/Aeon1c_MPG_2xL.png http://www.CampaignHD.com/images/Aeon1c_VC1_2xL.png

Also - looking at the grass with 2 times magnification, the BD version displays some typical Mpeg2 blocking artifacting, while the VC1 HD version looks pretty good...

BD/MPEG in the first shot, HD/VC1 in the second shot.

http://www.CampaignHD.com/images/Aeon1d_MPG_2x.png http://www.CampaignHD.com/images/Aeon1d_VC1_2x.png

Also, scaling with Lanczos, show the best result from the VC1 shot (second again)...

http://www.CampaignHD.com/images/Aeon1d_MPG_2xL.png http://www.CampaignHD.com/images/Aeon1d_VC1_2xL.png

noise and grain are NOT artifacts, high frequency details are not artifacts, VC1 is doing some slight filtering to eliminate high frequency details and noise including very fine film grain, screencaps are proving this time and time again, sadly this is what gives HD that HD look, not that HD-DVD is not good, but it isn't what it could be like D-Theater for example

-Gary

not that HD-DVD is not good, but it isn't what it could be like D-Theater for example

-Gary

Gary, time to let go, D-theater is dead. ;)

noise and grain are NOT artifacts, high frequency details are not artifacts, VC1 is doing some slight filtering to eliminate high frequency details and noise including very fine film grain, screencaps are proving this time and time again, sadly this is what gives HD that HD look, not that HD-DVD is not good, but it isn't what it could be like D-Theater for example

-Gary
I agree that High fequency detail is definitely something that one would wish to keep. But film grain and video noise are actually artifacts of those two respective mediums. While it is good to see some of them, too much is not necessarily a good thing.

The grain, naturally, is used to show the detail, but too much grain can spoil the image, and filtering this grain can spoil the image also.

However, the "detail" that is shown in these images to be superior to the other encodes means that simple "filtering" is not what VC1 is doing.

Insiders from MS have also said that VC1 is not doing this sort of basic filtering.

Temporal Analysis, or Temporal Interpolation, would lift additional detail from adjacent frames, while at the same time it would also reduce the influence of random grain not associated with picture details confirmed in adjacent frames.

This would give the appearance of showing less grain, while also providing more detail in each frame.

http://www.CampaignHD.com/images/Palace4_DE.png http://www.CampaignHD.com/images/Palace4_HD.png

http://www.CampaignHD.com/images/Palace4_DE_2xL.png http://www.CampaignHD.com/images/Palace4_HD_2xL.png

AVC is on the left, and VC-1 to the right? Then VC-1 seems to preserve all the roof lines... it's seems much better than AVC.

http://en.wikipedia.org/wiki/Lanczos_resampling

hmmm...

http://en.wikipedia.org/wiki/Lanczos_resampling

hmmm...

I was thinking the same, zoomed shots are a bad comparison

The Serenity shots can not be used to compare encoder/codec quality in general. The AVC is real time constant bit rate one pass BROADCAST and the VC-1 non real time VBR multi pass HD-DVD! :rolleyes:

The Serenity shots can not be used to compare encoder/codec quality in general. The AVC is real time constant bit rate one pass BROADCAST and the VC-1 non real time VBR multi pass HD-DVD! :rolleyes:

Enough with the valid counterpoints already, let rdjam have his fun.

Note to rdjam - Is this really the best example you could come up with?

if you have nothing to add, move on please

http://en.wikipedia.org/wiki/Lanczos_resampling

hmmm...Hmmm, not sure of the purpose of this post.

Lanczos was specifically created for "blowing up" images while preserving as much detail as possible.

Can't see anything in the artical linked that would say that the VC1 shots here do not have more detail than the AVC shots.

Enough with the valid counterpoints already, let rdjam have his fun.

Note to rdjam - Is this really the best example you could come up with?
Is there an example that you would prefer?

As you will notice from the "reserved" spots, there are more examples coming.

Why is VC1 is so much better?

Its not plain and simple.

The Serenity shots can not be used to compare encoder/codec quality in general. The AVC is real time constant bit rate one pass BROADCAST and the VC-1 non real time VBR multi pass HD-DVD! :rolleyes:
Are you stating that Serenity was encoded to AVC from a master in real-time as it was broadcast? Or is it more likely that an AVC encode was created at an earlier date and time to meet broadcast capabilities? In the latter, why would that have been done in only one pass and in real-time?

Is there an example that you would prefer?

As you will notice from the "reserved" spots, there are more examples coming.

How about an example that is clear when seen with the naked eye.

No matter what method you use to blow it up someone will infer that that has created or exaggerated the problem.

Off the top of my head I am not sure if there are any examples that could be used in this arguement that would be definitive proof either way.

Hmmm, not sure of the purpose of this post.

You are promoting the use of a specific kind of filter for a controversal topic, where the filter, itself, is subject to its own controversy. It's a precariously weak foundation to form an argument. That's all.

Are you stating that Serenity was encoded to AVC from a master in real-time as it was broadcast? Or is it more likely that an AVC encode was created at an earlier date and time to meet broadcast capabilities? In the latter, why would that have been done in only one pass and in real-time?

I think the point is that without knowing the metrics of the encodings, it isn't possible to determine whether these are valid comparisons or not.

You could take a shot from the DVD or broadcast/cable HD in North America, and call it the MPEG-2 example. Would that be valid?

Gary

I really don't think this is a good comparison since 1 is broadcast and the other is hd-dvd.


I think you need to have equaly tuned pictures from the original master. only then you can make a comparison.

Also I would like to know the bitrates at thos frames so that we can compare.

lol maybe you should compare the VC1 encodes on xbox live to a BD encode while we're at it.

I think the point is that without knowing the metrics of the encodings, it isn't possible to determine whether these are valid comparisons or not.

You could take a shot from the DVD or broadcast/cable HD in North America, and call it the MPEG-2 example. Would that be valid?

Gary
I have no argument with the criticism of the way that this thread's comparison has been set up. I was only asking a question about how the broadcasted movie was claimed to have been encoded as I found that statement to be contrary to what I would have assumed.

FYI - The bitrates are available on Madshi's and Xylon's threads, where these captures have been taken directly from.

NOTE - have added the bitrates to the posts...

BTW - Added MI3 examples...

Are you stating that Serenity was encoded to AVC from a master in real-time as it was broadcast? Or is it more likely that an AVC encode was created at an earlier date and time to meet broadcast capabilities? In the latter, why would that have been done in only one pass and in real-time?
The AVC encoding from which the pictures were taken are definitely one pass real-time encoding. I can say that because if you record the same movie at that AVC TV station multiple times, the average bitrate is different each time. I've already tested that.

The AVC encoding from which the pictures were taken are definitely one pass real-time encoding. I can say that because if you record the same movie at that AVC TV station multiple times, the average bitrate is different each time. I've already tested that.
Thanks, that's interesting. So the next obvious question is: encoded from what source material? Maybe I should start a new thread on how broadcast hi-def is encoded. I'm sure someone here knows how all of this works. I always assumed that an optimal broadcast encode was made for broadcast distribution and then that was re-encoded or changed in real-time based on the available bandwidth of that specific broadcaster.

King Kong added - with Mpeg2, AVC and VC1 shots.

I will also be adding the bitrates to all examples used, for comparion's sake...

EDIT - Bitrates have now been added...

The AVC encoding from which the pictures were taken are definitely one pass real-time encoding. I can say that because if you record the same movie at that AVC TV station multiple times, the average bitrate is different each time. I've already tested that.
Are you sure that Premier encodes AVC on the fly?

My understanding is that AVC encoders for HD are not normally realtime and that most AVC movies are pre-encoded, particularly in the case of SkyHD and Premier.

I believe Sky and Premiere both use the Tandberg EN5990 which is a realtime one pass encoder. Lots of people have also commented on different bitrates and end filesizes for the same movies/tv episodes on different hours/days.

I believe Sky and Premiere both use the Tandberg EN5990 which is a realtime one pass encoder.

Sky is definitely real-time, they use the same encoder for everything, sports, TV and movies. So it has to be fast enough (i.e. realtime) to encode sport and broadcast live.

@rdjam, thanks for the comparision! Must have been a lot of work. I love VC1, it is definately the best codec currently around. Even though I try to insult the creators of the VC1 codec with every post I write (see signature), I have to admit that their codec absolutely rulz!

Great post RDJAM! This explains what I have seen on my HT, and I agree 100%. VC-1 transfers look better. Just wish the Blu side would use it. :D

Gary, time to let go, D-theater is dead. ;)

No! Blasphemy!


"Hugs Master & Commander and Moulin Rogue tapes"

:o

Great post RDJAM! This explains what I have seen on my HT, and I agree 100%. VC-1 transfers look better. Just wish the Blu side would use it. :D

Deja Vu is excellent. Its like watching it from HD DVD :D

Great post RDJAM! This explains what I have seen on my HT, and I agree 100%. VC-1 transfers look better. Just wish the Blu side would use it. :D


I just wish when the Blu side does use VC-1, they would use Blu-Ray bandwidth. There are a few examples of this, BTW.

Now, I'm coming to this late but is this comparison between a broadcast encode and an HD-DVD (meaningless in other words)?

To respond directly to the topic title, "noise" may not be the best indicator or unconditionally reliable indicator, but it isn't exactly a "not good" indicator, either. It is agreed, that too much noise can be a detriment to the final presentation, though. Noise is much like the "fingerprint" of the different mediums the program has passed through from source footage to end user copy. The appearance of the noise can also be the tip-off of how/what ways the image has been processed/tampered/doctored/filtered. "Well-defined" noise ("detail" seems like an apt term, but is kind of awkward in the context of describing noise) suggests the underlying image is free of filtering/processing which could lose original detail of the program. Noise with a high-frequency roll-off characteristic can suggest a similar loss in detail in the image. Noise that is oddly "cleaned" from the image can suggest the image has undergone some dynamic sort of processing (noise reduction) with potential similar loss in detail in the image. If a codec is softening the encoded scene from the original scene due to bitrate/bandwidth issues, it is possible that the noise and detail undergo a similar softening effect. Temporal noise reduction may be good in static scene scenarios, but can introduce motion artifacts (smearing) in motion scenes.

Basically, a lot of things are possible. The safest bet with any good quality material is to affect it as little as possible (which includes the noise signature of the material). Once you go beyond that, any number of "bets can be off", regardless of how innocuous one thinks a certain filtering is.

I'm not sure what the point of this is other than to recap other threads with bigger pictures. Of course "high" bitrate multipass tweaked vc-1 on a disc is going to look better than a low bitrate single pass etc broadcast. If it didn't that would be a controversy.

And for the idea that vc-1 filtering away or smoothing over grain is a good idea... How is that being transparent to the source.? I thought the hype behind vc-1 was that it would show us what the master really looked like without artifacts.

Not like this:
http://img124.imageshack.us/img124/7530/mi3graintj4.png

Its not plain and simple.

You must have your tinfoil hat on being that you haven't been brainwashed like many here have. ;)

Kram, you're using "facts" again to prove a point...

Amir has said himself in dozens of threads that VC-1's "advantage" is that is has sub-frame grain optimization". RdJam, what do you suppose that is, if not a filter that effects grain within an image? To give them their due, I'm sure the mathematics are complicated, so RdJam is certainly right when he says it's not "a simple filter", but common, it's not transparent to the master. Numerous threads have now shown films where VC-1 produced a poor transfer (e.g. Lady in the Water), while many VC-1 encodes look fantastic.

If VC-1 fans want to argue that modifying the original film image to make it look "better" they are certainly free to (I even agree is some circumstances), but don't pretend like this aesthetic opinion is even a majority one, let alone so beyone reproach that you don't even have to question it in a post of screen grabs.

VC-1 is obviously a good codec, but several attempts to squeeze it into ever smaller file sizes have noticeably hampered the results. Rather than worrying about the future (e.g. movies on demand) uses, or trying to justify smaller HD DVD disk sizes, it would be nice if VC-1 were actually used to provide the absolute best image possible each and every time. While I see no evidence that VC-1 is in any way fundamentally superior to AVC, it is possible those differences would emerge if VC-1 released BD-optimized encodes (using the higher bitrates) on Blu-Ray disks like some AVS encodes do.

...but if they don't keep squeezing down vc-1 encodes ever further, the whole "efficiency" argument becomes jeopardized. ;) We should be keen to note that the better vc-1 encodes have not been in that "2-3x" range that is often cited. The spread seems to end up more around 1.5-2x between vc-1 vs. mpeg2 encodes. At times, they are even at similar bitrates (15-25 Mb/s), which suggests a very minor inherent efficiency advantage (<1.5x). Naturally, greater efficiency is possible if you are willing to tweak/optimize some settings aggressively. However, if any given compressionist doesn't bother or is unaware of such optimizations on any given vc-1 encoding they produce, the advantage becomes even more dubious.

Amir has said himself in dozens of threads that VC-1's "advantage" is that is has sub-frame grain optimization". RdJam, what do you suppose that is, if not a filter that effects grain within an image? To give them their due, I'm sure the mathematics are complicated, so RdJam is certainly right when he says it's not "a simple filter", but common, it's not transparent to the master. Numerous threads have now shown films where VC-1 produced a poor transfer (e.g. Lady in the Water), while many VC-1 encodes look fantastic.
I don't really follow what you're suggesting is going on. But in general, our design goal is figuring out how to keep the grain, not how to get rid of it or modify it.

If VC-1 fans want to argue that modifying the original film image to make it look "better" they are certainly free to (I even agree is some circumstances), but don't pretend like this aesthetic opinion is even a majority one, let alone so beyone reproach that you don't even have to question it in a post of screen grabs.
Our goal with VC-1 is to provide as accurate a rendition to the master as possible, without modifying the image.

VC-1 is obviously a good codec, but several attempts to squeeze it into ever smaller file sizes have noticeably hampered the results. Rather than worrying about the future (e.g. movies on demand) uses, or trying to justify smaller HD DVD disk sizes, it would be nice if VC-1 were actually used to provide the absolute best image possible each and every time. While I see no evidence that VC-1 is in any way fundamentally superior to AVC, it is possible those differences would emerge if VC-1 released BD-optimized encodes (using the higher bitrates) on Blu-Ray disks like some AVS encodes do.
I've often heard this assertion that some titles have compromised image quality due to lower bitrates. This is inaccurate - quality issues in HD DVD and Blu-ray movies encoded with VC-1 are almost always due to source issues in every title I've looked at.

...but if they don't keep squeezing down vc-1 encodes ever further, the whole "efficiency" argument becomes jeopardized. ;) We should be keen to note that the better vc-1 encodes have not been in that "2-3x" range that is often cited. The spread seems to end up more around 1.5-2x between vc-1 vs. mpeg2 encodes. At times, they are even at similar bitrates (15-25 Mb/s), which suggests a very minor inherent efficiency advantage (<1.5x). Naturally, greater efficiency is possible if you are willing to tweak/optimize some settings aggressively. However, if any given compressionist doesn't bother or is unaware of such optimizations on any given vc-1 encoding they produce, the advantage becomes even more dubious.
We train every shop doing studio encoding with VC-1, so there aren't any widely distributed titles released by operators not aware of these optimizations :).

Note that in most cases where titles get a higher VC-1 bitrate, that is driven more by how much capacity there is on the disc than anything else. The biggest single factor in dropping VC-1 bitrates from the 1st gen titles is that the 1st gen titles often didn't have anything else to spend bits on :).

Just a couple notes to the late entries.

Now, I'm coming to this late but is this comparison between a broadcast encode and an HD-DVD (meaningless in other words)?The comparisons are Mpeg, AVC and VC1, wherever possible from high bandwidth sources, including Bluray, such as the Mission Impossible example.

I will be adding more direct Bluray to HD DVD comparisons, so as to keep it relevant in like-for-like encodes with the different codecs.

To respond directly to the topic title, "noise" may not be the best indicator or unconditionally reliable indicator, but it isn't exactly a "not good" indicator, either. It is agreed, that too much noise can be a detriment to the final presentation, though. Noise is much like the "fingerprint" of the different mediums the program has passed through from source footage to end user copy.More noise, or even "well defined noise", does not mean more detail. The MI example is a good case of this - there is definitely more noise in the Mpeg Bluray version, but not more detail.

And for the idea that vc-1 filtering away or smoothing over grain is a good idea... How is that being transparent to the source.? I thought the hype behind vc-1 was that it would show us what the master really looked like without artifacts. The MI images of noise is a good example, even tho it comes from one of the other threads.

If you look only at the images you have copied, of an out-of-focus background, one might be inclined to think there was more detail in the Mpeg image, since there is so much more noise.

However, in reality, as shown in the MI post in the post above, (http://www.avsforum.com/avs-vb/showthread.php?p=10462466&&#post10462466) there is the same, or more, detail in the VC1 version, once you look at real objects in the scene, such as the stonework used in the example.

Too much noise in the picture can be a bit of a distraction, unless the director truly intended to use the noise to make a statement. In most cases, the noise is present as a result of the type of film used and the filming conditions, but are not actually part of the scene.

VC1 seems to be able to get the additional detail of the scene without exacerbating the film noise. It still keeps the grain, but doesn't let it wreck the scene.

How it does this is speculative, of course, but I really think the MS encoder may be looking at multiple frames and doing a bit of temporal analysis during the encoding process.

Either way, there doesn't seem to be much doubt that the VC1 encodes are genererally showing more detail.

I'll have more examples soon.

More noise, or even "well defined noise", does not mean more detail.

To keep the context of my post intact, the removal of noise is sometimes the "tampering fingerprint" that detail has already been compromised. In either event, removing noise is already moving the result 1 step away from the state of the master, as opposed to faithfully representing what is in the master.

If the noise is intact, and it has a natural signature (natural for the manner in which it occurs, in the first place), that can be a good sign that the utmost detail is intact as well. It's not a guarantee, but it is one of many signs.

More noise, or even "well defined noise", does not mean more detail. The MI example is a good case of this - there is definitely more noise in the Mpeg Bluray version, but not more detail.

You certainly did post them, but that in of itself does not mean it was a "good example". As identical the examples were, it was more a demonstration of subjective splitting of fine hairs, afair.

One can just as well as describe this as the BR version being detailed right down to the noise floor. The vc-1 fell short of that noise floor.

I hope this vc-1/grain thing will be figured out before we see insanely grainy films like Grindhouse, City of God, 28 Days Later (video noise), Aliens, and the last 9 Spielberg films on either HD format. Smoothed over isn't HD to me.

We train every shop doing studio encoding with VC-1, so there aren't any widely distributed titles released by operators not aware of these optimizations :).

I'm sure you do. However, being aware of these optimizations and using them are 2 different things.

I hope this vc-1/grain thing will be figured out before we see insanely grainy films like Grindhouse, City of God, 28 Days Later (video noise), Aliens, and the last 9 Spielberg films on either HD format. Smoothed over isn't HD to me.
Just to point out - 28 Days Later was shot using SD video cameras (though the final scene was shot on 35mm, since they changed the ending.) I highly doubt they'll bother releasing it on one of the HD formats - what would be the benefit?

I've seen 28 Days Later in HD and the titles, the last few minutes, and the credits are all HD sharp.

I've seen 28 Days Later in HD and the titles, the last few minutes, and the credits are all HD sharp.
So you would pay for an HD version where only the final scene and credits are in HD? Really?

How did the rest of the film look? Like an upconverted DVD, right?

There's really not much that can be done with the footage. There's only so much you can get from an XL1.

So you would buy an HD version for the final scene and the credits? Really?

Well, yeah. The impact when it goes from grainy DV to sharp as a tack 35mm is mindblowing.

How did the rest of the film look? Like an upconverted DVD, right?

Way way worse, but it's not that bothersome. It's definitely a look and style that works for the movie IMO. What you're seeing though is DV blown up on film. I doubt there was any upconverting going on.

Well, yeah.
To each his own. Imo, the final scene is the lamest part of the movie (the original ending was far more appropriate as far as I'm concerned.)
The impact when it goes from grainy DV to sharp as a tack 35mm is mindblowing.
Of course. How could it not be? The point is, I'm not willing to pay for 98% grainy DV and 2% HD footage when I already have 98% grainy DV and 2% upconverted SD footage right now. Especially when the 2% is credits and a lame, commercialized, up-beat ending that was tacked on at the last minute.

It just doesn't seem worth it to me.

I agree about an earlier post that comparing the HD DVD of serenity to an mpeg-2 and AVC broadcast is not right
the broadcasts are set to stay at a constant rate at all times due to broadcasts limitations while HD DVD vc1 can use less bits for some scenes or more bits for others
we need a better comparison method like gathering some master 1080p videos and encoding each based on HD DVD specs if we want a real comparison

ps. sorry for bad grammar and structure of sentances

I'm not sure what the point of this is other than to recap other threads with bigger pictures. Of course "high" bitrate multipass tweaked vc-1 on a disc is going to look better than a low bitrate single pass etc broadcast. If it didn't that would be a controversy.

And for the idea that vc-1 filtering away or smoothing over grain is a good idea... How is that being transparent to the source.? I thought the hype behind vc-1 was that it would show us what the master really looked like without artifacts.

Not like this:
http://img124.imageshack.us/img124/7530/mi3graintj4.png

Please take a close look at your example. I think what you are interpreting as additional grain and sharpness on MPEG-2 image are artifacts from visible macroblock edges.

Of course. How could it not be? The point is, I'm not willing to pay for 98% grainy DV and 2% HD footage when I already have 98% grainy DV and 2% upconverted SD footage right now.

So how about a movie with half grainy dv footage and half grainy film footage? At what percentage point does it become worth buying to you?

I must be crazy because I would buy a HD disc of a movie like Open Water which is all grainy dv with only the credits being HD sharp. Probably not for full MSRP though. ;)

So how about a movie with half grainy dv footage and half grainy film footage? At what percentage point does it become worth buying to you?
I don't know - but a single (crappy, tacked-on) scene plus credits? No thanks. YMMV.
I must be crazy because I would buy a HD disc of a movie like Open Water which is all grainy dv with only the credits being HD sharp. Probably not for full MSRP though. ;)
Could be. What would be the point of watching a movie shot on DV in HD? The resolution's just not there.

Please take a close look at your example. I think what you are interpreting as additional grain and sharpness on MPEG-2 image are artifacts from visible macroblock edges.

I'm looking at the dots/grain/ufos in the mpeg-2. It just doesn't look like compression noise. There is blocking within this grain but the grain is a different element.

What's really puzzling is the dozen or so patches of blocks in the vc-1 version that are just solid brown, like holes in the picture. No grain, no noise, just brown blocks with no detail. :confused: :eek:

What we be the point of watching a movie shot on DV in HD? The resolution's just not there.

There is no point. It would be like paying for a standard def upconversion on HD-DVD. ;)

I'm trying to remember how many mainstream dv shot movies there are out there that I think we'll see on either HD format. I can only think of a handful. I thought 28 Days later was in a list of titles announced by Fox but I guess not.

In looking at so many of the direct digital screen captures posted lately, I have been very surprised at how visible a difference there is between VC1 encodes and encodes in the other codecs, Mpeg2 and AVC.
..No offense rdjam but this smear campaign against MPEG-2 and MPEG-4 AVC is getting old. I think you know as well as anyone that there are many great to reference quality encodings with both MPEG-2 and MPEG-4 AVC that have received high picture quality reviews. If we look at the High-Def Digest review website (http://bluray.highdefdigest.com/reviews.html) at only the 4.5 and 5 star PQ ratings here is what I found:

Big Fish (http://bluray.highdefdigest.com/bigfish.html)
Black Rain (http://bluray.highdefdigest.com/blackrain.html)
Brothers Grimm (http://bluray.highdefdigest.com/brothersgrimm.html)
Closer (http://bluray.highdefdigest.com/closer.html)
The Covenant (http://bluray.highdefdigest.com/covenant.html)
Crank (http://bluray.highdefdigest.com/crank.html)
The Descent (http://bluray.highdefdigest.com/descent.html)
The Devil Wears Prada (http://bluray.highdefdigest.com/devilwearsprada.html)
Discovery Atlas: Australia Revealed (http://bluray.highdefdigest.com/discoveryatlasaustraliarevealed.html)
Discovery Atlas: Italy Revealed (http://bluray.highdefdigest.com/discoveryatlasitalyrevealed.html)
Dreamgirls (http://bluray.highdefdigest.com/dreamgirls.html)
Eight Below (http://bluray.highdefdigest.com/eightbelow.html)
Enemy of the State (http://bluray.highdefdigest.com/enemyofthestate.html)
Flyboys (http://bluray.highdefdigest.com/flyboys.html)
Happily N'Ever After (http://bluray.highdefdigest.com/happilyneverafter.html)
Ice Age: The Meltdown (http://bluray.highdefdigest.com/iceagemeltdown.html)
Identity (http://bluray.highdefdigest.com/identity.html)
Incubus: Alive at Red Rocks (http://bluray.highdefdigest.com/incubusaliveatredrocks.html)
Invincible (http://bluray.highdefdigest.com/invincible.html)
John Legend: Live at the House of Blues (http://bluray.highdefdigest.com/legendjohnliveatthehouseofblues.html)
Kingdom of Heaven: The Director's Cut (http://bluray.highdefdigest.com/kingdomofheavendc.html)
Kiss Kiss Bang Bang (http://bluray.highdefdigest.com/kisskissbangbang.html)
Ladder 49 (http://bluray.highdefdigest.com/ladder49.html)
Layer Cake (http://bluray.highdefdigest.com/layercake.html)
Men of Honor (http://bluray.highdefdigest.com/menofhonor.html)
Mission: Impossible III (http://bluray.highdefdigest.com/missionimpossibleiii.html)
Open Season (http://bluray.highdefdigest.com/openseason.html)
Payback (http://bluray.highdefdigest.com/payback_dc.html)
Phone Booth (http://bluray.highdefdigest.com/phonebooth.html)
The Prestige (http://bluray.highdefdigest.com/prestige.html)
The Punisher (http://bluray.highdefdigest.com/punisher.html)
The Pursuit of Happyness (http://bluray.highdefdigest.com/pursuitofhappyness.html)
Reign of Fire (http://bluray.highdefdigest.com/reignoffire.html)
Stealth (http://bluray.highdefdigest.com/stealth.html)
Tears of the Sun (http://bluray.highdefdigest.com/tearsofthesun.html)
Tony Bennett: An American Classic (http://bluray.highdefdigest.com/tonybennett_americanclassic.html)
Transporter (http://bluray.highdefdigest.com/transporter.html)
Transporter 2 (http://bluray.highdefdigest.com/transporter2.html)
The Wild (http://bluray.highdefdigest.com/wild.html)
Volver (http://bluray.highdefdigest.com/volver.html)
Windtalkers (http://bluray.highdefdigest.com/windtalkers.html)
World Trade Center (http://bluray.highdefdigest.com/worldtradecenter.html)


*Note that there are many great quality VC-1 encodings on Blu-ray as well but this post was made to show that there are great quality encodings with MPEG-2 and MPEG-4 AVC. Personally speaking I think that all three video codecs are capable of reference quality encodings and that those who attack MPEG-2 and MPEG-4 AVC do so simply because those video codecs are often used with Blu-ray.

Movie reviews don't compare MPEG2, AVC to VC-1 do they? They just tell if there are any major flaws and if the movie is pleasing to watch. Yes, MPEG encodes look good, but the question is, wouldn't they look better if VC-1 was used? No one is attacking Blu Ray as much as poor out dated encodes. Maybe the BD side does not want pay to use VC-1. :rolleyes:

Movie reviews don't compare MPEG2, AVC to VC-1 do they?Well the fact that there are so many great looking MPEG-2 and MPEG-4 AVC movies in terms of picture quality does show that those video codecs are capable of great quality encodings. The same idea of course can be used for VC-1 as well.


Yes, MPEG encodes look good, but the question is, wouldn't they look better if VC-1 was used?Would the best quality MPEG-2 and MPEG-4 AVC encodings look better with VC-1? Personally I think that is highly unlikely. Granted if they used a newly remastered source that would be different but based on the same exact source material I believe that there is no noticeable picture quality difference between the video codecs assuming that the encodings were all well made. Of course MPEG-2 would be less efficient at encoding than the other two advanced video codecs and would require more to get the same results. Also the very question has an implication behind it that with only minor editing can be used to imply the opposite:

Yes, VC-1 encodes look good, but the question is, wouldn't they look better if MPEG-4 AVC was used?Of course I expect many people to dislike this question but to me both questions would be based on bad assumptions.


No one is attacking Blu Ray as much as poor out dated encodes.I am pretty sure that a few posters in this thread are attacking MPEG-2 and MPEG-4 AVC simply because those video codecs are often used with Blu-ray.

rdjam,

You seem to be a big fan of Toshiba. And they have one of the premier AVC/MPEG4 encoders out there. Do you dislike it too?

BTW: The comparison I would like to see would be "Babel" in VC-1 vs "Babel" in AVC/MPEG4, given that Paramount and their post house chose AVC/MPEG4 for that title and Ben said that AVC/MPEG4 couldn't do film grain right (IIRC), but I don't know if "Babel" will be released in Europe and whether it will use VC-1 if it is released. They've been using VC-1 for other things as far as I know though.

--Darin

I'm looking at the dots/grain/ufos in the mpeg-2. It just doesn't look like compression noise. There is blocking within this grain but the grain is a different element.

What's really puzzling is the dozen or so patches of blocks in the vc-1 version that are just solid brown, like holes in the picture. No grain, no noise, just brown blocks with no detail. :confused: :eek:

VC-1's variable block sizes don't seem to help here as they create an unnatural look with the grain and probably too low a bit rate. While the MPEG2 has edges and some noise amplifying the grain the overall look is still like grain while for VC-1 it is not in this example. If we had the master it would be telling which is closer. They can't both be equally wrong.

VC-1's variable block sizes don't seem to help here as they create an unnatural look with the grain and probably too low a bit rate. While the MPEG2 has edges and some noise amplifying the grain the overall look is still like grain while for VC-1 it is not in this example. If we had the master it would be telling which is closer. They can't both be equally wrong.
Eh? How would variable block sizes hurt grain? Also, the codec will use 8x8 blocks like MPEG-2 where appropriate.

Even our released implementation of VC-1 installed with Windows Media Player 11 will do a much better job of retaining grain tahn any MPEG-2 implementation I've worked with. As always, I'm happy to help if anyone wants to do a head-to-head test with their own sources.

To keep the context of my post intact, the removal of noise is sometimes the "tampering fingerprint" that detail has already been compromised. In either event, removing noise is already moving the result 1 step away from the state of the master, as opposed to faithfully representing what is in the master.If one looks at post number 4 http://www.avsforum.com/avs-vb/showthread.php?p=10462464&&#post10462464 The left-most picture at the top is the Mpeg version. This is a good example of what you are talking about, as it strips almost all the grain to save bandwidth. The AVC and VC1 encodes seem to have a fair bit of grain left.

If the noise is intact, and it has a natural signature (natural for the manner in which it occurs, in the first place), that can be a good sign that the utmost detail is intact as well. It's not a guarantee, but it is one of many signs.Yes, my point in this thread is that noise is not necessarily an indicator of detail. Noise can be exaggerated in some encodes, without adding any detail, prehaps sometimes even obscuring details. So because one encode is "noisier" than others is definitely not a sign that it is a "truer" encode.

No offense rdjam but this smear campaign against MPEG-2 and MPEG-4 AVC is getting old.I'm sorry that you seem to misinterpret this thread, although most others have understood it.

This is a thread about VC1 having a better PQ on the whole than the other codecs, not a negative campaign against the others. This thread is more than likely in response to an apparent series of recent posts against VC1, and merely seeks to show comparisons so that folks can judge the VC1 encodes for themselves.

I think you know as well as anyone that there are many great to reference quality encodings with both MPEG-2 and MPEG-4 AVC that have received high picture quality reviews.Never said there weren't, but by and large, Mpeg has not done a very good job with hi def releases, even on Bluray. VC1, however, has shown the best overall quality of all the codecs.

Where these comparisons get very interesting is when we are able to compare VC1 encodes from one HD disc format, to Mpeg or AVC encodes from an other HD disc format. These are what I am working on right now...

... but the question is, wouldn't they look better if VC-1 was used?
Good post. And that is the point, really. I am buying these HD discs to get the very best version of these movies that are available anywhere, at the end of the day.

While not every single VC1 release is a home run (since source material can vary greatly), the VC1 encodes consistently show themselves to be the best versions, which is why I usually buy the VC1 releases.

Wouldn't it be more likely that an encoder with more grain and less detail is simply adding noise? In comparison, another encoder on the same material that has higher detail has simply avoided adding noise. Entropy - you can't unburn a log.

Have added comparisons from Aeon Flux, using Mpeg on Bluray and VC1 on HD DVD. These are directly comparable.

See post # 9 http://www.avsforum.com/avs-vb/showthread.php?p=10462475&&#post10462475

Eh? How would variable block sizes hurt grain? Also, the codec will use 8x8 blocks like MPEG-2 where appropriate..
It does not look appropriate here with the different block size flat regions cleary visible. It looks artificial and bad. The same block size on the MPEG2 looks better.

It does not look appropriate here with the different block size flat regions cleary visible. It looks artificial and bad. The same block size on the MPEG2 looks better.
? Are you talking about the out of focus background area, or the blowups of the masonry detail in the focal plane?

Here are more shots from Aeon Flux. Mpeg on the left, VC1 on the right.

Looking for detail in these, since some folks assert that VC1 is "filtering" detail out. This assertion is not supported by these shots:

http://www.CampaignHD.com/images/Aeon2b_MPG.png http://www.CampaignHD.com/images/Aeon2b_VC1.png

And here are the blow ups without Lanczos:

http://www.CampaignHD.com/images/Aeon2b_MPG_2x.png http://www.CampaignHD.com/images/Aeon2b_VC1_2x.png

And finally, with Lanczos:

http://www.CampaignHD.com/images/Aeon2b_MPG_2xL.png http://www.CampaignHD.com/images/Aeon2b_VC1_2xL.png

Here is another shot, with some motion. Our old mpeg friend "blocking artifacts" is here again. Mpeg on the left, VC1 on the right.

http://www.CampaignHD.com/images/Aeon3a_MPG.png http://www.CampaignHD.com/images/Aeon3a_VC1.png

And here are the images blown up with Lanczos to show the detail:

http://www.CampaignHD.com/images/Aeon3a_MPG_2xL.png http://www.CampaignHD.com/images/Aeon3a_VC1_2xL.png

And without Lanczos:

http://www.CampaignHD.com/images/Aeon3a_MPG_2x.png http://www.CampaignHD.com/images/Aeon3a_VC1_2x.png

for more photo comparisons

reserved for more

more photos coming

reserved for more comparos

more photos coming here

I'm sorry that you seem to misinterpret this thread, although most others have understood it.rdjam, trust me I understand what you are doing perfectly well. Your first post was more than enough to tell me what you are trying to do and it has nothing to do with comparing the video codecs.


This is a thread about VC1 having a better PQ on the whole than the other codecs, not a negative campaign against the others.No, if that was actually true you would be posting great looking screenshots of MPEG-2 and MPEG-4 AVC which are easy enough to find on Blu-ray.


Never said there weren't, but by and large, Mpeg has not done a very good job with hi def releases, even on Bluray. VC1, however, has shown the best overall quality of all the codecs.I thought you were comparing video codecs in this thread. Shouldn't you be trying to find the best sources of MPEG-2 and MPEG-4 AVC even if they are exclusive to Blu-ray? Or are you trying to make this into another Blu-ray vs HD DVD thread?


Where these comparisons get very interesting is when we are able to compare VC1 encodes from one HD disc format, to Mpeg or AVC encodes from an other HD disc format. These are what I am working on right now...Nonsense, if you were really interested in finding the best sources for each video codec it shouldn't matter if they are dual releases or if they are exclusive to one HD format. In fact you have already posted screenshots from VC-1 encodings exclusive to HD DVD and compared them to broadcast quality MPEG-2 and MPEG-4 AVC material. I honestly wouldn't mind giving you some titles and times to look at but I truly don't believe you have any interest in fairly comparing the video codecs.

I thought you were comparing video codecs in this thread. Shouldn't you be trying to find the best sources of MPEG-2 and MPEG-4 AVC even if they are exclusive to Blu-ray?

And how exactly is he supposed to compare anything when something's available on only one format?

? Are you talking about the out of focus background area, or the blowups of the masonry detail in the focal plane?
The two pics at the top of the page.

And how exactly is he supposed to compare anything when something's available on only one format?

You can't. thats why this thread premise is worthless. I could easily make a thread with the same premise titled "Why is MPEG-2 so much better" and prove my point by comparing Crank on BD to the VC-1 Titles on xbox live marketplace. now thats not really a fair and valid comparison is it?

You can't. thats why this thread premise is worthless. I could easily make a thread with the same premise titled "Why is MPEG-2 so much better" and prove my point by comparing Crank on BD to the VC-1 Titles on xbox live marketplace. now thats not really a fair and valid comparison is it?

That one would not necessarily* be valid because XBox videos are much smaller (Crank I believe uses 25 GB).

*of course, if the VC-1 version looked about as good or even better at only 10 GB (which is highly unlikely) then even this comparison would be quite telling.

The comparisons in this thread however are for the same movie, taking up roughly the same space on disc, with the only difference being the codec.

The comparisons in this thread however are for the same movie, taking up roughly the same space on disc, with the only difference being the codec.


Not entirely true...Are the Serenity comparisons valid (broadcast vs HDM)?

Not entirely true...Are the Serenity comparisons valid (broadcast vs HDM)?

No, because b/c is CBR and the HDM is VBR, so one scene may have more or less bits in the HDM, while the b/c is set at a certain level (about 18Mbps in the UK).

The comparisons in this thread however are for the same movie, taking up roughly the same space on disc, with the only difference being the codec.
There's more to it than that though, isn't there? There are so many variables in situations like this, such as the skill level of the compressionist, the quality/feature set of the encoding device, the compression method (real-time/pre-made, single/multiple pass, etc.) and so on. I don't doubt that VC-1 is an excellent codec, but until the variable of codec used can be truly isolated, you can't draw any definitive conclusions.

I won't deny that VC-1 is considerably more efficient than MPEG-2, but I'm not buying into the idea that neither MPEG-2 nor MPEG-4 AVC can achieve results as good as VC-1's. Not at this point, anyway. Even if that is the case, an apples-to-oranges comparison doesn't prove it.

I won't deny that VC-1 is considerably more efficient than MPEG-2, but I'm not buying into the idea that neither MPEG-2 nor MPEG-4 AVC can achieve results as good as VC-1's.

Has anybody ever said that? Of course MPEG-2 can achieve the same level of quality. It's mainly a matter of bitrate and available space.
But since every optical media is space constrained, the more efficient codec will often provide a better image.

Has anybody ever said that? Of course MPEG-2 can achieve the same level of quality. It's mainly a matter of bitrate and available space.
But since every optical media is space constrained, the more efficient codec will often provide a better image.

Many, many times.

Especially your HD DVD zealot corps, the whole point in this thread as stated by rdjam is to prove that VC-1 can produce a better picture.

This is a thread about VC1 having a better PQ on the whole than the other codecs, not a negative campaign against the others.

The fact is movies released on VC-1 just look better. You can analyze forever about all the varibles and what if's of MPEG! I just want the best looking version of any given release available. Lots of movie watchers including myself and industry experts feel the same. VC-1 consistently does a better job!

I have personally seen incredible transfers for all 3 commonly used HD codecs, so I would say that the best codec is the codec being used by the best author with the best preserved film to use.

Crank, BHD, etc. prove that MPEG2 is Awesome

Casino Royale proves AVC is Awesome

King Kong proves VC-1 is Awesome

~Josh

And how exactly is he supposed to compare anything when something's available on only one format?Well he has already managed to do that with both King Kong and Serenity and last I checked those were exclusive to HD DVD. The entire basis of this thread is rather illogical and all I see is someone who is looking long and hard to find the worst screenshots of MPEG-2 and MPEG-4 AVC that he can find.


That one would not necessarily* be valid because XBox videos are much smaller (Crank I believe uses 25 GB).Certainly, but comparing HD DVD encodings to broadcast material is not a valid comparison either. After all did they have the same source material to work with? Did they have the same peak bitrate to work with? Did they have a studio quality encoder to work with? Did they have the same level of skilled compressionists or the same amount of time to make the encoding? In fact as a previous poster pointed out even comparing two titles encoded on both Blu-ray and HD DVD with two different video codecs is not necessarily a valid comparison since we wouldn't know all of the factors involved in making the encodings.


The fact is movies released on VC-1 just look better.That is an opinion and personally speaking I have seen great looking encodings with all three video codecs.

rdjam, trust me I understand what you are doing perfectly well. Your first post was more than enough to tell me what you are trying to do and it has nothing to do with comparing the video codecs.Well perhaps you could tell me and everyone what you really think is going on then ;)

No, if that was actually true you would be posting great looking screenshots of MPEG-2 and MPEG-4 AVC which are easy enough to find on Blu-ray.Aeon Flux is one of the better BD discs out there, so it's a surprise to me too to see so much block artifacts in it. There'll be other comparisons also, using various encodes of the same movies.

I thought you were comparing video codecs in this thread. Shouldn't you be trying to find the best sources of MPEG-2 and MPEG-4 AVC even if they are exclusive to Blu-ray? Or are you trying to make this into another Blu-ray vs HD DVD thread?You have repeatedly stated in other threads in the past that comparisons can only be made with the same movies. It's hard to understand your shift in position, nor how you could make these comparisons with different movies. By this token, I may as well only show HD DVD exclusives. This is not about HD DVD versus Bluray, so much as it is about how VC1 is better. Where Bluray is using VC1, those releases are also shown to be fantastic quality.

Nonsense, if you were really interested in finding the best sources for each video codec it shouldn't matter if they are dual releases or if they are exclusive to one HD format.Again, one wonders how this could be so when you have stated in the past that comparisons cannot be made with different movies.

In fact you have already posted screenshots from VC-1 encodings exclusive to HD DVD and compared them to broadcast quality MPEG-2 and MPEG-4 AVC material. I honestly wouldn't mind giving you some titles and times to look at but I truly don't believe you have any interest in fairly comparing the video codecs.Because the thread is about VC1. I'd be happy to use VC1 BD exclusives in comparison to AVC and Mpeg2 also. But there are only a couple of these, and I don't have the screenshots yet.

You can't. thats why this thread premise is worthless. I could easily make a thread with the same premise titled "Why is MPEG-2 so much better" and prove my point by comparing Crank on BD to the VC-1 Titles on xbox live marketplace. now thats not really a fair and valid comparison is it?
No, it doesn't. Your premise would run into problems because the Xbox videos are 720p, so the comparison would be unfair.

There will be many more comparisons here, giving folks the opportunity to compare the VC1 versions with the other versions.

It may indeed be a good idea to compare Disney's VC1 BD exclusives with the AVC and Mpeg versions, as soon as they can all be rounded up.

At the end of the day, there is definitely something going on with VC1 which has made it the premier codec at this time. This thread is simply illustrating the differences so that we can understand it better.

Yes, it seems your chosen examples, so far, have shown how vc-1 may only be barely better, and then only if you blow it up 200% and manipulate the image with an oversampling filter. Otherwise, it seems to be pretty much identical under normal viewing conditions, but consumes 33% more bitrate than mpeg2, for the AF movie.

No, it doesn't. Your premise would run into problems because the Xbox videos are 720p, so the comparison would be unfair.

you compared broadcast serenity to hd dvd serenity. thats fair?

Aeon Flux is one of the better BD discs out there, so it's a surprise to me too to see so much block artifacts in it. There'll be other comparisons also, using various encodes of the same movies.
.
The MPEG2 BD of Aeon Flux is 15Gb. The VC-1 HD DVD of Aeon Flux is 20Gb.

It's not surprising that the VC-1 version is perceived to be better as it's a newer codec and has 33% higher bitrate. VC-1 may be better than MPEG2 but this comparison isn't going to provide evidence for either camp.

The MPEG2 BD of Aeon Flux is 15Gb. The VC-1 HD DVD of Aeon Flux is 20Gb.

It's not surprising that the VC-1 version is perceived to be better as it's a newer codec and has 33% higher bitrate. VC-1 may be better than MPEG2 but this comparison isn't going to provide evidence for either camp.

As stated in the Xylon's thread the VC-1 version is actually 16.8gb.

I believe MPEG-2 should only be used as long as it is on a BD50 with a high bitrate. However using the 'advanced' codecs at the same bitrates is more appopriate; I give Sony the thumbs up for using AVC with high bitrates on their latest releases.

Using MPEG-2 on a BD25 with a less than adequate bitrate (i.e. Aeon Flux) does nothing to help those who want the best picture quality possible.

The MPEG2 BD of Aeon Flux is 15Gb. The VC-1 HD DVD of Aeon Flux is 20Gb.
The BD version is on a single layer BD which is 5 gigs smaller than the HD DVD, so it's no surprise that the Mpeg encode was 5 gigs smaller also.

These are comparable releases, from the real-world. Both are what is available commercially for sale on their respective formats.

I've added more shots from Aeon Flux above.

[QUOTE=rdjam]The BD version is on a single layer BD which is 5 gigs smaller than the HD DVD, so it's no surprise that the Mpeg encode was 5 gigs smaller also.

These are comparable releases, from the real-world. Both are what is available commercially for sale on their respective formats.QUOTE]

Are we comparing Codecs or formats?

[QUOTE=rdjam]The BD version is on a single layer BD which is 5 gigs smaller than the HD DVD, so it's no surprise that the Mpeg encode was 5 gigs smaller also.

These are comparable releases, from the real-world. Both are what is available commercially for sale on their respective formats.QUOTE]

Are we comparing Codecs or formats?


Hopefully we're not comparing formats since both handle the same video codecs (the only difference being Blu-Ray's potentially higher bandwidth).

[QUOTE=rdjam]The BD version is on a single layer BD which is 5 gigs smaller than the HD DVD, so it's no surprise that the Mpeg encode was 5 gigs smaller also.

These are comparable releases, from the real-world. Both are what is available commercially for sale on their respective formats.QUOTE]

Are we comparing Codecs or formats?
Comparing codecs - but right now the best examples from any of the codecs are supposed to be what is on the Bluray and HD DVD formats, since they are pre-encoded with Variable Bit Rates (VBR).

So the best Mpeg and AVC examples are going to be on BD/HD shiny discs.

Where possible we are using Mpeg and AVC examples from braodcast as well (as in the first examples), which some folks also complain about also because they are CBR examples.

I feel that someone will always have some complaint or other about any comparisons, but we are using the best examples of each from whatever high-end sources are available.

Hopefully we're not comparing formats since both handle the same video codecs (the only difference being Blu-Ray's potentially higher bandwidth). Yes, the HD and BD formats can handle the same three codec choices, however, in reality most HD DVD material is in VC1, while most BD material is not.

Hopefully, this will change in time, but it hasn't yet

I have personally seen incredible transfers for all 3 commonly used HD codecs, so I would say that the best codec is the codec being used by the best author with the best preserved film to use.

Crank, BHD, etc. prove that MPEG2 is Awesome

Casino Royale proves AVC is Awesome

King Kong proves VC-1 is Awesome

~Josh

Agree. I think each codec has its strengths and weaknesses. To unreservedly declare that one codec is always so much better than another is absurd.

I don't think picking a movie that comes up 10 GB short of using the full capacity of a sl disc is "using the best example". I think what you have shown is that lower bitrate mpeg2 does admirably well at providing a nearly equivalent experience to what you would get on a 15 GB vc-1 production. Once you allow for higher bitrates, the potential in mpeg2 increases, as well.

I think the main point is that the detail is not lost in the VC1 version, as is stated by its detractors. If one looks at these images closely, one can see much more detail in the VC1 version:

http://www.CampaignHD.com/images/Aeon2b_MPG.png http://www.CampaignHD.com/images/Aeon2b_VC1.png

And here are the blow ups, with Lanczos:

http://www.CampaignHD.com/images/Aeon2b_MPG_2xL.png http://www.CampaignHD.com/images/Aeon2b_VC1_2xL.png

If vc-1 is operating in a comfortable bitrate regime, then there is no reason to suspect detail will be lost. Where people have suspected that detail is suffering is in those cases where bitrate is hovering in that 10 Mb/s and below range, in those attempts to be the compression king. As such, you have not chosen particularly good examples to address this.

We already know that if any codec of interest here is fed an appropriate bitrate, then it is capable of retaining detail.

As stated in the Xylon's thread the VC-1 version is actually 16.8gb.

I believe MPEG-2 should only be used as long as it is on a BD50 with a high bitrate. However using the 'advanced' codecs at the same bitrates is more appopriate; I give Sony the thumbs up for using AVC with high bitrates on their latest releases.

Using MPEG-2 on a BD25 with a less than adequate bitrate (i.e. Aeon Flux) does nothing to help those who want the best picture quality possible.
He posted the correction after I made my post.

I agree with your other comments. I'm amazed that the MPEG2 Aeon flux puts up so good a show to be honest.

If vc-1 is operating in a comfortable bitrate regime, then there is no reason to suspect detail will be lost. Where people have suspected that detail is suffering is in those cases where bitrate is hovering in that 10 Mb/s and below range, in those attempts to be the compression king. As such, you have not chosen particularly good examples to address this.

We already know that if any codec of interest here is fed an appropriate bitrate, then it is capable of retaining detail.

I agree with your post, but not the implication of your last statement. It seems to imply that all codecs are the same. If so, why bother trying to improve them?

Codecs, at the compression levels that are used on BD and HD-DVD are going to change the image quality. They may all look great on good quality encodes, but they aren't going to be identical to the master on the bitrates that are used (especially on screen shots).

The questions that can be asked are :

1) At what average bitrates do the encodes fall off a cliff and become unacceptable?
2) At good bitrates, what are the subtle artifacts that I expect to see for this codec?

VC-1 is shown to be best at 1) so far. One can argue that AVC is about the same, but there are differences. VC-1 can support arbitrary block-size, which can help preserve edge detail during low bitrates.

As for 2, each codec will tend to show different artifacts. I'm not an expert on each one, but I will tell you that in the still-image world, each codec produces its own unique effects. JPEG (based on DCT) tends to have ringing and blocking artifacts that can give a perception of added sharpness when reasonably good quality factors are used. The perception isn't real -- you are actually increasing noise. JPEG2000 (wavelet-based compression) tends to do the opposite. Even at high quality, they tend to soften an image. Now, JPEG2000 falls off a cliff at smaller file sizes (greater compression) than JPEG2, so it is a higher efficiency still-image codec. However, its differece in "look" at reasonable qualities is apparent.

I don't think any of these screenshot examples is really proving (2). Controlled tests, and a deeper understanding of the codecs and encoding parameter space is required. However, I applaud the OP with trying to show that grain doesn't mean sharpness.

I agree with your post, but not the implication of your last statement. It seems to imply that all codecs are the same.

No, I don't believe I said they are all the same. They are the same as far as they all have a "comfortable zone" for bitrate for a particular material. How they respond when dipping below that zone is certainly different, of course.



VC-1 can support arbitrary block-size, which can help preserve edge detail during low bitrates.

If you are at low bitrates, you will be macroblocking. In vc-1 it is just "masked macroblocking". Either way, the original detail in those local areas of the scene will be fubar.

The point seems to be to not get into that zone, in the first place (unless you are really bandwidth limited by the medium).

If you are at low bitrates, you will be macroblocking. In vc-1 it is just "masked macroblocking". Either way, the original detail in those local areas of the scene will be fubar.
"Masked macroblocking" isn't a meaningful term of art. Maybe you're thinking of a strictly postprocessing deblocking filter? Our loop filter measurably improves our accuracy to source measured in PSNR. A loop filter means fewer bits are needed to accurately reproduce flat/smooth areas, savings bits that can be spent in more complex areas.

It's important to understand compression isn't some flat thing. We vary the degree of compression down to the macroblock, so a feature that improves efficiency in one area of the image frees up bits to be used in other parts of the image.

A macroblock artifact with softened edges is still a macroblock. The detail which would be in place of that macroblock is no longer going to be accurate, even though other areas of the image are accurate. That's all I was saying.

Now if you are attempting to imply that there are zero circumstances (even with nonideal settings) where vc-1 would degrade to macroblock artifacts or softened detail or banding, that's another discussion, altogether.

With the capacity and bandwidth of DL BRD and HD-DVDs you'd think macroblocking would be a thing of the past.

With the capacity and bandwidth of DL BRD and HD-DVDs you'd think macroblocking would be a thing of the past.

If we were still displaying SD material yes. But we're using higher resolutions 4x higher.

I don't think we'll have cases where the average person would notice them anymore though. Heck I don't even notice them most of the time, unless someone freezeframes it and points it out to me.

I think if there was a concerted effort to make encodings that deliberately filled the target media disc in capacity and/or max bitrate, we could be through with macroblock artifacts and poor gradients. Sure, there would always be the glimmer of suspicion that the encode could have been smaller with no ill effect, but what's more important?...an encode that is the best it can possibly be given the format it resides on, or keenly straying that line between bitrate glut and bitrate starvation? I'll take the former, and leave the latter for academic research and digital satellite/cable/ip/cellphone operations. ;)

A macroblock artifact with softened edges is still a macroblock. The detail which would be in place of that macroblock is no longer going to be accurate, even though other areas of the image are accurate. That's all I was saying.
I think you're meaning something other than macroblock when you say macroblock. A macroblock is simply a 16x16 block. I think you're talking about block artifacts, but it'd help me understand what you're trying to say a lot better if you could be more specific.

Now if you are attempting to imply that there are zero circumstances (even with nonideal settings) where vc-1 would degrade to macroblock artifacts or softened detail or banding, that's another discussion, altogether.
My point is that a loop filter isn't just there to hide artifacts, but to allow parts of the image to encode more efficiently AND accurately in order for those bits to be provided to another part of the image.

Ben here you have a link of benefits with VC1.


http://www.thisishddvd.com/whatisvc1.aspx

No need writing in word what your company already has shown in pictures.

I think if there was a concerted effort to make encodings that deliberately filled the target media disc in capacity and/or max bitrate, we could be through with macroblock artifacts and poor gradients. Sure, there would always be the glimmer of suspicion that the encode could have been smaller with no ill effect, but what's more important?...an encode that is the best it can possibly be given the format it resides on, or keenly straying that line between bitrate glut and bitrate starvation? I'll take the former, and leave the latter for academic research and digital satellite/cable/ip/cellphone operations. ;)

Sony could bring you HD Superbit :)

Aeon Flux is one of the better BD discs out there, so it's a surprise to me too to see so much block artifacts in it.Aeon Flux is by no means one of the best looking Blu-ray discs and I hardly find it surprising that you could find poor looking screenshots of it given enough time.


You have repeatedly stated in other threads in the past that comparisons can only be made with the same movies.rdjam, I said that for individual titles and I don't think I ever said that such comparisons would be a way to judge an entire video codec. After all there are many factors involved in making an encoding and the video codec is only one of the factors involved.


By this token, I may as well only show HD DVD exclusives.You already have by posting screenshots from King Kong and Serenity.


This is not about HD DVD versus Bluray, so much as it is about how VC1 is better.VC-1 is an efficient video codec but from what I have seen all three video codecs are quite capable of producing reference quality encodings. Granted MPEG-2 is less efficient at it than the other two video codecs but even MPEG-2 is capable of reference quality encodings with sufficient bitrate.


Again, one wonders how this could be so when you have stated in the past that comparisons cannot be made with different movies.When exactly did I state that and what exactly did I state? Also technically speaking I do change some of my opinions over time since both me and world change. Personally I think only fools try to remain the same while wise men try to improve themselves.


Because the thread is about VC1.Yes, but that is the problem. Instead of trying to objectively judge the video codecs the entire point of this thread is instead about trying to prove a belief that you already had.


I think the main point is that the detail is not lost in the VC1 version, as is stated by its detractors.Just to say this but personally I believe that all three video codecs are capable of encoding fine detail.

No probs, Richard. We can probably leave it there. Yes, all codecs are "capable" given the right conditions, but in real life, we are seeing differences.

I'll have some more soon, and will also be including VC1 on Bluray for you.

Xylon has posted some new comparisons between VC1 and Mpeg2 releases on his thread.

With Training Day, for example, these direct-mode 200% blowups show the difference caused by Mepg2 macro-blocking quite starkly.

Notice how much cleaner and accurate the VC1 version is.

Bluray on the left, HD DVD on the right:

http://www.CampaignHD.com/images/Trainingday_bd_1a_2x.png http://www.CampaignHD.com/images/Trainingday_hdvd_1a_2x.png

http://www.CampaignHD.com/images/Trainingday_bd_1b_2x.png http://www.CampaignHD.com/images/Trainingday_hdvd_1b_2x.png

Xylon has posted some new comparisons between VC1 and Mpeg2 releases on his thread.

With Training Day, for example, these direct-mode 200% blowups show the difference caused by Mepg2 macro-blocking quite starkly.

Not wishing to argue with you, but isn't this a slightly loaded comparison?

Isn't the VC-1 version of Training day araound 17 gigs, whereas the BD version is 15 gigs?

No offence but 'newer and more efficient codec at higer bitrate' beats 'older inefficient codec at lower bit rate' is hardly news.

Xylon has posted some new comparisons between VC1 and Mpeg2 releases on his thread.

With Training Day, for example, these direct-mode 200% blowups show the difference caused by Mepg2 macro-blocking quite starkly.

Notice how much cleaner and accurate the VC1 version is.

Bluray on the left, HD DVD on the right:

http://www.CampaignHD.com/images/Trainingday_bd_1a_2x.png http://www.CampaignHD.com/images/Trainingday_hdvd_1a_2x.png

http://www.CampaignHD.com/images/Trainingday_bd_1b_2x.png http://www.CampaignHD.com/images/Trainingday_hdvd_1b_2x.png


Considering that the MPEG-2 encode of Training Day is actually SMALLER than the VC-1 encode by 2GB, and they look that similar to the point I can't even tell the difference with a single frame blown up on my monitor, you are really shooting yourself in the foot. I would expect such a comparison from someone demonstrating how unnecessary VC-1 is!

It seems the vc-1 encode edged out ahead in this comparison, but at negative efficiency. :p

Actually, the thread is comparing VC1 to other codecs - the difference in encode size is negligible in this case. In some cases, the sizes have been the other way around.

Both these releases are direct "competitiors" to each other, using VBR on next-gen discs, and so are absolutely fair to compare.

After all, it would hardly be "fair" to compare an Mpeg2 encode of twice the size (which doesn't exist), just because Mpeg2 is not as good.

Sure it would be fair, if the "2x efficiency" claim was really true. Unfortunately, this really gets blown out of the water, when we find examples (such as the one you have chosen to cross-post here and various other pet topics) where vc-1 is actually eating up more bitrate than the older codec- hence, "negative efficiency".

where vc-1 is actually eating up more bitrate than the older codec- hence, "negative efficiency".Because the BD version is on a smaller disc, I suspect, as are 70% of bluray releases.

And, BTW, my post was on-topic.

Nevertheless, the vc-1 is eating up more bitrate in the example you chose.

If you want to infer the influence of disc sizes, perhaps you should also consult the free space left on the respective discs, first?

rdjam,

You aren't helping your argument with your small thumbnail sized examples. I wouldn't want to watch any part of a movie zoomed into a thumbnail the size of my cell-phone. Why don't you chose a newer title like Flag's of our Fathers to post examples from? Oh wait, that is being done already by Xylon and I much prefer his larger screen-caps.

Isn't VC-1 basically WM-9? Has anyone done a comparison between T2 MPEG2 encode on BluRay and T2 WM-9 HD encode from the T2 Extreme Edition? From what I have read they appear to have used the same master. They also must have used quite a bit of compression in order to make the movie fit on a standard DVD. Might be interesting to see that comparison.

If anyone has any insight there, please share.

If it's like any of the other demo wmv-hd clips I've seen, I'd say the likelihood of banding and softening are imminent. :p (I'm sure T2 is better than that, though, but isn't that only 720p?)

If it's like any of the other demo wmv-hd clips I've seen, I'd say the likelihood of banding and softening are imminent. :p (I'm sure T2 is better than that, though, but isn't that only 720p?)
T2 was paleolithic era. Main Profile, not Advanced. CBR, not VBR. And many, many generations of the codec ago. Plus playback constraints were worse back then, so those encodes didn't use all the modern tricks we have today.

T2 itself was 1440x800 anamoprhic IIRC.

So you have everything to gain by reissuing a new encoding of T2 using the latest settings and posting that sample on your website. Surely the differences will be jarring, while holding or decreasing the bitrate even further, with the newest state of the codec...

So you have everything to gain by reissuing a new encoding of T2 using the latest settings and posting that sample on your website. Surely the differences will be jarring, while holding or decreasing the bitrate even further, with the newest state of the codec...
Perhaps you haven't heard.

T2 is on HD DVD in Europe, and in glorious VC1. It'll be a very easy comparison, and soon ;)

rdjam,

You aren't helping your argument with your small thumbnail sized examples. I wouldn't want to watch any part of a movie zoomed into a thumbnail the size of my cell-phone. Why don't you chose a newer title like Flag's of our Fathers to post examples from? Oh wait, that is being done already by Xylon and I much prefer his larger screen-caps.
Well, if you want the full 1920 by 200 to be blown up 200% and 4 times the size, I can sure do that.

But if you are implying that blowing these up for examination is not relevant to real-world viewing, then you probably don't have a 120" screen where one sees all of this detail. It's part of the joys of being a perfectionist...

Perhaps you haven't heard.

T2 is on HD DVD in Europe, and in glorious VC1. It'll be a very easy comparison, and soon ;)

At the same bitrate as the old wmv-hd sample? I doubt it, but that would be quite something to behold.

Perhaps you haven't heard.

T2 is on HD DVD in Europe, and in glorious VC1. It'll be a very easy comparison, and soon
At the same bitrate as the old wmv-hd sample? I doubt it, but that would be quite something to behold.
What relevance is the old DVD WMV version to Bluray and HD DVD?

The old Windows Media codec used for that DVD release is now over half decade old, an antique by today's standards. VC1 is not the same.

No, the comparison shall be between the VC1 on HD DVD versus the codec used on the bluray version - which codec was that again?

The context is referring to the wmv-hd sample that appears on MS's demo website. If the codec has matured so much, then it should have no problem showing considerable improvement over the older encode, while using even less or just using the same bitrate.

If you are going to compare to a relatively high bitrate hdvd version, then you have no idea if it was because of codec maturation or just plainly higher bitrate. The most you can declare is that one is an apple and one is an orange. That's it.

The context is referring to the wmv-hd sample that appears on MS's demo website. If the codec has matured so much, then it should have no problem showing considerable improvement over the older encode, while using even less or just using the same bitrate. If you are going to compare to a relatively high bitrate hdvd version, then you have no idea if it was because of codec maturation or just plainly higher bitrate. The most you can declare is that one is an apple and one is an orange. That's it. I would respectfully submit that this argument merely seeks to change the goal posts.

The WMV encode from that old DVD was created as an added value to fit on a regular DVD, in addition to the regular movie.It's a completely different ballgame to an HD DVD or Bluray release where one can use as much space as desired.

So to say that if the VC1 version blows away the Bluray version it's still a "failure" in your eyes because it is bigger than the old DVD wmv version is just another "inventive irrelevancy" that you seem so fond of.

The Bluray and HD DVD versions of this movie are directly comparable competitors and have nothing to do with previous versions, no more than the Bluray version should be smaller than the Satellite Mpeg version.

But perhaps you are concerned that the Bluray version will not win this comparison? I thought your confidence in Mpeg2 was supreme?

You are coming around now. It's apples to oranges. So it can only be guessed what improvements came from brute bitrate increases or codec maturation. The real test would be to see what improvements are possible within the same bitrate envelope.

Whether or not a br or hdvd release comes out is an entirely separate matter altogether. I'm sure they will look great.

You are coming around now.
Thanks, but I haven't changed my position, nor do I agree with your statements... and the answer to your above question is: "Codec Evolution"

The best you can do is an assumption. You can never know for sure unless an apple to apple test is done.

The best you can do is an assumption. You can never know for sure unless an apple to apple test is done.
They've been done repeatedly on a number of threads. Mpeg lost. Catch the re-runs this weekend, perhaps?

I have no idea what you refer to with this comment.

...clearly

They've been done repeatedly on a number of threads. Mpeg lost. Catch the re-runs this weekend, perhaps?MPEG what?

They've been done repeatedly on a number of threads. Mpeg lost. Catch the re-runs this weekend, perhaps?What exactly do you think that MPEG-2 lost at? Given sufficient bitrate and capacity I don't think it is video quality since there have been reference quality encodings done with MPEG-2. On the other hand I would agree with the fact that the two advanced video codecs are mroe efficient than MPEG-2.

Don't know what the debate is on that one. The Mpeg shots have lost to VC1 pretty consistently.

It's easy to complain this was because of the bitrate being too low for Mpeg to "breathe", but these are Bluray releases, direct competitors for the HD DVD VC1 releases, and are what is being sold in real-life.

About 70% of bluray titles available are Mpeg2 and 70% of bluray titles are on Single Layer discs.

Given what we've seen so far, it's a very good reason to go for AVC and VC1 on bluray titles (or the HD DVD versions).

However, the final comparison I'm waiting for is "Dreamgirls" - which is Dual Layer Mpeg2 for bluray and VC1 for HD DVD.

If mpeg can't beat VC1 on this title, with so much more space, I think we can end the debate for good.

If mpeg can't beat VC1 on this title, with so much more space, I think we can end the debate for good.

It doesn't necessarily have to beat it...it only has to match it to end the debate (assuming the debate is about what "looks" better).

[QUOTE=kitzi]
Comparing codecs - but right now the best examples from any of the codecs are supposed to be what is on the Bluray and HD DVD formats, since they are pre-encoded with Variable Bit Rates (VBR).

So the best Mpeg and AVC examples are going to be on BD/HD shiny discs.

Where possible we are using Mpeg and AVC examples from braodcast as well (as in the first examples), which some folks also complain about also because they are CBR examples.

I feel that someone will always have some complaint or other about any comparisons, but we are using the best examples of each from whatever high-end sources are available.

Yes, the HD and BD formats can handle the same three codec choices, however, in reality most HD DVD material is in VC1, while most BD material is not.

Hopefully, this will change in time, but it hasn't yet


And they complain with right too. I work for a European PayTV Channel, and we have yet to come across a single HD movie (delivered on Sony HDCam tapes) that looks good. HDCam has a bitrate of over 100mbs, but any HD-DVD or Blu-ray disc is far superior to what's being delivered to broadcasters. I can't shake the feeling that this is intentional (since there is no way of copy protecting tapes for broadcasters).

And they complain with right too. I work for a European PayTV Channel, and we have yet to come across a single HD movie (delivered on Sony HDCam tapes) that looks good. HDCam has a bitrate of over 100mbs, but any HD-DVD or Blu-ray disc is far superior to what's being delivered to broadcasters. I can't shake the feeling that this is intentional (since there is no way of copy protecting tapes for broadcasters).
That's interesting. You don't happen to work for Premiere, do you?

That's interesting. You don't happen to work for Premiere, do you?

No, but close. I work for Teleclub (Switzerland). We were the first PayTV channel in Europe. Premiere is actually a spinoff (so to speak) from TC. It's all about movie rights, but I can't go into details here. We still get most of our movies sent to us from Premiere and we offer a few channels that are direct feeds off of Premiere. You can check it out on www.teleclub.ch. It's in German though.

No, but close. I work for Teleclub (Switzerland). We were the first PayTV channel in Europe. Premiere is actually a spinoff (so to speak) from TC. It's all about movie rights, but I can't go into details here. We still get most of our movies sent to us from Premiere and we offer a few channels that are direct feeds off of Premiere. You can check it out on www.teleclub.ch. It's in German though.
Thanks! I'm German, so no problem with that site! With which bitrates do you send? Premiere HD recently lowered their bitrates... :( I guess only people from Switzerland can receive your channels, right?

Say, do you have any influence on movie/master selection? Recently Premiere HD has shown some movies (e.g. Serenity, Pride & Prejudice, ...) cropped instead of OAR. I really hate that. So if you can do anything to change that, that would be most welcome... :D

About 70% of bluray titles available are Mpeg2 and 70% of bluray titles are on Single Layer discs..

Can you compare 100% of them to their VC-1 encoded counterparts?

If not, this statement is banal fluff.

Can you compare 100% of them to their VC-1 encoded counterparts?

If not, this statement is banal fluff.
In your opinion...

Given the performance demonstrated by Mpeg on BD/HD discs so far, it's quite relevant to those wanting the best presentation. A quick look at the bitrates thread on blu-ray.com shows that most of the BD Mpeg releases are only 16 to 23 mbps, on average, so the samples shown on Xylon's threads so far are pretty representative of most of them...

Unfortunately, not all BD films are available on as VC1 HD DVD releases - though it would be nice. However, I have most of them as TS files from broadcast HD at almost the same bitrates ;)

It's easy to complain this was because of the bitrate being too low for Mpeg to "breathe", but these are Bluray releases, direct competitors for the HD DVD VC1 releases, and are what is being sold in real-life.In real life we have seen reference quality encodings in MPEG-2, MPEG-4 AVC, and VC-1 on Blu-ray. All three video codecs have shown themselves capable of reference quality encodings.


However, the final comparison I'm waiting for is "Dreamgirls" - which is Dual Layer Mpeg2 for bluray and VC1 for HD DVD.

If mpeg can't beat VC1 on this title, with so much more space, I think we can end the debate for good.Well if this debate was about efficiency I would agree with that but you have been debating that MPEG-2 can not look as good as VC-1. A claim that professional reviews have long disagreed with and is not going to be supported by Dreamgirls since both the VC-1 version of Dreamgirls (http://hddvd.highdefdigest.com/dreamgirls.html) and the MPEG-2 version of Dreamgirls (http://bluray.highdefdigest.com/dreamgirls.html) were considered reference quality.

Thanks! I'm German, so no problem with that site! With which bitrates do you send? Premiere HD recently lowered their bitrates... :( I guess only people from Switzerland can receive your channels, right?

Say, do you have any influence on movie/master selection? Recently Premiere HD has shown some movies (e.g. Serenity, Pride & Prejudice, ...) cropped instead of OAR. I really hate that. So if you can do anything to change that, that would be most welcome... :D
Yes, we broadcast in Switzerland only. Unfortunately I have no influence on any format decisions (I work in the Graphics departement). As of now, TC doesn't broadcast in HD, but we will see a shift towards it on TV over DSL (which we do for the national Telecom company). There is a lot of testing going on a the moment, no codecs or bitrates have been decided on yet (the HDcam movies we get, because the new movies are only delivered this way. We convert them down to SD for our channels). Somebody on another forum mentioned 20mbit/s was European standard for HD broadcast, but there is no such thing as a standard. Supposedly our encodes will be around 6 to 9 mbit/s.
Premiere and TC are completely independent companies by the way.

But back on topic: Since there is no telling if broadcasters receive the same masters as the studios use to encode for BD or HD-DVD, any comparison of optical media to broadcast content seems to be guesswork.
And just like Richard Paul said, when there are (current) encodes on both formats (and this is what this is all about in the end) reviewers all over the net agree that any codec can render great results, VC1 or AVC or MPEG2.

In your opinion...

Given the performance demonstrated by Mpeg on BD/HD discs so far, it's quite relevant to those wanting the best presentation. A quick look at the bitrates thread on blu-ray.com shows that most of the BD Mpeg releases are only 16 to 23 mbps, on average, so the samples shown on Xylon's threads so far are pretty representative of most of them...

Unfortunately, not all BD films are available on as VC1 HD DVD releases - though it would be nice. However, I have most of them as TS files from broadcast HD at almost the same bitrates ;)

I'm not sure that it is relevant at all rdjam. The majority of the releases you're speaking of come from June-October 2006 and we're not living in the past still are we? Next you'll be telling us that most BD50's are really BD25's as BD50's don't exist.

If you include ALL the scored Blu-ray reviews from High Def Digest, Home Theater Spot, Home Theater Forum, DVD Talk and Upcoming discs, including all the MPEG-2 releases from the first few months of Blu-ray's launch, and include titles that never even got released like the 1 star rated Robocop, you get the following figures.

Keep in mind several titles have 5 reviews here, so this is not refelctive of how many titles there actually are on Blu-ray.

Of 680 reviews: 431 MPEG-2, 125 AVC/MPEG-4, 124 VC-1
Average star rating

1) 4.22 stars AVC/MPEG-4
2) 3.94 stars VC-1
3) 3.78 stars MPEG-2

Why is VC-1 so much better? You've got me there! I really don't know.

I'm not sure that it is relevant at all rdjam. The majority of the releases you're speaking of come from June-October 2006 and we're not living in the past still are we? Next you'll be telling us that most BD50's are really BD25's as BD50's don't exist.

If you include ALL the scored Blu-ray reviews from High Def Digest, Home Theater Spot, Home Theater Forum, DVD Talk and Upcoming discs, including all the MPEG-2 releases from the first few months of Blu-ray's launch, and include titles that never even got released like the 1 star rated Robocop, you get the following figures.

Keep in mind several titles have 5 reviews here, so this is not refelctive of how many titles there actually are on Blu-ray.

Of 680 reviews: 431 MPEG-2, 125 AVC/MPEG-4, 124 VC-1
Average star rating

1) 4.22 stars AVC/MPEG-4
2) 3.94 stars VC-1
3) 3.78 stars MPEG-2

Why is VC-1 so much better? You've got me there! I really don't know.

:D excellent post!

By those numbers AVC is kicking vc-1 ass more than vc-1 is kicking mpeg-2's.

I'm not sure that it is relevant at all rdjam. The majority of the releases you're speaking of come from June-October 2006 and we're not living in the past still are we? Next you'll be telling us that most BD50's are really BD25's as BD50's don't exist.

If you include ALL the scored Blu-ray reviews from High Def Digest, Home Theater Spot, Home Theater Forum, DVD Talk and Upcoming discs, including all the MPEG-2 releases from the first few months of Blu-ray's launch, and include titles that never even got released like the 1 star rated Robocop, you get the following figures.

Keep in mind several titles have 5 reviews here, so this is not refelctive of how many titles there actually are on Blu-ray.

Of 680 reviews: 431 MPEG-2, 125 AVC/MPEG-4, 124 VC-1
Average star rating

1) 4.22 stars AVC/MPEG-4
2) 3.94 stars VC-1
3) 3.78 stars MPEG-2

Why is VC-1 so much better? You've got me there! I really don't know.

Hmm Maybe cause of all formats VC-1 has the most use. Its almost exclusively used on HD-DVD, Used by Warner in BD, while LG and Disney have also both used it.

Mpeg2 is almost completely used by Sony, Highly used by LG, it alternates on Fox, Disney, and Paramount movies.

Mpeg4 is used By Disney but not all of the time, Fox but not all of the time, and sometimes by LG and Paramount.

The fack that VC-1 can take up 60-70% of the Codec usage and still maintain an almost 4 star rating deserves some credit. Specially considering how many older catalog movies done on HD-DVD that sometimes because of their lack of success don't get the treatment all movies deserve.

The Big help with Mpeg4 is the few movies that have used it have been big titles on the BD side and therefore have received extra treatment.

By those numbers AVC is kicking vc-1 ass more than vc-1 is kicking mpeg-2's.
One should keep in minds, these reviewers, like us, are judging PQ, not codec performace, old films tend to get lower scores simply because they are not as sharp/clean as newer ones. In addition, there is director's intent to be taken into consideration.

To really evaluate a codec's performance, one should really compare encodes with their masters, I doubt any of these reviewers have access to masters.

I don't really care about codecs, as long as PQ is top notch. If PQ is same, I prefer low bit rate, smaller file size makes it easier for HTPC.

How is 124 VC1 titles out of a total of 680 discs 60%-70%?

Hmm Maybe cause of all formats VC-1 has the most use. Its almost exclusively used on HD-DVD, Used by Warner in BD, while LG and Disney have also both used it.

Mpeg2 is almost completely used by Sony, Highly used by LG, it alternates on Fox, Disney, and Paramount movies.

Mpeg4 is used By Disney but not all of the time, Fox but not all of the time, and sometimes by LG and Paramount.

The fack that VC-1 can take up 60-70% of the Codec usage and still maintain an almost 4 star rating deserves some credit. Specially considering how many older catalog movies done on HD-DVD that sometimes because of their lack of success don't get the treatment all movies deserve.

The Big help with Mpeg4 is the few movies that have used it have been big titles on the BD side and therefore have received extra treatment.

That sounds like a lot of grasping to me. The fact that MPEG-2 has all those crappy titles from the first few months yet still holds a 3.78 star rating is credit to it to say the least. IN fact it's fair to say that all the codecs deserve credit, is it not?

Youmention catalogue titles, but as pointed out I jhaven't left anything off the MPEG-2 side either; not T5e, not Kiss of the Dragon or whatever it's called, not HoFD, not the unreleased yet reviewed Robocop.

If it's only big movies on the BD side that have used MPEG-4 then how come the number of MPEG-4 relviews equals the number of VC-1 reviews?

How is 124 VC1 titles out of a total of 680 discs 60%-70%?

It wasnt 680 disc

It was 680 reviews. (doubble dipping)

Not sure about the numbers but majority of the HD DVD titles is VC1 and on the BD side you got Warners that does VC1 and some disney titles. So they should have over 50% VC1 titles.

[QUOTE=rdjam]
And they complain with right too. I work for a European PayTV Channel, and we have yet to come across a single HD movie (delivered on Sony HDCam tapes) that looks good. HDCam has a bitrate of over 100mbs, but any HD-DVD or Blu-ray disc is far superior to what's being delivered to broadcasters.
HDCAM is not 1080p 4:2:0 but 1440*1080 3:1:1 so no wonder it does not look as good as a good HD disc even with higher bit rate.

[QUOTE=HDJK]
HDCAM is not 1080p 4:2:0 but 1440*1080 3:1:1 so no wonder it does not look as good as a good HD disc even with higher bit rate.

Even if they have the same number of Chroma pixels 960*540 vs 480*1080

The main difference is that HDCAM is and intraframe format, wich means each frame is encoded on its own. Thats really throws away the bits.

Plus HD discs have 33% more lumapixels then HDCAM.

The biggest problem for HD Discs would be to encode a HD CAM master.

1440*1080 (480*1080) trough 1920*1080 (960*540) = 1440*1080 (480*540)

It wasnt 680 disc

It was 680 reviews. (doubble dipping)

Not sure about the numbers but majority of the HD DVD titles is VC1 and on the BD side you got Warners that does VC1 and some disney titles. So they should have over 50% VC1 titles.

Actually if you want to count double-dipping that just makes it worse as there's far more VC1 titles that are dual format than there's mpeg2 and AVC ones. Heck, I'll be generous. Lets for the sake of argument assume that all the mpeg2 and AVC titles are dual format and none of the VC1. So you then have 124 VC1 titles out of a total of 402 - that's still not 60%-70%, or even 35%.

That sounds like a lot of grasping to me. The fact that MPEG-2 has all those crappy titles from the first few months yet still holds a 3.78 star rating is credit to it to say the least. IN fact it's fair to say that all the codecs deserve credit, is it not?

Youmention catalogue titles, but as pointed out I jhaven't left anything off the MPEG-2 side either; not T5e, not Kiss of the Dragon or whatever it's called, not HoFD, not the unreleased yet reviewed Robocop.

If it's only big movies on the BD side that have used MPEG-4 then how come the number of MPEG-4 relviews equals the number of VC-1 reviews?

My response to that would be the age of titles used compaired on both formats and the codec used. The fact is very few catalog titles use Mpeg4 and the older it is the more unlikely they used it. Not saying they can't or that those movies won't look great but but its just been the way things have worked. Two studios really seem to look at their catalog list more then anyone else Warner and Universal, Warner treats classics like they deserve, but both and anyone else tend to show a little less care on non-mainstream titles. They still look great but not GP or Robin Hood type levels even for semi recent movies and especially comedies. So VC-1 through Universal and Warner by far has more older and not cared for as much catalog titles second being Mpeg2 which also has the distinction of having the worst HD releases attached to its name like you mentioned.

Finally its about efficiency, as I am sure all of them can look exactly the same given both effort and bandwidth. The Problem is A.) as everyone has stated storage and bandwidth is important why use a codec that requires you to use more of it if something else can do it just as good with less. B.) Some Codecs also require more effort so you have to manage time resources.

How is 124 VC1 titles out of a total of 680 discs 60%-70%?

Where did you get that number Hell Universal alone has almost 100 movies and only one is not VC-1. All warner movies are VC-1 and all but 1 Paramount movie is VC-1.

Since titles released are about the same you are saying that out of 200+ release movies by HD-DVD group that only just about half of them are VC1. Who is releasing these non VC-1 movies.

Actually if you want to count double-dipping that just makes it worse as there's far more VC1 titles that are dual format than there's mpeg2 and AVC ones. Heck, I'll be generous. Lets for the sake of argument assume that all the mpeg2 and AVC titles are dual format and none of the VC1. So you then have 124 VC1 titles out of a total of 402 - that's still not 60%-70%, or even 35%.

I really don't know where you are pulling numbers from but non of you math, or even what you are saying is making sense.

Actually if you want to count double-dipping that just makes it worse as there's far more VC1 titles that are dual format than there's mpeg2 and AVC ones. Heck, I'll be generous. Lets for the sake of argument assume that all the mpeg2 and AVC titles are dual format and none of the VC1. So you then have 124 VC1 titles out of a total of 402 - that's still not 60%-70%, or even 35%.

I wasnt talking about double format releases.

I was refering to that they counted different reviews of the same movie.


If you include ALL the scored Blu-ray reviews from High Def Digest, Home Theater Spot, Home Theater Forum, DVD Talk and Upcoming discs, including all the MPEG-2 releases from the first few months of Blu-ray's launch, and include titles that never even got released like the 1 star rated Robocop, you get the following figures.

Of course different reviews were used - 5 sites.

If you count just one site there's far too much subjectivity involved. I'll post the list later.

In real life we have seen reference quality encodings in MPEG-2, MPEG-4 AVC, and VC-1 on Blu-ray. All three video codecs have shown themselves capable of reference quality encodings.Ah, yes. Each HAS shown at least one brilliant release. But VC1 has been brilliant for the majority of releases - unlike Mpeg2...

Well if this debate was about efficiency I would agree with that but you have been debating that MPEG-2 can not look as good as VC-1.Not true. I have not said that Mpeg2 cannot look as good as VC1. I have said that VC1 releases consistently look better than Mpeg2 releases, on the HD and BD formats. Which is absolutely true.

VC-1 version of Dreamgirls (http://hddvd.highdefdigest.com/dreamgirls.html) and the MPEG-2 version of Dreamgirls (http://bluray.highdefdigest.com/dreamgirls.html) were considered reference quality .Which is why I have said that I am looking forward to seeing screen captures on Xylon's thread, so that we can see what the very best Mpeg2 presentation of Bluray looks like in comparison to VC1. ;) Certainly, one would not expect to see Mpeg2 blocking artifacts on this STERLING bluray release, then... correct?

I'm not sure that it is relevant at all rdjam. The majority of the releases you're speaking of come from June-October 2006 and we're not living in the past still are we? Next you'll be telling us that most BD50's are really BD25's as BD50's don't exist.Clever rhetoric - but the fact is, those BD25 single layer discs and Mpeg BD releases have not gone away - much as you'd like to pretend that they are not relevant. They each make up about 70% of the purchase choices in the Bluray format. I've never said that BD 50's don't exist, tho I HAVE said they are a pain in the tush to make, and have problems (even today).

Of 680 reviews: 431 MPEG-2, 125 AVC/MPEG-4, 124 VC-1
Average star rating

1) 4.22 stars AVC/MPEG-4
2) 3.94 stars VC-1
3) 3.78 stars MPEG-2

Why is VC-1 so much better? You've got me there! I really don't know.Again, this is just clever wording. How come there are 435 Mpeg2 reviews in your sample, compared to only 124 VC1 reviews? Given that most HD VD releases are in VC1.

Answer: Simple... Much of your reviews included in your "total" come from 2 quite biased "review sites" which primarily review Bluray titles and have much fewer HD DVD titles. So your numbers become much less about a scientific sample, than about a sample that has been "weighted" towards Bluray.

An anecdotal point, though: I was looking at these "Bluray reviews score higher than HD DVD reviews" threads on blu-ray.com over the last few weeks, and I noticed that some time back, the average PQ score of HD DVD went back into the lead, so I wondered how long it would tak for someone to start "disecting" some small subset of the Bluray scores to try to make Bluray look better.

The fact is, that HD DVD scores higher overall on PQ than Bluray. So try trying to separate a subset of Bluray reviews to make it look better will be seen as a transparent attempt to do just that.

Of course different reviews were used - 5 sites.

If you count just one site there's far too much subjectivity involved. I'll post the list later.Sure you will. The fact is that if one looks at the 200-odd "discs" in each format and compares PQ, HD DVD comes out way ahead. But by using multiple reviews for certain discs in your "sample" you end up, whether deliberately or not, fudging the numbers.

And when one realizes that the Bluray-biased sites mostly only have Bluray reviews, one can EASILY see where the BD scores get a boost. Mixing the sites this way simple invites criticism. If you compare the average scores of each format *by Site*, you will have a MUCH more objective result.

Oops - but then we would see which site sticks out like a sore thumb with very high BD PQ scores...

To reinforce my point:

In your last posts on the matter on bluray .com (which you keep in your signature file), HD DVD scored a 3.91 on PQ, which beat bluray's PQ average.

So in you next post, you "chose" to only look at "the last 30 releases" - and with this "sample", you were able to bring the HD DVD PQ average down to 3.76, and again were able to "prove" that Bluray was superior. This is hilarious...

BTW - it's really strange that bluray is so much better, considering that HD DVD is pretty much kicking its tail in almost every screen capture comparison... ;)

http://forum.blu-ray.com/showpost.php?p=36615&postcount=22

Rdjam, if it was single layer HD-DVD vs single layer BRD for all those mpeg-2 vs vc-1 comparisons how do you think that would've played out?

It should also be noted that the context to rdjam's statement should also include that the majority of titles in the screenshot thread are chosen such that the mpeg2 program size examples are compressed greater than the vc-1 program sizes, by an average factor of 0.9x, and going as far as 0.57x. Should it be a mystery to anyone that mpeg2 does not look all that good when compressed to vc-1 levels, let alone beyond vc-1 levels?

To say that the comparison thread is rigged, would be putting it kindly.

To say that the comparison thread is rigged, would be putting it kindly.

I can hardly see why it would be Xylon's fault if BD puts out many examples of MPEG-2 releases with insufficient bit rate.

- Tom

I think Hanky was referring to this thread... I think. Xylon's thread is certainly not rigged. It's showing us the horrors of bad decisions.

Let's not forget that STUDIOS put out those lousy mpeg-2 encodes that use a pathetic amount of the discs capacity. Studios like Warner Bros, Paramount, and even Sony. They could've given us vc-1 or avc but they went out of their way to give us inferior products. A shame too because there's a few movies that got the SL/mpeg-2 treatment that I wanted to buy. Now they'll just sit on the shelf at Best Buy.

BD fans need to complain to the studios that are releasing poor quality transfers! Don't blame VC-1 if it looks better!

Rdjam, if it was single layer HD-DVD vs single layer BRD for all those mpeg-2 vs vc-1 comparisons how do you think that would've played out?But it wasn't, was it? I don't know of any movies that are BD Dual Layer and HD DVD Single layer, with the exception of Chronos (and we all know that story). The comparisons that Xylon has done have been based on what is actually available in the real world, not what "one side" or "the other" would "prefer" to have compared.

I can hardly see why it would be Xylon's fault if BD puts out many examples of MPEG-2 releases with insufficient bit rate.

- TomExactly right.

I think Hanky was referring to this thread... I think. Xylon's thread is certainly not rigged. It's showing us the horrors of bad decisions.This thread is based on Xylon's, just to avoid claims of bias.

All shots here came from Xylon's captures, with the exception of the Serenity shots, which came from Madshi's. I don't think Madshi's thread is biased either.

BD fans need to complain to the studios that are releasing poor quality transfers! Don't blame VC-1 if it looks better!
Too right! :)

But it wasn't, was it? I don't know of any movies that are BD Dual Layer and HD DVD Single layer, with the exception of Chronos (and we all know that story). The comparisons that Xylon has done have been based on what is actually available in the real world, not what "one side" or "the other" would "prefer" to have compared.
Yes, but this thread isn't titled "Why HD DVD 30 is better than BD 25".

It's called "Why is VC-1 so much better?". VC-1 may well be better but matching it against ill-advised MPEG-2 transfers isn't proving much.

It's like matching a heavyweight champion boxer against a flyweight. The outcome is never in doubt but it's impossible to make a useful assessment of the champions skills.

Yes, but this thread isn't titled "Why HD DVD 30 is better than BD 25".

It's called "Why is VC-1 so much better?". VC-1 may well be better but matching it against ill-advised MPEG-2 transfers isn't proving much.

It's like matching a heavyweight champion boxer against a flyweight. The outcome is never in doubt but it's impossible to make a useful assessment of the champions skills.

Comparing a lot of better VC-1 transfers against poor MPEG-2 transfers does not much tell us why VC-1 is better. But with enough of those comparisons it certainly does start to provide evidence that it is.

The next time you have a choice of buying a 30 GB VC-1 or a 25 GB SL MPEG-2 hidef disc are you really going to state it probably doesn't matter? (yes, BD50 excluded here, there's enough space for MPEG-2 there)

- Tom

Comparing a lot of better VC-1 transfers against poor MPEG-2 transfers does not much tell us why VC-1 is better. But with enough of those comparisons it certainly does start to provide evidence that it is.

The next time you have a choice of buying a 30 GB VC-1 or a 25 GB SL MPEG-2 hidef disc are you really going to state it probably doesn't matter? (yes, BD50 excluded here, there's enough space for MPEG-2 there)

- Tom
When did I say it didn't matter?

30 GB VC1 HD DVD vs 25 GB MPEG BD should a walkover for VC1 so posting comparisons of this here harldy breaks new ground or tells us anything that we didn't know. this would be fine if the thread title wasn't "Why is VC1 so much better?".

The thread title implies that VC1 one is better than any other codec yet the comparisons are all loaded in it's favour. Why not compare World Trade Centre? This might be more interesting as it's MPEG2 on BD50. I also look forward to some VC1 vs AVC comparisons. These may actually go towards making a case for VC1 being better.

World Trade Center is coming.

I will try my best (with everyones help) to compare BD50 against HDDVD30.

Thank you, Xylon.

World Trade Center is coming.

I will try my best (with everyones help) to compare BD50 against HDDVD30.Cool, what are bitrates and actual file sizes for them?

Cool, what are bitrates and actual file sizes for them?

I'm still waiting for my copies to arrive. One thing is for sure BD versions will have the edge in size.

I just hope it's not a repeat of the We Were Soldiers fiasco where the mpeg-2 is too artifacty and the vc-1 is filtered to hell. We lose.

filtered to hell.
Wanna exaggerate a bit more?

Yes, but this thread isn't titled "Why HD DVD 30 is better than BD 25".

It's called "Why is VC-1 so much better?". VC-1 may well be better but matching it against ill-advised MPEG-2 transfers isn't proving much.

It's like matching a heavyweight champion boxer against a flyweight. The outcome is never in doubt but it's impossible to make a useful assessment of the champions skills.

You make too much sense.

;) Thanks.

Yes, but this thread isn't titled "Why HD DVD 30 is better than BD 25".

It's called "Why is VC-1 so much better?". VC-1 may well be better but matching it against ill-advised MPEG-2 transfers isn't proving much.

It's like matching a heavyweight champion boxer against a flyweight. The outcome is never in doubt but it's impossible to make a useful assessment of the champions skills.
Actually, I'm not the one who has taken the discussion off topic - I've merely responded to the previous posts.

The thread is still about VC1 - the only problem is that because VC1 is normally used on HD DVD titles, we've been hearing a lot about how the "other" format might be able to beat VC1 HD DVDs with higher bandwidths on dual layer BD discs.

As for "flyweights", I'm not sure I would have believed that BD folks would have described 70% of BR releases in those terms, but hey ho...

I just hope it's not a repeat of the We Were Soldiers fiasco where the mpeg-2 is too artifacty and the vc-1 is filtered to hell. We lose.
Actually - it seemed pretty obvious that the problem with WWS was simply a "not very good" master. Both BD and HD releases suffered from this. So much so, that there were even Mpeg blocking artifacts in the VC1 version which seemed to come straight from said same master.

World Trade Center is coming.

I will try my best (with everyones help) to compare BD50 against HDDVD30.Are you planning on doing a comparison with "Flags of our Fathers" (which would have advanced codecs for both)?

Thanks,
Darin

Actually, I'm not the one who has taken the discussion off topic - I've merely responded to the previous posts.

The thread is still about VC1 - the only problem is that because VC1 is normally used on HD DVD titles, we've been hearing a lot about how the "other" format might be able to beat VC1 HD DVDs with higher bandwidths on dual layer BD discs.


I don't think you understood Kid Banana's concerns. You're comparing HD DVD 30 with BD 25, and THEN making the conclusion that VC-1 is better (the very title of your thread). Do you forget?

I don't think you understood Kid Banana's concerns. You're comparing HD DVD 30 with BD 25, and THEN making the conclusion that VC-1 is better (the very title of your thread). Do you forget?
I trust that you haven't forgotten that there are also comparisons against a variety of other situations too, including high bitrate AVC from Serenity, near the Opening Post ;)

I trust that you haven't forgotten that there are also comparisons against a variety of other situations too, including high bitrate AVC from Serenity, near the Opening Post ;)

So you finally agree that HD DVD 30 vs BD 25 comparisons are invalid for concluding that VC1 is "so much better"?

Actually - it seemed pretty obvious that the problem with WWS was simply a "not very good" master. Both BD and HD releases suffered from this. So much so, that there were even Mpeg blocking artifacts in the VC1 version which seemed to come straight from said same master.

How do you know they were mpeg-2 blocking artifacts?

So you finally agree that HD DVD 30 vs BD 25 comparisons are invalid for concluding that VC1 is "so much better"?
Well, it's hard to say that, so I don't think I would.

There are SOME who still claim that VC1 only has a 10% advantage over Mpeg, and some of the Mpeg BD25 discs out there actually have higher data rates than the VC1 versions.

So I think your "inference" that "all BD25 Mpeg2 releases are lower bitrates", and therefore should not be compared, is false...

VC1 has beaten higher bitrate Mpeg releases in these comparisons also...

How do you know they were mpeg-2 blocking artifacts?
Blocking artifacts are not hard to spot. They are big and square. Take a look at the nose of the helicopter, in front of the windshield, on the HD DVD version. The blocking artifacts there are also present in the Mpeg version (along with a few extra artifacts).

Obvious (large) macroblocking like that is atypical of VC1 encodes, which will typically have smaller-sized blocks if pushed (and usually only after an encoding error). Seeing as how these large blocks are present in the exact same spot on the Mpeg version, it is a very logical conclusion that it was in the master.

Obvious (large) macroblocking like that is atypical of VC1 encodes, which will typically have smaller-sized blocks if pushed (and usually only after an encoding error). Seeing as how these large blocks are present in the exact same spot on the Mpeg version, it is a very logical conclusion that it was in the master.Blocks in the master? Give me a break.

The video quality of the AVC and VC1 videos tends to alternate from pic to pic. Sometimes AVC is better, sometimes it's VC1. Frankly, it's not enough to really conclude that VC1 is better.

Theoretically AVC is the better codec based upon it's more advanced feature set.

However, what I've heard is that Microsoft distributes VC1 encoding packages to production houses that are equipped with all sorts of tools for the developers to use along with support. I don't believe there is currently any H.264/AVC packages that are as good as Microsofts right now. Production houses are just as interested in the tools and support that come with the codecs as they are in the codecs themselves.

However, what I've heard is that Microsoft distributes VC1 encoding packages to production houses that are equipped with all sorts of tools for the developers to use along with support.Did you check this thread
http://www.avsforum.com/avs-vb/showthread.php?t=840193
in which a Microsoft guy said VC-1 tools lack a de-banding filter?

Well, it's hard to say that, so I don't think I would.

There are SOME who still claim that VC1 only has a 10% advantage over Mpeg, and some of the Mpeg BD25 discs out there actually have higher data rates than the VC1 versions.

So I think your "inference" that "all BD25 Mpeg2 releases are lower bitrates", and therefore should not be compared, is false...

VC1 has beaten higher bitrate Mpeg releases in these comparisons also...

You're shifting your stance ever so slightly again.

What I infer is indeed that low bitrate MPEG2 should not be compared to high bitrate VC1 encodes. It is a farcical practice.

If some claimed that VC1 had only a 10% advantage over Mpeg2, you have not shown it to be false. If "some Mpeg BD25 discs out there actually have higher rates than the VC1 versions", it is irrelevant if you hadn't actually used these specific examples for your comparisons.

Where you have provided the bitrates, you pitted an 11mpbs Mpeg2 source against a 17-19mpbs VC1 source, to conclude that VC1 is better. Well done, mate.

You're shifting your stance ever so slightly again. Nay, I say you are shifting yours.

So you finally agree that HD DVD 30 vs BD 25 comparisons are invalid for concluding that VC1 is "so much better"? Here you state that BD25 Mpeg releases cannot be compared to HD DVD 30 Gig VC1 releases, since you feel it is unfair.

Yet I have already stated that VC1 has won comparisons against highr bitrate mpeg encoding as well. I also refused to change my position, quite clearly, in direct response to your shapeshifting.

Yet, when I make a point that you are trying to state that ALL BD25 Mpeg releases should not be compared, EVEN when some of them are higher bitrates than the equivalent VC1 encodes, any objective person would agree that they CAN indeed be compared - especially given that the BD25 releases are ON SALE in direct competition to the equivalent VC1 releases on HD DVD.

Yet, in your reply, you dig your heels in rather than concede a valid point. However, you "shift your stance" ever (not) so slightly, by changing your wording significantly: What I infer is indeed that low bitrate MPEG2 should not be compared to high bitrate VC1 encodes. It is a farcical practice.A completely different meaning, which means you back off of your original "all encompassing" statement.

Then, having just stated that there are comparisons of VC1 where the VC1 encode was at a significantly lower bitrate than the mpeg encode, you reply: If some claimed that VC1 had only a 10% advantage over Mpeg2, you have not shown it to be false. If "some Mpeg BD25 discs out there actually have higher rates than the VC1 versions", it is irrelevant if you hadn't actually used these specific examples for your comparisons....wherein you completely ignore the fact that we are not JUST comparing VC1 to Mpeg in this thread, but also to AVC. This mpeg fixation has got to be the funniest thing. It's not the only codec available you know...

Where you have provided the bitrates, you pitted an 11mpbs Mpeg2 source against a 17-19mpbs VC1 source, to conclude that VC1 is better. Well done, mate.I suggest you read the thread again, mate. The AVC encode in the comparison you mentioned is over 19 mbps.

Frankly, all this talk about Mpeg2 is mind boggling, as it is not even CLOSE to keeping up with VC1.

The main target for the comparisons is AVC, which is touted as the direct competitior to VC1.

So lets recap:

Serenity: AVC=17.89 mbps and VC1=15.77 mbps, yet VC1 kicks butt. http://www.avsforum.com/avs-vb/showthread.php?p=10462464&&#post10462464

King Kong: The AVC and VC1 bitrates are almost identical, yet VC1 kicks AVCs butt. http://www.avsforum.com/avs-vb/showthread.php?p=10462469&&#post10462469

Mission Impossible: The BD Mpeg encode is 21.5 Gigs and VC1 is 20.1 Gigs. yet VC1 wins again. http://www.avsforum.com/avs-vb/showthread.php?p=10462466&&#post10462466

I won't talk about the Harry Potter comparison, as that one WAS a wipeout and probably not fair to compare. You'll note I didnt bother to include it in this thread, right? It was also interesting that the AVC version of Fantastic Four looked better in many respects to the Mpeg Bluray version, even though the Mpeg version had twice the bitrate. I'm looking forward to seeing this in VC1 :)

Then there was Aeon Flux BD Mpeg vs HD VC1: Guess who won that one too? http://www.avsforum.com/avs-vb/showthread.php?p=10472031&&#post10472031

Then, one of the bigger gaps, Training Day: BD filsize 15 gigs, HD size 17 gigs. But the Mpeg wasn't even close - probably would have needed 10 gigs more to match. http://www.avsforum.com/avs-vb/showthread.php?p=10601999&&#post10601999

Anyway - it sounds like Xylon will have some BD50s to compar very soon, which will give the Mpeg stuff even more "breathing room".

It'll be interesting....

Don't even try to muddle the situation.

What everyone is telling YOU is it is nonsense to compare low bitrate MPEG2s to VC1, and then claim that VC1 is better because of the results. It matters not that I used BD25 as the equivalent term for "low bitrate MPEG2" - the meaning should not be lost on anyone but the thickest, even if there are rare exceptions to the term.

Yet for the exceptions, you feel you should then be allowed to compare ALL BD25 vs HD30 regardless of bitrate (just because they are "real") - this is a complete 180 degrees to how much importance you placed in my specific use of the terms "BD25/HD30" versus "low bitrate MPEG2/high bitrate VC1".

Hooray for your examples, as we can see that VC1 needed similar or significantly higher bitrates to beat MPEG2. And of course, 17.89mbps AVC versus 15.77mbps VC1 is a significant advantage to AVC, but 17mbps AVC versus 19mbps VC1 is "almost identical" bitrates? You don't do AVC any justice. Let alone the fact that your judgement of "better" is personal.

The following quote for your convenience.

Serenity: AVC=17.89 mbps and VC1=15.77 mbps, yet VC1 kicks butt. http://www.avsforum.com/avs-vb/show...&&#post10462464

King Kong: The AVC and VC1 bitrates are almost identical, yet VC1 kicks AVCs butt. http://www.avsforum.com/avs-vb/show...&&#post10462469

Mission Impossible: The BD Mpeg encode is 21.5 Gigs and VC1 is 20.1 Gigs. yet VC1 wins again. http://www.avsforum.com/avs-vb/show...&&#post10462466

I won't talk about the Harry Potter comparison, as that one WAS a wipeout and probably not fair to compare. You'll note I didnt bother to include it in this thread, right? It was also interesting that the AVC version of Fantastic Four looked better in many respects to the Mpeg Bluray version, even though the Mpeg version had twice the bitrate. I'm looking forward to seeing this in VC1

Then there was Aeon Flux BD Mpeg vs HD VC1: Guess who won that one too? http://www.avsforum.com/avs-vb/show...&&#post10472031

Then, one of the bigger gaps, Training Day: BD filsize 15 gigs, HD size 17 gigs. But the Mpeg wasn't even close - probably would have needed 10 gigs more to match. http://www.avsforum.com/avs-vb/show...&&#post10601999

So lets recap:
Serenity: AVC=17.89 mbps and VC1=15.77 mbps, yet VC1 kicks butt. http://www.avsforum.com/avs-vb/showthread.php?p=10462464&&#post10462464

Yes, let's: Real time versus >= 2 pass hand tuned. Unfair and pointless.

King Kong: The AVC and VC1 bitrates are almost identical, yet VC1 kicks AVCs butt. http://www.avsforum.com/avs-vb/showthread.php?p=10462469&&#post10462469

Ditto.

Mission Impossible: The BD Mpeg encode is 21.5 Gigs and VC1 is 20.1 Gigs. yet VC1 wins again. http://www.avsforum.com/avs-vb/showthread.php?p=10462466&&#post10462466

I'm so impressed. VC-1 wins against MPEG2 at same bit rate? WOW!
Show us something relevant, please. Something like AVC versus VC-1 both >= 2 pass hand tuned and about the same bit rate. Or MPEG2 at 50GB disc rates versus supposedly transparent VC-1 at 10-15 Mbit/s average.

Are you planning on doing a comparison with "Flags of our Fathers" (which would have advanced codecs for both)?

Thanks,
Darin

Yes.

Yes, let's: Real time versus >= 2 pass hand tuned. Unfair and pointless.

Ditto.

I'm so impressed. VC-1 wins against MPEG2 at same bit rate? WOW!
Show us something relevant, please. Something like AVC versus VC-1 both >= 2 pass hand tuned and about the same bit rate. Or MPEG2 at 50GB disc rates versus supposedly transparent VC-1 at 10-15 Mbit/s average.Wow. I've not read this thread yet but if what you point out is what the OP is based on, why can this thread stay here this long? Where are mods with common sense?

Don't even try to muddle the situation.

What everyone is telling YOU is it is nonsense to compare low bitrate MPEG2s to VC1, and then claim that VC1 is better because of the results. At least you are sticking to your re-stated statement. ;)

No one is deliberately selecting lower-bitrate versions of Mpeg from these Bluray discs. But if 70% of Bluray discs are Single Layer BD25, it makes it a bit hard for folks like Xylon to find matching releases that meet "your" requirments.

Essentially, you would prefer that no one compare these BDs with HD DVDs because it would be unfair, yet I see they are chargin the same price for these releases...

not be lost on anyone but the thickeststeady on with the personal stuff there, big fella... not only do I understand the sentence, but I have pointed out how it differed from your original statement. But again, these folks are doing their best to make meaningful comparisons- all I'm doing is taking those and blowing them up to show the encoding differences vis-a-vis VC1. No need to take it out on me personally.

Yet for the exceptions, you feel you should then be allowed to compare ALL BD25 vs HD30 regardless of bitrate (just because they are "real") - this is a complete 180 degrees to how much importance you placed in my specific use of the terms "BD25/HD30" versus "low bitrate MPEG2/high bitrate VC1".And yet some of the comparisons were very comparable bitrates, with the edge going to Mpeg rates in some. What would seem fair to you, DOUBLE the bitrate for Mpge, or something?

Never fear, more comparisons are coming which I'm sure will meet your desires...

Then, one of the bigger gaps, Training Day: BD filsize 15 gigs, HD size 17 gigs. But the Mpeg wasn't even close - probably would have needed 10 gigs more to match. http://www.avsforum.com/avs-vb/showthread.php?p=10601999&&#post10601999

It's 13.8gig vs 17.1gig, but why let the facts get in the way. The comparison proves nothing other than VC1 is better than MPEG2 when VC1 has 20% larger bitrate. This is not news to anyone here.

Anyway - it sounds like Xylon will have some BD50s to compar very soon, which will give the Mpeg stuff even more "breathing room".
MORE breathing room? Which of your quoted examples have given MPEG2 20% headroom over the VC-1 version like the advantage that VC-1 had in Training Day?

I look forward to furher comparisons and also wish to pass on my apprecaition for all the hard work Xylon puts into these. Many thanks Xylon.

No one is deliberately selecting lower-bitrate versions of Mpeg from these Bluray discs. But if 70% of Bluray discs are Single Layer BD25, it makes it a bit hard for folks like Xylon to find matching releases that meet "your" requirments.
No one here is critising Xylon. I'm sure everyone is grateful for the hard work he puts in.

Essentially, you would prefer that no one compare these BDs with HD DVDs because it would be unfair, yet I see they are chargin the same price for these releases...

There is a separate thread where Xylons screen captures show us the differences between the various releases and codecs. It is very informative.

What I, and most other posters are objecting to, is that you are extracting comparisons from that thread and placing them in your 'Why VC-1 is better' thread. the comparisons you are using are not providing useful data for your argument.

If someone here started a thread called 'Why AVC at 40 mbs is better than VC1 at 15mbs' you would rightly claim this as an unfair comparison and yet you are using a similar tactic in this VC-1 advocacy thread (I'm aware that this is an extreme analogy but I'm illustrating a point, I'm not literally saying Rdjam is posting 40mbc VC-1 against 15mbs Mpeg2).

I think VC-1 is a very good codec but the 'evidence' in this thread isn't proving it's superiority as the competing codecs have not received similar optimisation to VC1.

At least you are sticking to your re-stated statement. ;)

No one is deliberately selecting lower-bitrate versions of Mpeg from these Bluray discs. But if 70% of Bluray discs are Single Layer BD25, it makes it a bit hard for folks like Xylon to find matching releases that meet "your" requirments.

It should be obvious that "my" requirements are not unreasonable. The lack of better samples does not make the current ones valid for YOUR purposes. Xylon's hard work was not intended to push your agenda.


Essentially, you would prefer that no one compare these BDs with HD DVDs because it would be unfair, yet I see they are chargin the same price for these releases...

So now it seems that the basis for a valid comparison is price. Let's forget bitrates.


steady on with the personal stuff there, big fella... not only do I understand the sentence, but I have pointed out how it differed from your original statement. But again, these folks are doing their best to make meaningful comparisons- all I'm doing is taking those and blowing them up to show the encoding differences vis-a-vis VC1. No need to take it out on me personally.

No, you refused to understand my original statement.


And yet some of the comparisons were very comparable bitrates, with the edge going to Mpeg rates in some. What would seem fair to you, DOUBLE the bitrate for Mpge, or something?

woah, some edge there to Mpeg2. You don't even do Mpeg2 any justice.

I'm glad others are keeping a watchful eye on this development, as much as me! :cool:

Clever rhetoric - but the fact is, those BD25 single layer discs and Mpeg BD releases have not gone away - much as you'd like to pretend that they are not relevant. They each make up about 70% of the purchase choices in the Bluray format. I've never said that BD 50's don't exist, tho I HAVE said they are a pain in the tush to make, and have problems (even today).

Again, this is just clever wording. How come there are 435 Mpeg2 reviews in your sample, compared to only 124 VC1 reviews? Given that most HD VD releases are in VC1.

Answer: Simple... Much of your reviews included in your "total" come from 2 quite biased "review sites" which primarily review Bluray titles and have much fewer HD DVD titles. So your numbers become much less about a scientific sample, than about a sample that has been "weighted" towards Bluray.

An anecdotal point, though: I was looking at these "Bluray reviews score higher than HD DVD reviews" threads on blu-ray.com over the last few weeks, and I noticed that some time back, the average PQ score of HD DVD went back into the lead, so I wondered how long it would tak for someone to start "disecting" some small subset of the Bluray scores to try to make Bluray look better.

The fact is, that HD DVD scores higher overall on PQ than Bluray. So try trying to separate a subset of Bluray reviews to make it look better will be seen as a transparent attempt to do just that.

Sure you will. The fact is that if one looks at the 200-odd "discs" in each format and compares PQ, HD DVD comes out way ahead. But by using multiple reviews for certain discs in your "sample" you end up, whether deliberately or not, fudging the numbers.

And when one realizes that the Bluray-biased sites mostly only have Bluray reviews, one can EASILY see where the BD scores get a boost. Mixing the sites this way simple invites criticism. If you compare the average scores of each format *by Site*, you will have a MUCH more objective result.

Oops - but then we would see which site sticks out like a sore thumb with very high BD PQ scores...


Boy those BD50's must sure be a pain to make eh?
January - 30 Blu-ray's released, 14 of them were BD50's.

You try haplessly to reinforce a point you made that wasn't true. You've stated that the two sites that are biased to Blu-ray have, let me quote you "And when one realizes that the Bluray-biased sites mostly only have Bluray reviews, one can EASILY see where the BD scores get a boost"

I'd now like you to explain that for all of us please. YOu can see how many reviews are from each site in this chart.

HDD 132HD/169BD, HTS 137HD/167BD, HTF 36HD/59BD, UD 101HD/89BD, Talk 172HD/203BD
04.25.07
HD DVD PQ SQ TOTAL Blu-ray PQ SQ TOTAL
HighDef 3.95 3.64 3.80 HighDef 3.96 3.85 3.90
HTSpot 4.02 3.88 3.95 HTSpot 3.96 4.23 4.09
DVDTalk 3.66 3.51 3.58 DVDTalk 3.60 3.67 3.63
HTForum 4.13 3.76 3.94 HTForum 4.19 4.08 4.14
UpDisc 4.03 3.87 3.95 UpDisc 3.99 4.16 4.08
Totals 3.91 3.70 3.80 Totals 3.88 3.95 3.91

The only site I see that DOESN'T have more Blu-ray reviews than HD DVD is Upcoming Discs, so are you suggesting that all four of the other sites are biased?

Well, presuming that Upcoming Discs is one of the sites you're talking about, which of the other ones is biased to Blu-ray?

High Def Digest - Blu-ray ahead by 0.01
Home Theater Spot - HD DVD ahead by 0.06
Home Theater Forum - Blu-ray ahead by 0.06
Upcoming Discs - HD DVD ahead by 0.04
DVD Talk - HD DVD ahead by 0.06
Total PQ - HD DVD ahead by 0.03

Wow rdjam, I thought you said that if you compared all the reviews by site, HD DVD would be far ahead in quality! Isn't that what you said? The fact is that if one looks at the 200-odd "discs" in each format and compares PQ, HD DVD comes out way ahead. But by using multiple reviews for certain discs in your "sample" you end up, whether deliberately or not, fudging the numbers.


So if you take each site individually, which one is it that you note as being "Far ahead" in overall ratings for HD DVD?

Which site do you note as being "sticking out like a sore thumb" in favour of Blu-ray?

Is it fair to say you were just assuming without actually taking the time to research it yourself? Or was it just FUD?

Also you took exception to the fact that there were so few VC-1 reviews compared to the MPEG-2 releases. Dear fellow it was clearly explained that the comparison was for Blu-ray reviews only. That much should have been clear. The codec comparison wasn't "weighted towards Blu-ray"....it was ONLY about Blu-ray.

Not true. I have not said that Mpeg2 cannot look as good as VC1.Which is not the impression I have gotten from reading your threads/posts against MPEG-2 and MPEG-4 AVC. As such are saying that you believe that all three video codecs are capable of reference quality encodings?


I have said that VC1 releases consistently look better than Mpeg2 releases, on the HD and BD formats. Which is absolutely true.Individually that depends on the title, but overall VC-1 does have a slightly better record than MPEG-2.


Which is why I have said that I am looking forward to seeing screen captures on Xylon's thread, so that we can see what the very best Mpeg2 presentation of Bluray looks like in comparison to VC1. ;) Certainly, one would not expect to see Mpeg2 blocking artifacts on this STERLING bluray release, then... correct?Actually compression artifacts can be found even on reference quality encodings, regardless of video codec, if you are willing to spend enough time looking for them. They tend to be small artifacts though that aren't noticeable during normal viewing.


Wow. I've not read this thread yet but if what you point out is what the OP is based on, why can this thread stay here this long?Many posters have mentioned the absurdity of comparing multi-pass HD DVD encodings to real time broadcast encodings but markrubin didn't seem to have any problem with it (http://www.avsforum.com/avs-vb/showthread.php?p=10463127&&#post10463127). Personally I don't understand why though since such unfair comparisons are in my opinion rather misleading.

Show us something relevant, please. Something like AVC versus VC-1 both >= 2 pass hand tuned and about the same bit rate.

FWIW it seems that Paramount has migrated to AVC across the board for its Blu-ray releases: Flags of Our Fathers, Norbit and Trading Places are all AVC.

Boy those BD50's must sure be a pain to make eh?
January - 30 Blu-ray's released, 14 of them were BD50's.

You try haplessly to reinforce a point you made that wasn't true. You've stated that the two sites that are biased to Blu-ray have, let me quote you "And when one realizes that the Bluray-biased sites mostly only have Bluray reviews, one can EASILY see where the BD scores get a boost"

I'd now like you to explain that for all of us please. YOu can see how many reviews are from each site in this chart.

HDD 132HD/169BD, HTS 137HD/167BD, HTF 36HD/59BD, UD 101HD/89BD, Talk 172HD/203BD
04.25.07
HD DVD PQ SQ TOTAL Blu-ray PQ SQ TOTAL
HighDef 3.95 3.64 3.80 HighDef 3.96 3.85 3.90
HTSpot 4.02 3.88 3.95 HTSpot 3.96 4.23 4.09
DVDTalk 3.66 3.51 3.58 DVDTalk 3.60 3.67 3.63
HTForum 4.13 3.76 3.94 HTForum 4.19 4.08 4.14
UpDisc 4.03 3.87 3.95 UpDisc 3.99 4.16 4.08
Totals 3.91 3.70 3.80 Totals 3.88 3.95 3.91

The only site I see that DOESN'T have more Blu-ray reviews than HD DVD is Upcoming Discs, so are you suggesting that all four of the other sites are biased?

Well, presuming that Upcoming Discs is one of the sites you're talking about, which of the other ones is biased to Blu-ray?

High Def Digest - Blu-ray ahead by 0.01
Home Theater Spot - HD DVD ahead by 0.06
Home Theater Forum - Blu-ray ahead by 0.06
Upcoming Discs - HD DVD ahead by 0.04
DVD Talk - HD DVD ahead by 0.06
Total PQ - HD DVD ahead by 0.03

Wow rdjam, I thought you said that if you compared all the reviews by site, HD DVD would be far ahead in quality! Isn't that what you said? The fact is that if one looks at the 200-odd "discs" in each format and compares PQ, HD DVD comes out way ahead. But by using multiple reviews for certain discs in your "sample" you end up, whether deliberately or not, fudging the numbers.


So if you take each site individually, which one is it that you note as being "Far ahead" in overall ratings for HD DVD?

Which site do you note as being "sticking out like a sore thumb" in favour of Blu-ray?

Is it fair to say you were just assuming without actually taking the time to research it yourself? Or was it just FUD?

Also you took exception to the fact that there were so few VC-1 reviews compared to the MPEG-2 releases. Dear fellow it was clearly explained that the comparison was for Blu-ray reviews only. That much should have been clear. The codec comparison wasn't "weighted towards Blu-ray"....it was ONLY about Blu-ray.

It is safe to assume that rdjam did not do any form of research in his statements. It is only pure passion and making figures up to make HD-DVD look better so that the casual forum onlooker will read it and believe it to be true. Unfortunately this happens quite a lot and the good thing about this forum is we occasionally have people who set those make-believers in their place by showing them actual indisputable facts.

It is safe to assume that rdjam did not do any form of research in his statements. It is only pure passion and making figures up to make HD-DVD look better so that the casual forum onlooker will read it and believe it to be true. Unfortunately this happens quite a lot and the good thing about this forum is we occasionally have people who set those make-believers in their place by showing them actual indisputable facts.

I wouldn't call PQ marks by so-called "professional reviews" from highdefdigest & Co "indisputable facts" though ;)


FWIW Buena Vista/Disney's recent Apocalypto release uses high-bitrate (~26MBit/s avg., peaks at around 40MBit/s +/- buffer) AVC for the main feature and high-bitrate (about the same avg. and peak rates) VC-1 for the documentary... You go figure.

VERY ODD decision indeed considering "VC-1 is so much more better", isn't it :confused:

Buena Vista/Disney seems to be the ONLY unbiased (=unrestricted by exclusivity contracts) studio (with the exception of Paramount perhaps) when it comes to codec choice. Lately they clearly prefer AVC for their high profile releases, although VC-1 is undoubtedly "so very much more better" :p

It is very comforting to have such knowledgeable members like rdjam who just know everything better than some major studios - nevertheless these studios somehow manage to release one reference title after another with subpar codecs (look at PotC I+II, Apocalypto. These releases rival (to say the least) King Kong and I-Robot (cbr Mpeg2 :eek: DVHS, can you believe it?)).

btw rdjam, I suggest you watch PotC: DMC . It will be a release after your taste - lots of DNR applied in order to make it much more "smooth" and clean looking compared to the (IMO outstanding, in your opinion awfully grainy) HD trailer. PotC I looks "way better" - wonderfully preserved film grain -> it still looks like film and not as clean as your daily soap for a pleasant change.

Same goes for Matrix on HD-DVD btw - wonderful "film-like" releases (despite some very obvious Edge Enhancement and pumped up contrast in Matrix I) - not the usual clean, DNRed Warner stuff .

It's 13.8gig vs 17.1gig, but why let the facts get in the way. The comparison proves nothing other than VC1 is better than MPEG2 when VC1 has 20% larger bitrate. This is not news to anyone here.


MORE breathing room? Which of your quoted examples have given MPEG2 20% headroom over the VC-1 version like the advantage that VC-1 had in Training Day?

I look forward to furher comparisons and also wish to pass on my apprecaition for all the hard work Xylon puts into these. Many thanks Xylon.
Again - you are picking ONLY one example.

There were other examples, where Mpeg had a higher bitrate, and others where even AVC had a higher bitrate.

Folks who continuously choose to claim that all examples gave VC1 a bitrate advantage are just "innacurate", IMO.

What I, and most other posters are objecting to, is that you are extracting comparisons from that thread and placing them in your 'Why VC-1 is better' thread. the comparisons you are using are not providing useful data for your argument. So how is the info useful in one thread, but suddenly biased and useless in another? ;)

By blowing up the images this way, one is able to show how each of the codecs differ in their detail, and how they respond to pressure. There will be many more comparisons of varying bitrates, so I don't see what all the panic is about, really.

FWIW it seems that Paramount has migrated to AVC across the board for its Blu-ray releases: Flags of Our Fathers, Norbit and Trading Places are all AVC.
Thank God for that.

The less we all see of mpeg2 on either of these new disc formats the better.

All this defending and apologising for Mpeg2 is just getting old.

Buena Vista/Disney seems to be the ONLY unbiased (=unrestricted by exclusivity contracts) studio (with the exception of Paramount perhaps) when it comes to codec choice. Lately they clearly prefer AVC for their high profile releases, although VC-1 is undoubtedly "so very much more better" :p
And there are many studios that have chosen VC1 also. In fact most HD DVD releases are VC1.

Again - you are picking ONLY one example.

There were other examples, where Mpeg had a higher bitrate, and others where even AVC had a higher bitrate.

Folks who continuously choose to claim that all examples gave VC1 a bitrate advantage are just "innacurate", IMO.

For mpeg2/vc-1 comparisons, there were 3 out of 9 where mpeg-2 had the technically higher program size. Of the 3, there were 2 where the program size was just a nick higher- essentially they were equal size scenarios (1.07x and 1.01x). So that leaves 1 where there was a clear increase in program size, and it wasn't by a particularly huge amount (1.33x).

So "1 out of 9" would seem to run counter to your claim.

FWIW Buena Vista/Disney's recent Apocalypto release uses high-bitrate (~26MBit/s avg., peaks at around 40MBit/s +/- buffer) AVC for the main feature and high-bitrate (about the same avg. and peak rates) VC-1 for the documentary... You go figure.Have you looked at that one? At around the 24 minute mark in the documentary there is a section from the movie where they show the sun. The banding is really bad. I checked the same scene in the movie (in AVC) and the bitrate there is higher and I don't see the banding. I don't have access to the master, so can't compare with that, but these look very different as far as the banding. The extra is probably 1080i60 and maybe there are other factors here, but it would be interesting to find out why they look so different (at least with my display).

--Darin

So how is the info useful in one thread, but suddenly biased and useless in another? ;)
It's because the two threads have different purposes. Xylon's thread shows us comparisons between various formats and software. The size and provenance of each capture is clearly stated. In short, it's a brilliant thread and one of the all-time best on AVS.

Your thread is called "Why is VC-1 better". Most people visiting this thread would expect to see VC-1 taking all commers and beating them. Instead what we are presented with is a series of comparisons where VC-1 often has an inbuilt advantage, i.e. beating MPEG2 when VC1 has a comprable or larger bitrate, beating realtime AVC tv broadcasts etc.

If VC-1 is better, then show it beating equally-sized and optimised AVC encodes or MPEG-2 with a comfortably higher (30% more) bitrate.

And there are many studios that have chosen VC1 also. In fact most HD DVD releases are VC1.

And is pretty much a must for HD-DVD to compete.

But of late, it looks like AVC is becoming the quality codec choice for Blu Ray (no problem with high bitrate VC-1, either will do) and optimised low bitrate VC-1 should stay on HD-DVD, it helps with the formats limitations.

I would like to see more high bitrate VC-1 encodes so we can compare them with AVC, to see if VC-1 really is all that.

...and others where even AVC had a higher bitrate.You fail to mention the fact though that when you compare VC-1 and MPEG-4 AVC you are comparing multi-pass HD DVD encodings to real time broadcast encodings. It is hardly a surprise that a multi-pass encoding made by skilled compressionists using a studio quality VC-1 encoder would do better than a real time broadcast quality MPEG-4 AVC encoding.


Folks who continuously choose to claim that all examples gave VC1 a bitrate advantage are just "innacurate", IMO.The problem is rdjam you have given VC-1 unfair advantages in most of the comparisons you have posted. Whether it is in bitrate or in comparing multi-pass HD DVD encodings to broadcast encodings. As mentioned their is only one comparison that even came close to fair, which was the Mission Impossible 3 comparison (http://www.avsforum.com/avs-vb/showthread.php?p=10462466&&#post10462466), yet most people would never even notice the differences between the two encodings. In fact most professional reviewers gave the two encodings the same picture quality score even though the MPEG-2 encoding was only 7% larger than the VC-1 encoding. Considering that Amir once said (http://www.avsforum.com/avs-vb/showthread.php?p=8345206&&#post8345206) that VC-1 was close to 3 times as efficient as MPEG-2 that is actually rather impressive for poor old MPEG-2 ;).

I think it was the MI3 comparison was +7% and WWS was +30%, fwiw. ;)

EDIT: Oops! Unless you meant a different MI comparison from the first 2 movies.

Personally I think VC1 is better than MPEG-2 simply because it can do equivalent quality in fewer bits. Since the point of compressing is to compress I think that is sufficient, at least to me.

But I have not seen anything yet that proves it is particularly better or worse by this yardstick than AVC. AVC mature encoders are arriving and beginning to be used. So I think we will still have to wait for the answer to this one.

- Tom

I think it was the MI3 comparison was +7% and WWS was +30%, fwiw. ;)Good catch, and it is only a 7% difference.

But I have not seen anything yet that proves it is particularly better or worse by this yardstick than AVC.
I think it's a safe assumption to say that AVC and VC-1 are pretty much equally capable, and both are considerably better than MPEG-2.

Personally I think VC1 is better than MPEG-2 simply because it can do equivalent quality in fewer bits. Since the point of compressing is to compress I think that is sufficient, at least to me.
So why is Rdjam using comparisons where VC-1 is given roughly equal or (n the case of Training Day) substantially larger bitrates than MPEG-2?

I'll say it again. VC-1 may well be better than MPEG-2 but these comparisons aren't highlighting where it's strengths are.

From Rdjams comparisons we learn that VC-1 will beat Mpeg-1 (if VC-1 has an equal or higher bit rate) and that 2 pass optimised VC-1 can beat an encoded-on-the-fly AVC tv broacast. No-one here ever disputed these facts.

Many posters have mentioned the absurdity of comparing multi-pass HD DVD encodings to real time broadcast encodings but markrubin didn't seem to have any problem with it (http://www.avsforum.com/avs-vb/showthread.php?p=10463127&&#post10463127). Personally I don't understand why though since such unfair comparisons are in my opinion rather misleading.

Richard

that is a bit unfair: my post was as Moderator trying to keep the peace: there is nothing wrong with a good, spirited technical discussion on these boards;

as long as they are within AVS rules: challenge the info in the post: not the poster

I have not yet posted or taken a position on the technical merits of the issue

Thank you

From Rdjams comparisons we learn that VC-1 will beat Mpeg-1 (if VC-1 has an equal or higher bit rate) and that 2 pass optimised VC-1 can beat an encoded-on-the-fly AVC tv broacast. No-one here ever disputed these facts.

It seems some Hollywood studios either disputed or were unaware of the first part (assuming they actually cared) since they proceeded to release some single sided MPEG-2 BD's of full length movies. Most here would probably admit that was ill advised.

But point taken on the second part.

- Tom

Thank God for that.

The less we all see of mpeg2 on either of these new disc formats the better.

All this defending and apologising for Mpeg2 is just getting old.


Who is still defending or apologizing for Mpeg2 on this forum? I'll tell you that Tears of the Sun is a very high quality Mpeg2 disc that fits on a 25GB SL Blu-Ray disc, however it would have looked even better had the same people that worked on POTC DMC got a hold of it with their knowledge and experience today and it was encoded on BD50 with AVC... And that is why rdjam I feel your thread is misleading at best. The Blu-Ray exclusive studio's are way past the usage of Mpeg2 already on Blu-Ray discs as they are also using BD50's for the majority of films and at a 100% rate for the blockbusters. It is silly to keep on comparing old, non-optimized on the fly encodes against newer optimized VC-1 encodes. It is also silly to compare non-optimized AVC that is also on the fly against optimized VC-1 where encoders take their time and re-watch the encodes for inconsistencies.

If the point of your thread was to really lessen the use of Mpeg2 around the world, then this thread is useless and doesn't need to go any further because that is already happening.

If the point of this thread is to try and convince casual AVS’ers who don’t know any better that VC-1 is more efficient than Mpeg2 and AVC so it is the greatest thing since the invention of sliced bread. Then you fail at that since you are selectively choosing biased challenges where everyone on this forum is pointing out to you as being so. What is keeping you from choosing Flags of our Fathers as a one on one comparison? Or screenshots of POTC DMC vs King Kong with scenes that are in the jungle and have similar colors or ocean scenes..etc..It can be done…

As for facts, AVC has proven itself worthy of being called the codec with the best looking release at tier 0 in POTC DMC if you put any faith on the tier list whatsoever. King Kong which was the higest rated non-animated HD-DVD has dropped down to tier 1 because of artifacts in the film taking it out of tier 0 status. These are just two films however if you want to compare two recent titles that got the best treatment, then you have to pick two titles that got the best treatment with their respective codecs. If you fail to do this, then it simply proves to everyone that AVC with higher bitrates on a larger BD50 medium and proper encoding has a greater chance at surpassing a VC-1 encode on a DL HD-DVD30 under the same encoding conditions. My last sentence is what is happening in the real world today when it comes to blockbuster releases today..

This is a silly thread and I wonder what ties the orig poster has with the VC1 codec since he seems to champion it in every single thread but hey, whatever.

All I know is many discs done in VC1 look good, many in AVC look good and many in mpeg2 look good. I say enjoy the movies. If you have no financial ties to the codec being used then who cares? Clearly the codec being used does not automatically equate to top PQ (since we have seen mediocre jobs with the diff codecs) so why champion one over the other if you have no financial gain from it? I just do not understand the point of championing for a single codec all the time...

Oh and if you really want to compare VC1 then do it vs a bluray release done in AVC like Flags of Our fathers or Coming to America or whatever other releases.

Well, up to now, I have taken my material for this thread from what been posted in other threads by Xylon and Madshi. However, it is unfortunate that I have been accused of deliberately biasing the comparisons, when all I have done is used whatever became available.

I did not control what anyone else chose to capture.

However, due to popular demand expressed here, I will be doing a load of captures from various BR and HD DVD releases, so that we can have comparable bitrate AVCs, and higher bitrate Mpegs, compared to VC1 titles in this thread.

Hopefully folks will be pleased with these. I expect them to be ready within a week.

I appreciate everyone's patience in the meanwhile...

This is a silly thread and I wonder what ties the orig poster has with the VC1 codec since he seems to champion it in every single thread but hey, whatever. ...If you have no financial ties to the codec being used then who cares? ...so why champion one over the other if you have no financial gain from it? I just do not understand the point of championing for a single codec all the time...

Wow - so much doubt...

Just to repeat, for the record. I don't work for Microsoft or Toshiba, or HD DVD, etc etc etc. I am not even in the Consumer Electronics industry, but rather, in travel and hospitality. I have no financial interest in who "wins" or "loses" this "war".

I am simply a lover of the very best Hi Def versions of the movies I love, and the bottom line is that HD DVD has been producing the best PQ, and rather consistently, far more so than the other format. Sure, not every title has been the best, but many more so than BR. If the BR release is shown to be the best version, and it is a title I want, then I will go with that also. But so far for any title that appears in both formats, my choice has been HD DVD, ostly because of VC1.

Finally, as stated above, I will be doing a lot of my own screen captures from now on, so as to provide much more material for these comparisons, and so satisfy the popular demand.

rdjam....How will you be capturing the shots of the films btw? My suggestion is to take the movie you feel that looks the best on HD-DVD that uses VC-1 and get screenshots of it on the very best scenes you feel it offers. Then Do the exact same thing with the very best Blu-Ray has to offer on a similarly lighted scene and we can all compare side by side between the two movies if you want to do something different. Xylon is already taking care of side by sides of Flags of our Fathers and those movies that have been encoded with VC-1 and AVC for the same movie so another thread of the same is not necessary. However a direct comparison of the Best of the Best both formats have to offer even if they are different films might be more useful in your case. You have to leave bias out the window when doing these comparisons as this thread will get old very fast if all you post are items that benefit HD-DVD. Then this will become a pointless thread..

Thanks for the suggestions Supermans.

I hope that that is OK with everyone, as I can see the complaints in my head already: "Why aren't you comparing the same frames?"

I will do a combination of what you say, as well as comparing exact frames, I think, to suit both sides.

Where encodes are available from the same movie in more than one codec, I'll compare the same frames - (a) in scenes with little or no movement, to show the best of the images, and (b) in very difficult scenes to show the worst of each codec.

And to your suggestion, I will take the best and the worst of other encodes wherever I can find them, if it is not possible to get the same frames (ie not on both formats).

How's that sound?

By the way - reviews of the VC1 and AVC versions of "Coming To America" are up. And the VC1 version has beaten up the AVC version.

http://www.avsforum.com/avs-vb/showthread.php?t=855440

http://www.***************.com/fusionbb/showtopic.php?tid/137357/

The AVC/MPEG-4 encode on the BD disc, just doesn’t match up the the VC-1 presentation on the HD DVD. In comparison, it looks rather flat, especially when it comes to the color saturation and in the detail department. The overall look of this encode is much softer and lacks the depth that the competing format has presented with what I assume is the same master of the film. The only plus that the Blu-ray encode has over the HD DVD is that it doesn’t look quite as compressed. The film has a bit more grain on the HD DVD encode compared to the Blu-ray, and I am thinking that some type of filtering in the AVC encoding process may be the culprit of the softer looking presentation.

By the way - reviews of the VC1 and AVC versions of "Coming To America" are up. And the VC1 version has beaten up the AVC version.That's one professional review, on a 720p display. Not to mention it's a BD25 vs HD30.

Please, at least attempt to be not transparent and dishonest.

WELL, Well, well...

We now have some great shots from Flags of Our Fathers for comparison.

The above protesters should be VERY pleased with this set of shots, since The Bluray version is on AVC, and, GET THIS, the AVC version is using over 58% MORE BITS than the VC1 version.

It's a Bluray-lover's dream come true. Surely if ANYTHING will show VC1 up, then THIS will be it. The VC1 encode is "only" 20.80 gigs, yet the AVC encode is a WHOPPING 32.90 gigs.

AND YET! The two encodes are virtually identical. Now this means one of two things, in my view: (1) That AVC needs a lot more data to look the same, and may look worse at the same rate, or (2) Worse! That there is no additional benefits to a high data rate on the next gen codecs at all, and that the extra bandwidth of Bluray is NOT NEEDED when using the advanced codecs. Now which of the two options is it? More comparisons of other movies may help us decide...

Anyway, without further ado - here are some 200% blowups for comparison, again using Lanczos to maximise anydetail stored by each encoder.


BD AVC on left - HD DVD VC1 on Right:

http://www.CampaignHD.com/images/Flags_bd_1a_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_1a_2xL.png

http://www.CampaignHD.com/images/Flags_bd_1b_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_1b_2xL.png

http://www.CampaignHD.com/images/Flags_bd_1c_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_1c_2xL.png

Massively impressive performance by VC1 against AVC with 58% more horsepower given to it...

And here are some more!

BD AVC 32.9 Gigs on left - HD DVD VC1 20.8 Gigs on Right:

http://www.CampaignHD.com/images/Flags_bd_2a_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_2a_2xl.png

http://www.CampaignHD.com/images/Flags_bd_2b_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_2b_2xl.png

http://www.CampaignHD.com/images/Flags_bd_2c_2xL.png http://www.CampaignHD.com/images/Flags_hddvd_2c_2xl.png

A VERY impressive performance by VC1 against AVC with 58% more horsepower given to it...

I am more impressed by AVC and the extra bandwidth
http://img516.imageshack.us/img516/3422/flagscuffrt6.png

rdjam,
Again, you lack the necessary information to tell if the advantage is 58% or 0%. You conveniently missed the context that a linear behavior for pq may not be occuring between the 2 encoding sizes you cite. That's why you need more data, before you can conclude "58%".

I will agree with rdjam on one thing: for only 20gb the vc-1 encode doesn't look horrible. It just doesn't look as good the avc.

I'm sure if the vc-1 had used all of it's disc space for the actual movie then the gap would be really small or non-existent. Of course if the avc encode used all of the 50gb on the BRD...

rdjam,
Again, you lack the necessary information to tell if the advantage is 58% or 0%. You conveniently missed the context that a linear behavior for pq may not be occuring between the 2 encoding sizes you cite. That's why you need more data, before you can conclude "58%".
32.9 Gigs, divided by 20.8 gigs, Equals = 1.5817, equals 58.17% bigger for the AVC version.