...at least in my current experience. I am very excited to have gotten a PS3 to go along with my HD-A2, but I am having a hard time finding superlative titles to go with it. I have rented Ice Age, Ultraviolet, Talladega Nights, Fifth Element, and have just bought Pearl Harbor and Black Hawk Down (still in the mail! ugh!). I will rent Kingdom of Heaven next, but I honestly expected Pearl Harbor to dazzle me. the reviews have been pretty great.

Well, I still see MPEG block in this one. BAD in some places. I have yet to see this problem with any HD-DVD transfer. I am going to rent some VC-1 Blu-Rays next to see if it's the codec or the player...but I already know it's the codec. Well, dual layer BD-50 "should" be able to avoid this. Bit rate seems to hover around 25 most of the time, and hits the 30s during some scenes. Maybe the grain is what killed it...IDK. Just so you BR Fanboys don't flame me with suggestions about how it's my display, my eyes, or some HD-DVD love affair I took some pics at random parts of the film. These are shot close enough to show the detail of what I'm talking about. they are by NO means the worst cases of it in this film. I have yet to find a BR title that doesn't have this artifact. It's not as bad as HD cable, but it's enough that for these expensive players and movies it shouldn't be happening. The sound is unreal however, but why is BR still not matching the PQ of HD-DVDs, even with the extra space? i think it's time to stop defending MPEG2 and move on to the advanced codecs.

The pics are JPG, but I made sure to use a large enough file size to avoid adding additional artifacts (at least noticable ones). Ever see this on a VC-1 encode?

Wow. That last one looks horrible! I can safely say, on my 10' wide screen, I've yet to see a block with HD-DVD, for what that's worth.

No more codec threads, please. Codecs are not the issue.

Let's be honest Rob, anything that criticises anything to do with BD appears to be "not the issue" for you. The OP's topic is obviously of interest.

The pics are JPG, but I made sure to use a large enough file size to avoid adding additional artifacts (at least noticable ones). Ever see this on a VC-1 encode?
Yep, those are classic MPEG-2 block artifacts. MPEG-2 encodes everything in 8x8 blocks, and if individual blocks don't get enough bits to encode the high frequency data in them, they can wind up looking quite different from their neigbors. This becomes a bigger problem when you have two adjoining blocks with the same problem, since the errors won't match, and the underlying 8x8 block pattern pops out at you.

You're not going to see VC-1 look that bad in that way, even at insanely low data rates, since we have a "loop filter" that smooths out those harsh right-angle artifacts between 8x8 blocks. Of course, HD DVD uses higher VC-1 bitrates where the blocks don't get overly compressed to that degree anyway. But we degrade more smoothly than MPEG-2 under bandwidth stress.

Sony talks about the speed of their MPEG-2 encoder more than they talk about how good an MPEG-2 encoder it is. These are pretty dramatic problems for a title using BD-50.

Yep, those are classic MPEG-2 block artifacts. MPEG-2 encodes everything in 8x8 blocks, and if individual blocks don't get enough bits to encode the high frequency data in them, they can wind up looking quite different from their neigbors. This becomes a bigger problem when you have two adjoining blocks with the same problem, since the errors won't match, and the underlying 8x8 block pattern pops out at you.

You're not going to see VC-1 look that bad in that way, even at insanely low data rates, since we have a "loop filter" that smooths out those harsh right-angle artifacts between 8x8 blocks. Of course, HD DVD uses higher VC-1 bitrates where the blocks don't get overly compressed to that degree anyway. But we degrade more smoothly than MPEG-2 under bandwidth stress.

Sony talks about the speed of their MPEG-2 encoder more than they talk about how good an MPEG-2 encoder it is. These are pretty dramatic problems for a title using BD-50.
^^Ya, but what about the JELLY FISH!! :eek:

HAHA, sorry, that comment made by, who was it, some Sony exec still cracks me up, even if there is some artifact that could be described in that was that pops up at some stage of the VC-1 encoding process. I've never seen anything like that on any of the HD-DVDs I watch at home, while the reported blockiness of even the best looking titles of BD mare the final product. That is one thing that has allowed me to hold off on shelling out the expense of getting a BD player. I get all the blockiness I need with HD cable.

I've read this disc has issues with chroma noise as well. What really gets me is that the Blu-Ray folks believe that this is a Tier 1 title (http://www.avsforum.com/avs-vb/showthread.php?t=753726) with "No visible compression, sharp image with many examples of 3D." Blu-Ray must have very low standards if this is Tier 1.

hah, i don't think 15 year old worn out VHS tapes look that bad on my 60" sxrd screen. I guess blu-ray users should all sell their players and buy VHS players then, right?

I thought MPEG2 was limited to 16x16 blocks. ??? At any rate, keep in mind that the playing movie doesn't look like this. These are detail shots just to show the effect. They demonstrate a 10-15% image crop.

But what's the deal with this? I expect a BD50 disc to have enough space to avoid this. I read so many messages about people defending MPEG2 (even saying at 25+ mbps it is better than VC1), so why this?

What's a good BD VC1 encoded disc to try out?

I thought MPEG2 was limited to 16x16 blocks. ??? At any rate, keep in mind that the playing movie doesn't look like this. These are detail shots just to show the effect. They demonstrate a 10-15% image crop.

But what's the deal with this? I expect a BD50 disc to have enough space to avoid this. I read so many messages about people defending MPEG2 (even saying at 25+ mbps it is better than VC1), so why this?

What's a good BD VC1 encoded disc to try out?
On my 10' wide screen, I'm sure they'd be visible enough. :(

The only VC-1 encoded BDs are from Warner, who simply port over their HD-DVD encodes. So if there are no blocks in the HD-DVDs, which there aren't, there won't be any in their BD counterparts.

I have to say it's a little disheartening to know that BD exclusive studios feel that encodes with even a little bit of blockiness is acceptable, and that they apparently don't feel the need to go with VC-1, or at least AVC for all their titles.

What bitrate is this, and how much motion is there?

Yeah nice! Would love some time codes!:)

Yeah nice! Would love some time codes!:)
Heh. I was going to ask that, but decided it was moot for me, since I hate that movie anyway and wouldn't bother ever watching it again. :p

However, it would be interesting if someone could post the timecodes, so someone could verify the bitrates at those times.

No more codec threads, please. Codecs are not the issue.


Codec is certainly responsible for the issue I've illustrated above.

With time codes you would not have to watch the entire thing over again! That's why I asked!:D

Heh. I was going to ask that, but decided it was moot for me, since I hate that movie anyway and wouldn't bother ever watching it again. :p

However, it would be interesting if someone could post the timecodes, so someone could verify the bitrates at those times.

No need for timecodes. Any time there is fast motion you can pause the frame and find this. If you don't want to get too far into this movie though, pause on the wide shot of the plane landing in the wheatfield as it pans right at the very beginning. It's about 1:35 into the film. PS3 reports the bitrate as 24mbps here. Half the screen will be filled with blocks.

How about 2:53 as the plane is turning around. Bitrate is 26.4 and again squares throughout the grass and bushes.

Just about anywhere there is action you can pause and find this artifact. Bitrates go over 30 at some points.

They are indeed 8x8 squares BTW. They are discernable enough that I can count their pixel dimensions.

If mpeg2 really is that bad, then why are iRobot and U-571 on DVHS still considered some of the best looking HD of any format? Are these blocking artifacts present there as well?

If mpeg2 really is that bad, then why are iRobot and U-571 on DVHS still considered some of the best looking HD of any format? Are these blocking artifacts present there as well?
Maybe for the same reason that Pearl Harbor is considered a "Tier 1" Blu-Ray release.

If mpeg2 really is that bad, then why are iRobot and U-571 on DVHS still considered some of the best looking HD of any format? Are these blocking artifacts present there as well?

Just my speculation: better source and better mpeg2 encoder from JVC.
Need to ask dr1394...

No need for timecodes. Any time there is fast motion you can pause the frame and find this. If you don't want to get too far into this movie though, pause on the wide shot of the plane landing in the wheatfield as it pans right at the very beginning. It's about 1:35 into the film. PS3 reports the bitrate as 24mbps here. Half the screen will be filled with blocks.
How about 2:53 as the plane is turning around. Bitrate is 26.4 and again squares throughout the grass and bushes.
Just about anywhere there is action you can pause and find this artifact. Bitrates go over 30 at some points.
Can you see blocks in real time? Are you sure the blocks are not an artifact of the pause function (incomplete decoding)? There should be no blocking at these rates unless the encoding was sloppy.

Can you see blocks in real time? Are you sure the blocks are not an artifact of the pause function (incomplete decoding)? There should be no blocking at these rates unless the encoding was sloppy.

That's what I was thinking too... this can't be "real", nobody in their mind would ship a movie with such artefacts. You get artefacts such as this with JPEG if you turn the usual 0-100% quality parameter to something like 10-15. It's quite possible to be a problem with the pause function (stupid, stupid Sony - many users are likely to be comparing/examining paused frames, even though nobody watches paused movies), or something is wrong with your particular PS3.

Not all video decoders are the same, it looks like the PS3 software decoders aren't up to scratch.

Why does this one persons experience take precedent over everyone else? Oh I forgot. Because Blu-ray is evil and will eat our children.

This exact same thing has happened to HD DVD. I recall a thread analyzing Phantom Of The Opera practically on the pixel level looking for "jellyfish". Stop the partisan BS and get used to it. If there are no issues with PH, then there will be plenty of objective AVSer's who will offer their support. And if there are issues, then you take it from there.

Let's be honest Rob, anything that criticises anything to do with BD appears to be "not the issue" for you. The OP's topic is obviously of interest.

I have to agree with Rob, codec is not the issue here..

This is just another desperate trick from HD-DVD campers which really starts to get boring.

Sorry but I see no such thing on my end. I went thru frame-by-frame on both of those timecodes you mentioned. And I looked at a lot of other scenes too. I'm using PC-based playback with PowerDVD Ultra 7.2. If other people give their reports with different setups then we can safely say this is something down to your setup.
It is possible the PC player employs MPEG-2 noise filter, ala Samsung in the old firmware. Such a filter is designed to smear the block edges.

My main suspect would be the PS3 obviously. You mentioned that you don't see this when the movie is playing so I would say that the PS3 pause function is not up to snuff.
No Pause function explains what is posted here. It would explain jaggies but not obvious compression artifacts.

I can pause on any frame in PowerDVD and can do perfect frame-advance and the image never falls apart.
Assuming your PC display is 1920x1080 and not some other resolution. Most PC displays are 16:10 and if you run them full screen, you get filtering/interpolation which softens the video. So be sure to run 1:1 and report back. Also look in the player to see if there is an advanced setting for noise reduction and turn that off and report what you see.

Don't you think all the reviewers and regular users who have watched this movie would have seen such huge defects? Why does this one persons experience take precedent over everyone else? Oh I forgot. Because Blu-ray is evil and will eat our children. :rolleyes:
To be honest, this is a torture test. This was a movie that Disney picked as their test clip in DVD Forum codec shoot out for HD DVD. So if there is a movie that is going to show problems from their portfolio, this is probably it.

And yes, I agree it is pushing it to claim that every frame of a movie must be perfect. They never will be. The question is though, now that you have seen the artifact, whether you can spot it in normal viewing.

BTW, I don't agree that professional viewers have not seen this artifact. When MPEG-2 becomes mildly blocky, it can give a hard-edge look to the video with the blocks being readily visible (well, to folks who don’t do compression for a living). All of those sharp edges as they come and go rapidly in the 8x8 blocks, increase the high frequency component of the video, getting rid of the smooth "film look" we all want. You can perhaps see this in Peter Bracke's review: http://bluray.highdefdigest.com/pearlharbor.html

"Unfortunately, chroma noise can be a glaring irritant. It is especially evident on primary colors, and combined with the harsh quality I generally find with MPEG-2 transfers, the image often has an unappealing hardness."

Although he does go on to say that he doesn't see blocking artifacts during fast motion which tells me the artifacts are not as visible in full motion, but cause the harshness I talk about.

I can see this thread going on for 10-12 pages :(

If the blocks looked that bad then I don't think any amount of smearing could fix it. I don't see anything near that bad of an image. I don't see anything wrong with it at all in fact. No blocking, no smearing, no noise filtering that I can see.
Actually, deblocking filters are very effective in hiding such artifacts at the cost of reducing resolution. If you don't mind, please check the advanced settings in your player.

We can't rule out that the PS3 decoder is not up to par either as someone already mentioned.
I don't see any flaws in the decoder that can cause such an artifact. There can be decoder issues at all, but this is such an obvious artifact of bit started MPEG-2 that it is hard to put blame anywhere else in the decoder.

The only possibility that could point the finger at PS3 is if it sharpens the video on the way out. Doing so really highlights blocking artifacts. I don't have a PS3 yet unfortunately so don't know if there is a setting for sharpness in it. I think Stacey tested the PS3 and found it to not have any ringing problems so I am not sure there is a lot of blame to go there. BTW, trust me, I love to see something wrong in PS3 :). But I just don't think that is the burning bush here.

I've seen plenty of other defects from both formats for me to complain about but not blocking. And certainly nothing to this degree.
I see blocking artifacts frequently in BD titles. The early launch titles are good examples as are the demo loops in stores/shows. But that is neither here nor there wrt to the specific title here.

My display is a 4:3 CRT and it is set to 1920x1440.
Ah, that explains it :). I love CRTs and they provide an excellent picture. Alas, they lack the resolution and sharpness of a fixed pixel display. Indeed, if I want to find blocking artifacts, I always play things on a 1080p LCD with 1:1 pixel mapping, not a CRT. The CRT softens the edges automatically because there is no such thing as a pulse going from zero to max in an analog device whereas a digital display happily goes from black to full white in one pixel.

Alas, I can't find my copy of Pearl Harbor either. When I do, I will give it a try on the Ruby and report back if there is interest.

I thought MPEG2 was limited to 16x16 blocks. ??? At any rate, keep in mind that the playing movie doesn't look like this. These are detail shots just to show the effect. They demonstrate a 10-15% image crop.
It uses 8x8 blocks. But color is only sampled once every 2x2 block, so one 8x8 block of chroma covers 16x16 pixels. That's called a macroblock.

But what's the deal with this? I expect a BD50 disc to have enough space to avoid this. I read so many messages about people defending MPEG2 (even saying at 25+ mbps it is better than VC1), so why this?
Yeah, those are pretty bad artifacts. I'm not sure how much of it is the content's unsuitbiility to MPEG-2, and how much on Sony's emphasis on fast encoding over video quality.

What's a good BD VC1 encoded disc to try out?
Check out any of the recent Warner titles - they use VC-1.

That's what I was thinking too... this can't be "real", nobody in their mind would ship a movie with such artefacts.

While not a High Definition release, the recent Sony DVD for "Transformers: The Movie - 20th Anniversary Special Edition" is plagued by horrible blocking artifacts of the type we're discussing here during any fast motion, some of the worst I've ever seen on DVD. I'd post some screen caps, but the damned thing won't load in my PC.

So yes, discs do still get released with such terrible encoding.

It looks sort of like a few of those lower bitrate downloadable WMV HD trailers that were posted a while back. Miami Vice:

http://www.eugbanana.com/files/Pix/MiamiVice1.jpg

I wonder how much of it is due to the decoder though, as this was captured on my Mac with QuickTime and the Flip4Mac plug-in. Flip4Mac is distributed by Microsoft though.


if there is interest.
Of course. :)

So yes, discs do still get released with such terrible encoding.Well you've reviewed about 30 Blu-Ray movies for DVD Talk. Have you seen the severe artifacts that are the subject of this thread in those movies or any other high def Blu-Ray release?

It uses 8x8 blocks. But color is only sampled once every 2x2 block, so one 8x8 block of chroma covers 16x16 pixels. That's called a macroblock.


Yeah, those are pretty bad artifacts. I'm not sure how much of it is the content's unsuitbiility to MPEG-2, and how much on Sony's emphasis on fast encoding over video quality.


Check out any of the recent Warner titles - they use VC-1.
The 8x8 blocks of chroma that covers a 16x16 macroblock is known as the 4:2:0 YUV color space. All shipping consumer grade video SD and HD uses this color space. Even HD DVD and BD use YUV 4:2:0

There is 8x8 DCT used by MPEG2 whereas AVC High Profile allows 8x8 or 4x4 transforms that are DCT like.

In many cases, if the master is D5, the MPEG2 DCT should be a closer match to the D5 tape as I believe that 200Mbps format is DCT based as well. AVC 8x8 transforms under stress streams will likely quantize the 8x8 blocks as well, while 4x4 transforms for HD should be used judiciously. (Conjecture, I have not looked at D5).

I have no idea how VC-1 does this, I expect it enjoys many of the same advantages of AVC except for the intra modes may be different and perhaps it does only 4x4 transforms?

The jpeg capture for #3 is the worst form of macroblocking I've seen only in badly quantized pictures, and if true would mean there is at least least one frame that is extremely heavily quantized... So heavily quantized that noone should see this in anything shipping today. If it does exist, Disney or sony should look at that frame in the D5 and see if someone screwed up either the encode or the mastering.

#1 & #2 also looks a bit like macroblocking, but it might also be a display artifact. Timecodes and location on frame would be good.

On my setup, I watched the segments near the beginning and the attack scenes on both the samsung and the PS3. I did notice the PS3 showed more grain effect than the samsung, and I do see the wheat field in the beginning has areas that suspiciously look smoothed out, and has some macroblocking, but not anywhere near the level of what is shown in the jpegs. The samsung appears to filter things out so I concentrated on the PS3 playback to get more detail out of the pictures.

Using PS3, I tried to playback parts of the attack to try to find where people are running around, and I used slow motion on those scenes. I am puzzled as what I see are motion blur effects in many frames (which is not great but expected in fast movement), but none of the macroblocking that is discussed above.

If there is a timestamp for these frames, especially bigger chunks of the frames to identify where they came from, that would help chase this down.

When I watched the attack scene, I thought the bombings of the ships, all the way through to dog fights, were actually quite well done, even if there was significant grain in there. Nothing like what the jpegs are showing.

Are you implying that D-Theater is not as good as everyone says it is? Sounds like you've never even seen a D-Theater title. And as far as I can tell you have never seen many Blu-ray titles either. And yet you feel that you can comment on the quality of both.
You are correct, I have not seen a D-Theater title and should have been clearer, I apologize. I have watched Blu-Ray however, which is why I believe though Blu-Ray is good, HD DVD is better. What I meant was Pearl Harbor is considered Tier 1 because perhaps most obviously have not noticed the PQ issues being discussed, whether it is smoothed somewhat due to the display and/or player being used. D-Theater, being around prior to HD DVD and Blu-Ray, had no other standard to compare against, so it would certainly been considered as having the best PQ at the time. My question would be -- does D-Theater have any sort of smoothing built into the player that would hide something like this? Has anyone looked at the PQ of D-Theater as in depth as the OP has done with this Blu-Ray title?

Sorry Amir, I normally agree with you, but this could easily be a decoder issue. I've seen plenty of cases (especially with software decoders on PCs) where a compression artefact shows up with one decoder and not with another.

if it is a decoder issue, then PS3 is not up to snuff for optimal BD playback.
if it is a codec issue, then MPEG2 should not be used for any HD material.

The 8x8 blocks of chroma that covers a 16x16 macroblock is known as the 4:2:0 YUV color space. All shipping consumer grade video SD and HD uses this color space. Even HD DVD and BD use YUV 4:2:0
Correct.

There is 8x8 DCT used by MPEG2 whereas AVC High Profile allows 8x8 or 4x4 transforms that are DCT like.
Correct. VC-1 also has variable block sizes (including 4x8 and 8x4).

In many cases, if the master is D5, the MPEG2 DCT should be a closer match to the D5 tape as I believe that 200Mbps format is DCT based as well. AVC 8x8 transforms under stress streams will likely quantize the 8x8 blocks as well, while 4x4 transforms for HD should be used judiciously. (Conjecture, I have not looked at D5).
Note that D5 uses 4:2:2 color, so there isn't a simple macroblock correspondance from it to any of the delivery codecs.

I have no idea how VC-1 does this, I expect it enjoys many of the same advantages of AVC except for the intra modes may be different and perhaps it does only 4x4 transforms?
Nope, we had variable block size before H.264 (H.264 added those with High Profile, after VC-1 beat it in initial DVD Forum testing). We also have a DCT-like transform, loop filter, etcetera.

Using PS3, I tried to playback parts of the attack to try to find where people are running around, and I used slow motion on those scenes. I am puzzled as what I see are motion blur effects in many frames (which is not great but expected in fast movement), but none of the macroblocking that is discussed above.
Curious. Perhaps PS3 is doing some post-processing as well?

I am the OP. Honestly, my intent is not to show HD-DVD as the better format in a "my dad can beat up your dad" fashion. I have both formats, and my love is for HD, not either side. Still, I only see these artifacts with the Blu-Ray discs I have watched. I DO see them in motion, although they are nowhere near as disgusting in normal motion. They are VERY evident in slow-mo, and obviously more evident on the 52" Sharp than the 42" (blocks are bigger).

Also, keep in mind that the Sharp is displaying 1080p 1 for 1 without overscan or any form of scaling. I understand a good majority of sets do not behave this way. Also, LCD shows sharp edges very cleanly because of the elimination of the "focus" issue that CRTs and DLP sets (projectors too) have so I expect them to look their worst on an LCD.

It could be the PS3 at fault, but these are the same artifacts I see on the Samsung and Sony in-store demos, so if the problem is in the PS3 it's in he other players too. Also, these artifacts are quite recognizable and common to MPEG2 encodes over broadcast, and on SD-DVD so blaming them on anything but the codec seems to be a stretch to me. I imagine the excessive grain is wasting bit-depth here, probably making the issue worse, but that doesn't explain why I see it in Ice Age, and The Transporter (which I have playing as I type this).

Again, I am not pronouncing the Doom of Blu-Ray here. I am just pointing out a glaring and noticable (to my discriminating eye) shortcoming of MPEG2, which I believe Sony is pushing in order to justify the superiority of Blu-Ray's extra capacity. Why not use the extra space for insane bit-dpeth VC1 encodes, extras, and uncompressed audio...or concentrate on BD25 discs and make them more affordable. Pushing BD50 and releasing titles with artifacts like this seems like a backwards move to me.

For those of you wanting time codes, honestly you don't need them. I can find offending frames every minute or so in just about every title I've seen.

FWIW, I've seen those artifacts in broadcast MPEG2 (or else something very similar).

I don't have enough experience with Blu-ray to assess this for Blu-ray MPEG2 titles though.

That statement basically discredits your entire experience as far as I am concerned. And now I am really starting to think those pictures are not what they claim to be.

What is it you think they are? I played the Blu-Ray disc on the PS3 at 1080p 1/1 on a Sharp LCD, paused the movie, used a Canon Digital SLR to snap a picture of the screen, reduced that 6mp file to 1/4 it's pixel size, and posted it here.

Knowing that each MPEG square is 8x8 pixels in size should give you a pretty good idea of how close of a crop the shots are. If I get bored tonight I'll do some more. I'll make sure to take note of the time, although I don't know of a way to provide timecode data from the PS3's info display.

I'm guessing based on your response hat you are a Blu-Ray fan. Again, I have nothing against the format at all, and plan on having quite a few titles. I'd just like to see them use the newer codecs to do it. Regardless of what you are sensitive to, I see the squares. My screenshots prove that they are in fact there as far as I'm concerned. You can theorize my motives all you want, but anyone who follows the steps above should be able to reproduce those shots without any trickery.

I'm not a fan of anything

You don't have a preference of one format over another? Your sig tells a differing story...

Frankly, the problem isn't with a format, but with either a codec or a player. I guess I have a CRT, but I've been macroblocking on HDTV, why couldn't I see it with MPEG-2 HD from a high bitrate source? It may not look as clearly defined as with an LCD, but it should still be there, right?

Please. i don't want to turn this into a contest. My intent was really to voice my vote for wider acceptance of the newer codecs on Blu-Ray, regardless of the spave they take up on the disc. I did some more shots just now, pretty much at random points. The last few are of The Transporter, just to show that it isn't limited to PH. I tried to show the time the shots are taken, as well as the bitrates. you can find the shots here...

http://www.pbase.com/sbushman/bluray_mpeg2_analysis&page=all

For the lazy, here are a few....

http://www.pbase.com/sbushman/image/72711024.jpg
http://www.pbase.com/sbushman/image/72711025.jpg

http://www.pbase.com/sbushman/image/72711020.jpg
http://www.pbase.com/sbushman/image/72711021.jpg

http://www.pbase.com/sbushman/image/72711026.jpg
http://www.pbase.com/sbushman/image/72711027.jpg

These aren't during high action either, where the problem gets worse, but is possibly less noticable because my eye is following the action. On solid patches such as sky area it is most distracting to me.

Follow the link above for the others too.

I checked on my PS3 and its just not there....and I am sure if it were...it would not have taken this long for somone to point it out.

One thing I do know that MPEG-2 has over VC-1.. is 8 out of the top 10 titles on the BD P/Q tier. ;)

I checked on my PS3 and its just not there....and I am sure if it were...it would not have taken this long for somone to point it out.

What's your display environment? 1:1 1080p displays are only just now becoming widespread.

Those pictures show totally textbook MPEG-2 artifacts. I can't think of anywhere else something like that could be coming from.

Someone without an agenda and a similar display needs to confirm these findings. I intend to get a 1080P display this year and if I have to buy a blu-ray player then I don't want it to look like this. I am picky at these prices. :D I want to see pictures though. I have to thank the OP for at least showing his proof instead of announcing something without backing it up. Your choice to believe him or not.

Someone without an agenda and a similar display needs to confirm these findings.
And better get skills like he has to take decent pictures with his DSLR.... :)

One thing I do know that MPEG-2 has over VC-1.. is 8 out of the top 10 titles on the BD P/Q tier.
Then I would be extremely interested if the OP would take pics of these 8 titles and show us what they all look like (or anyone with the same display and camera).

Ohhh.

It's a Sharp TV.

Anyone's who has seen them against similar models from Mitsubishi, Sony, or Samsung KNOW that this is no TV to demo the best of the best on.

Never mind the fact that pixel response time, lack of proper black level...what are the settings on the TV? Is NR on? Sharpness settings, etc?

How about some taken on ANOTHER TV?

And better get skills like he has to take decent pictures with his DSLR.... :)

My design of HD capture unit can acquire 24-bit RGB losslessly from any analogue or digital source, including HDCP-encrypted HDMI.

King Kong, captured from the 360 add-on via VGA (so may not be 100% precise in terms of colour calibration):

http://img155.imageshack.us/img155/690/kong1wx3.th.jpg (http://img155.imageshack.us/my.php?image=kong1wx3.jpg) http://img294.imageshack.us/img294/2837/kong2lf0.th.jpg (http://img294.imageshack.us/my.php?image=kong2lf0.jpg)

Ridge Racer 7, captured from PS3 via HDMI (100% exact to the machine's framebuffer, JPEG compression apart):

http://img294.imageshack.us/img294/9502/r1dq3.th.jpg (http://img294.imageshack.us/my.php?image=r1dq3.jpg) http://img294.imageshack.us/img294/4111/r2pj3.th.jpg (http://img294.imageshack.us/my.php?image=r2pj3.jpg)

I'll have a retail PS3 unit early next month and will be happy to capture any image from any BD that is causing controversy, though the thought of spending any kind of money on Pearl Harbor rankles somewhat.

The only way to get to the bottom of this is to take out the display from the chain of devices rendering the image and get at the image from a lossless source, or at least a direct one. Then we can have a 20-page thread on the merits (or otherwise) of the player that produced it in the first instance :D

And those new screenshots actually indicate the display being the culprit in my mind.

On the one hand, I would disagree as those artefacts look uncannily like the result of MPEG2 compression.

On the other hand, I'm more inclined to agree as the presence of macroblocking even on scenes with little motion indicates either a really poor encoder or else a terribly low amount of bandwidth being afforded the picture. I mean, come on, even something like Canopus Procoder or the TMPEG offering could do better than that.

Ohhh.

It's a Sharp TV.

Anyone's who has seen them against similar models from Mitsubishi, Sony, or Samsung KNOW that this is no TV to demo the best of the best on.

Never mind the fact that pixel response time, lack of proper black level...what are the settings on the TV? Is NR on? Sharpness settings, etc?

How about some taken on ANOTHER TV?

Sharp Aquos are among the best if not the best consumer LCDs, just check any reveiw site like CNET - I have a Sony 46in 1080p Bravia LCD. I would've opted for the Sharp due to better colours and contrast but wanted something bigger than 42in.

King Kong, captured from the 360 add-on via VGA (so may not be 100% precise in terms of colour calibration):

http://img155.imageshack.us/img155/690/kong1wx3.th.jpg (http://img155.imageshack.us/my.php?image=kong1wx3.jpg) http://img294.imageshack.us/img294/2837/kong2lf0.th.jpg (http://img294.imageshack.us/my.php?image=kong2lf0.jpg)

Finally some real screenshots of a HD-DVD. It looks even sharper when black level corrrected. Thank you.

Well you've reviewed about 30 Blu-Ray movies for DVD Talk. Have you seen the severe artifacts that are the subject of this thread in those movies or any other high def Blu-Ray release?

The pulled release of RoboCop had severe pixelation artifacts, but it didn't quite look like what's pictured here (which I've seen many times on HD cable, a recent example being The Island on HBO). I can't think of any Blu-ray releases that have looked like this.

I have not seen the Pearl Harbor or Transporter Blu-rays.

I've seen no blu-ray look like this either thus far. Screen is 15' wide so would have expected macro blocking this severe to be obvious...

Mark

Ohhh.

It's a Sharp TV.

Anyone's who has seen them against similar models from Mitsubishi, Sony, or Samsung KNOW that this is no TV to demo the best of the best on.

Never mind the fact that pixel response time, lack of proper black level...what are the settings on the TV? Is NR on? Sharpness settings, etc?

How about some taken on ANOTHER TV?
Sharp LCD TVs are generally very well respected.


And those new screenshots actually indicate the display being the culprit in my mind.
Except that they look like MPEG2 artifacts.

With all the advanced videophiles on this forum (I'm not one), I'm surprised that at least one doesn't take a picture to either prove or disprove what the OP has shown.

well, to be fair, I also always see block noise when there is quick fade in/out on BOTH HD DVD and Blu-ray Disc. But I always thought it's my TV.

Mark_H, do you still use the -1 sharpness settting on the Lumagen?

JeffY - no, not now I'm all-digital - I use 0. At 1 or -1 I can see image degradation in test patterns.

Mark

Well we've got at least 4 or 5 people now who don't see anything like this. And one of them also has a PS3. Another is a reviewer. And another has a 15' screen. People can draw their own conclusions from this information now.

And I'll remind you that the OP wasn't just talking about Pearl Harbor. He claims to see it on ALL Blu-ray titles. It could be that his Sharp LCD is such a high-end display that it allows him to see artifacts that the rest of us can't.
I suspect the best way to test this is with a 1080p LCD TV with 1:1 pixel mapping over HDMI or DVI. I will note that the original poster has such a TV.

Personally, I'd be surprised if this were a TV issue. My guess it's either an encoding issue (with his MPEG2 discs), or a decoding issue.

Sharp LCD TVs are generally very well respected.



.

Except that Sharp overprocess all images and the OP is not legitimate blaiming MPEG2 for the faults.

Change display and come back later, it's not BD related as i posted earlier.

It's the got damn LCD, Sharp in this case.

Actually, codecs and codec support are the entire issue.

As to why do D-VHS tapes look better.

D-VHS = Constant Bit-Rate (high)
BD-Disc = Variable Bit Rate (scene dependant)

If you don't think that the codec choice is the ENTIRE ISSUE - then check out this (admittedly old) independant report on the various codecs. [BTW: VC-1 has evolved significantly since this was done] - http://video.ldc.lu.se/pict/WM9V-MP4AVC-MP2V_comparison-Goldman.pdf

Except that Sharp overprocess all images and the OP is not legitimate blaiming MPEG2 for the faults.

Change display and come back later, it's not BD related as i posted earlier.

It's the got damn LCD, Sharp in this case.
He claims he doesn't see this on HD DVD titles with VC-1 on his Toshiba HD-A2. If it were the TV, he would see the same problem.

He claims he doesn't see this on HD DVD titles with VC-1 on his Toshiba HD-A2. If it were the TV, he would see the same problem.

Not taking sides here, but to the OP, how about a pic of a HD DVD that is VC-1 encoded displayed on the same TV?

Checked the PH scene out on my PS3 and while there is very noticeable grain, I could not see the artifacts the OP sees and shows in his screenshots. Here is a quick screenshot taken from my Benq 8720 (118", calibrated).

http://64.253.117.170/pics/IMG_1942.jpg
http://64.253.117.170/pics/IMG_1943.jpg

Once again, I think this is a display issue (remember the infamous KK image quality fight of Nov 06!). Some displays will show it, others won't. If yours does, then you see it. If it doesn't, then great, count your blessings.

OMD, PH has a noticeable grain to it, but doesn't display the artifacts the OP has seen or posted. Lucky me!

-Alex-

Except that Sharp overprocess all images and the OP is not legitimate blaiming MPEG2 for the faults.

Change display and come back later, it's not BD related as i posted earlier.

It's the got damn LCD, Sharp in this case.

You think it's the Set? Hmmm. My set must be Pro HD-DVD only then, because it reserves this behavior for the Blu-Ray, as I've said before.

Before I got the Sharp I had a Samsung LN4095 (which didn't meet my expectations), returned that for a Sony Bravia (v2500) which also didn't make the cut. If you think the Sharp (with NR off of course, Sharpness at 0) is showing artifacts that the Samsung and Sony won't show you must not have as discriminating an eye as I. Both the Samsung and Sony show pretty obvious Black Crush on all sources, and the Samsung can't even display 1/1 1080p. All sources are overscanned. Both show very noticable color casting in the shadow areas, and while there is NR that must be turned off on the Sharp to get a true display, there are about a dozen such features on the Sony and half as many on the Samsung. Black level is another story, as I prefer TRUE black levels rather than contrast enhanced artificial black level, and in this respect the Sharp beats the others hands down (and the 52" set beats any LCD I've ever seen).

As far as motion artifacts, they are eliminated by turning off NR on most of the sets...but in a true response time comparison the Sharps are 8ms/4ms (42/52 respectively). If you think MPEG2 artifacts are being cause by a slow response time you have missed the point completely.

I see artifacts in my Blu-Ray discs, took pictures of the artifacts to prove they are there, and posted them for all to see here. I am certainly open to suggestions, but as I've said before VC-1 doesn't suffer from this...on the same display. I can verify that my display is providing 1/1 output as well.

I am still open to reasonable suggestions, but how about somebody who "doesn't" have the blocks doing some comparable screenshots (not captures) of your PS3 or other player paused on similar frames. Maybe we can determine if it has to do with my sensitivity level vs. yours.

I am an audio engineer, and for years I couldn't hear the difference between a Mackie and API/Neve audio mic preamp. Once my ears finally awoke though listening to the Mackie almost made me think I had a bad mic cable when compared to the others. My point is this. Sometimes you just don't notice something until you do. After that it's glaring. I noticed the MPEG2 artifacts on the first Samsung Blu-Ray/1080p display I ever saw. I chalked it up to the demo...then blamed the Samsung...then the display. Seeing them on the Sony demo made me question things more. Seeing them at home brought me to this.

I have seen people walk up to an LCD set on "display" nuclear mode, watch a few seconds of stretched SD-DVD and go on for 20 minutes about how unbelievable the picture is. I doubt any of you would do that, but my point is that people see things differently. I am pretty sure the pictures I took with my camera are of MPEG2 blocks...& not phantom MPEG2 blocks either.

My point is this. Sometimes you just don't notice something until you do. After that it's glaring.

This is very true. I have spent the past six years staring at hideous MPEG2 encoding on standard definition DVD and I *know* what I'm looking for and I simply do not see these issues discussed here with Blu-Ray. This is on a system which is 1:1 mapped from source to 1080p display (unless I engage the anamorphic lens in which case there is some scaling).

That is not to say that MPEG in HD is perfect. I currently much prefer the images I see when VC-1 is used, but I suspect that is because MPEG is being bit-starved in many of the current releases so while the transfers often look clean (of mpeg artefacts compared to DVD) they also seem to lack the detail we see in VC-1 and have an often over processed and artificial look to them. VC-1 looks and feels very filmic in comparison.

Mark

I have put Aeon Flux at #1 on my list from BB online. I will try to do a direct comparison between the HD-DVD and Blu-Ray in a direct comparison on the same set, under the same conditions. That will at least eliminate the display as a factor I think.

-Bairda - THanks for doing some screenshots to compare to. Although I can still see a very small amount of blocking in some areas on the face closeup, most is indeed not there (top right forehead, bottom left cheek, etc.). Of course, if not mistaken I think your display is a native 1280x720 dlp projector so downsampling, and focus error is definitely working to help eliminate artifacts.

Also, I think people don't realize that these artifacts change in how they look when they are in motion. The most recognizable artifact of these blocks is large peices of straight edge "digital" grain, that in no way resembles film grain. It is caused by seeing the "errors" in the blocks fluctuating around. I still think most who claim they don't see it either don't have native 1/1 1080p displays, or don't know exactly how to see it. IMO if you don't recognize it you should try to keep it that way. I promise, once you see it you will be ruined forever.

I'd like to see more people do what Bairda has done, especially those with 1/1 1080p displays. Pictures do a lot to back up what people say they are seeing.

If the PC software allowed me to take screenshots :( I would have flooded this forum with about a gazillion of them by now from both HDDVD and BD.

Doesn't FRAPS or any of the other DirectShow-friendly capture tools work?

Checked the PH scene out on my PS3 and while there is very noticeable grain, I could not see the artifacts the OP sees and shows in his screenshots. Here is a quick screenshot taken from my Benq 8720 (118", calibrated).

http://64.253.117.170/pics/IMG_1942.jpg
http://64.253.117.170/pics/IMG_1943.jpg

Once again, I think this is a display issue (remember the infamous KK image quality fight of Nov 06!). Some displays will show it, others won't. If yours does, then you see it. If it doesn't, then great, count your blessings.

OMD, PH has a noticeable grain to it, but doesn't display the artifacts the OP has seen or posted. Lucky me!

-Alex-

I can see a hint of the compression artefacts but with that level of downscaling it's really hard to see.

Checked the PH scene out on my PS3 and while there is very noticeable grain, I could not see the artifacts the OP sees and shows in his screenshots. Here is a quick screenshot taken from my Benq 8720 (118", calibrated).
Your image looks A LOT smoother and less artifacted. Still I can make out some block lines. But they're very very faint and almost not noticable. See attachment.

It looks to me as if the decoding algorithm in the original poster's setup is running foul. Or there's a very bad sharpening algorithm running in the display. Or maybe a weird combination of both.

Category 5, does your display have settings about sharpening and/or about MPEG block reduction? Can you turn all those stuff off, if you can find it in the display's OSD somewhere?

Here is a quick screenshot taken from my Benq 8720 (118", calibrated). Comparing a Benq PE8720 Wxga (1280 X 720) DLP to a Sharp 1920x1080 LCD? Does the downscaling hide artifacts?

I've tried FRAPS, Snagit, Camtasia, Hypercam and some others I can't even remember. The fact that PowerDVD uses overlay video pretty much eliminates the possibility of taking screenshots in the first place. The only screenshots I've been able to take are with WinDVD at 540p :p because it doesn't use overlay. When I tried the 1080p hack/update for WinDVD it uses overlay so once again I was out of luck. I've tried everything I could think of and its a no go. And believe me I really would like to take some 1080p screenshots. If anyone knows the secret I sure would love to know...
Have you tried setting hardware acceleration to 0%?

The pulled release of RoboCop had severe pixelation artifacts, but it didn't quite look like what's pictured here (which I've seen many times on HD cable, a recent example being The Island on HBO). I can't think of any Blu-ray releases that have looked like this.

I have not seen the Pearl Harbor or Transporter Blu-rays.

The Island on HBO-HD was one of the worst examples of MPEG artifacting I have ever seen! Truly amazing how bad it was. The scenes that had a lot of movement were virtually unwatchable! A true disgrace! :mad:

Yep, the 8720 is a 720p display. So there will be downsampling in any image I post. I am not sure how much this will affect the image. Based on the 1080p/720p comparisons done in this forum it is probably less then what most would think (refer to this thread for more info: http://www.avsforum.com/avs-vb/showthread.php?t=767929).

In addition, the image scaling on the posted JPG is adding some artifacts. Not many but some.

Not sure what my "focus error" would be. My 8720 is pretty razor in my eyes with individual mirrors visible from 6-8 ft. Optics on the Benq 8700 series are considered some of the best (see this review: http://www.projectorreviews.com/Manufacturers/benq/PE8720_final/index.asp).

Once again if your display and setup doesn't show it, then Great!

However, this discussion now has me hunting for artifacts. Guess I am doomed to finding them now!

-Alex-

Comparing a Benq PE8720 Wxga (1280 X 720) DLP to a Sharp 1920x1080 LCD? Does the downscaling hide artifacts?
It does. Any kind of filtering (which a downscaling is) smears the sharp block edges making them harder to see.

Amir were you able to test Pearl Harbor yet?
No I did not. Had to pack and get ready for CES. As a matter of fact, I am flying out in a few minutes so I don't know when I get to doing this. If the discussion still not settled by the time I get back end of this week, I will for sure give it a try.

Sorry about that.

My setup is a ps3 to a rx-v2700 to a Benq PE8700 all HDMI at 1080i. The image attached is taken from The Transporter. It is similar to one of the photos the OP presented. The photo was taken using a Canon 1ds2 on a tripod, mirror lockup enabled, f8, ISO 200, 2 second exposure, 400mm lens from about 15 ft, 10 second self timer.

I would like some more details about how the photo was taken, since the exposure time was long (1/25 sec). I'm not sure if the digital rebel the OP used has a mirror lockup feature or not, but my understanding is that at shutter speeds from 1/60 sec to 1 second mirror shake is important and lockup should be enabled. In any event I'd like to know if the shot was taken from a tripod using a self timer or cable release, and if MLU was enabled. We should at least eliminate the camera as a possible source of the streaking.

By the way, I profess no expertise in video, but I am a good landscape photographer, see www.thewave.info.

yup, I see that block noise too especially on top of the letter "o".

just a note, 2 sec exposure is too long and will diminished the block noise. For video, ideally 1/60 sec is recommended. Tripod and cable shutter release is a must, MLU is optional. f/8 is definitely going to smear the block edges too. Since the object is flat, f/2.8 or wider is recommended.

Another opinion from someone with a Ruby:

http://www.hometheaterspot.com/htsthreads/dvd-review.php?sequence=1796
Indeed, if what is reported in this thread is true, that would put Chad in a tough spot with his new found love for MPEG-2 :).

Per David's suggestion, I reshot (see attached) at 1/25 second f2.8 I didn't see any significant difference between this shot at the one exposed for two seconds in terms of sharpness. Neither shows the streaking shown in the OPs photos.

Thank you Bill0711, your second picture (to my eyes) shows more blocks. The one big block over the letter "o" now looks like 4 mini blocks, I can now also see blockiness on the right bottom corner from the letter "t". Also more block can be seen on the bottom right corner at the edge of the letter "t". The same goes on the bottom left of the letter "l".

My setup is a ps3 to a rx-v2700 to a Benq PE8700 all HDMI at 1080i. The image attached is taken from The Transporter. It is similar to one of the photos the OP presented. The photo was taken using a Canon 1ds2 on a tripod, mirror lockup enabled, f8, ISO 200, 2 second exposure, 400mm lens from about 15 ft, 10 second self timer.

I would like some more details about how the photo was taken, since the exposure time was long (1/25 sec). I'm not sure if the digital rebel the OP used has a mirror lockup feature or not, but my understanding is that at shutter speeds from 1/60 sec to 1 second mirror shake is important and lockup should be enabled. In any event I'd like to know if the shot was taken from a tripod using a self timer or cable release, and if MLU was enabled. We should at least eliminate the camera as a possible source of the streaking.

By the way, I profess no expertise in video, but I am a good landscape photographer, see www.thewave.info.

Unless you have some special Mod I do not know of the PE8700 is neither 1080 or HDMI... I have one and it is 720 and DVI, and I love it BTW :)

Another opinion from someone with a Ruby:

http://www.hometheaterspot.com/htsthreads/dvd-review.php?sequence=1796

Says it all really. :D

Chad couldn't Spot an Mpeg2 artefact if it hit him on the head.

Yes you are correct, I use a HDMI to DVI cable to feed it and it is a 720P projector but I feed it 1080i.

Yes you are correct, I use a HDMI to DVI cable to feed it and it is a 720P projector but I feed it 1080i.

The 720p output will mask mpeg2 compression artefacts.

the thing is, even with 1080i converted down to 720p, one can still see the MPEG2 artifacts quite clearly.

Says it all really. :D Exactly. As I said, codecs are not the issue.

My setup is a ps3 to a rx-v2700 to a Benq PE8700 all HDMI at 1080i. The image attached is taken from The Transporter. It is similar to one of the photos the OP presented. The photo was taken using a Canon 1ds2 on a tripod, mirror lockup enabled, f8, ISO 200, 2 second exposure, 400mm lens from about 15 ft, 10 second self timer.

I would like some more details about how the photo was taken, since the exposure time was long (1/25 sec). I'm not sure if the digital rebel the OP used has a mirror lockup feature or not, but my understanding is that at shutter speeds from 1/60 sec to 1 second mirror shake is important and lockup should be enabled. In any event I'd like to know if the shot was taken from a tripod using a self timer or cable release, and if MLU was enabled. We should at least eliminate the camera as a possible source of the streaking.

By the way, I profess no expertise in video, but I am a good landscape photographer, see www.thewave.info.
Yep. Block artifacts are definitely there, although it's much less obvious due to the 1280x720 downscaling (and the fact that it's a photograph of a projected image).

I note that you have a PS3 as well. I wonder if this is strictly a PS3 decoder issue or if it's an MPEG2 encoding issue.

Exactly. As I said, codecs are not the issue.

If codec is not an issue, why do I still see compression artefacts?

Actually, codecs and codec support are the entire issue.

As to why do D-VHS tapes look better.

D-VHS = Constant Bit-Rate (high)
BD-Disc = Variable Bit Rate (scene dependant)

Also, I doubt D-VHS tapes are getting this kind of scrutiny much. D-VHS was great for the displays of a few years ago, but I don't think people are watching them much with 1:1 on 1080p displays. They'll definitely see artifacts with D-VHS there as well.

I note that you have a PS3 as well. I wonder if this is strictly a PS3 decoder issue or if it's an MPEG2 encoding issue.
It's unlikely to be a decoder issue. For all codecs, the baseline codec behavior is pretty clearly defined. There's extra things that a decoder can do to hide errors, but it's have to be a serious bug to actually introduce them.

And the images are of straight-up MPEG-2 artifacts.

Benwaggoner, I hear what you are saying, but if one player has an excellent decoder that hides errors (without loosing detail) and another player has a more basic decoder that doesn't, I still blame the decoder.

We had this with early software decoders in PCs, they didn't create compression artefacts, they were just crap at dealing with them.

Benwaggoner, I hear what you are saying, but if one player has an excellent decoder that hides errors (without loosing detail) and another player has a more basic decoder that doesn't, I still blame the decoder.

We had this with early software decoders in PCs, they didn't create compression artefacts, they were just crap at dealing with them.
Well, I guess the point is that with these next gen HD formats, artifacts as bad as illustrated in the first post shouldn't be there in the first place.

...but if one player has an excellent decoder that hides errors (without loosing detail)...

the keyword is "hides". If a decoder needs to hide anything, then the data itself is less than ideal, the problem IS there to begin with.

Regardless how you want to view it, it's still a codec problem.
If it is not a codec problem, the edge of the blocks won't need to be hidden.

I'm not saying VC1 doesn't have that problem, but I do see this types of 'problem' mostly on DVD, then BD, then the least amount (but STILL visible to my eyes) on HD DVD.

the keyword is "hides". If a decoder needs to hide anything, then the data itself is less than ideal, the problem IS there to begin with.

Regardless how you want to view it, it's still a codec problem.
If it is not a codec problem, the edge of the blocks won't need to be hidden.

I'm not saying VC1 doesn't have that problem, but I do see this types of 'problem' mostly on DVD, then BD, then the least amount (but STILL visible to my eyes) on HD DVD.

True Dat!!!!!

Furthermore, if a decoder actually hides blocking artifacts, the very same decoder might hide details in a GOOD video stream, too. I much prefer an honest decoder, even if it shows artifacts more clearly in bad video streams. However, there are ways artifacts can be multiplied, e.g. by sharpening algorithms. Of course a decoder should in no way intensify artifacts.

Furthermore, if a decoder actually hides blocking artifacts, the very same decoder might hide details in a GOOD video stream, too. I much prefer an honest decoder, even if it shows artifacts more clearly in bad video streams. However, there are ways artifacts can be multiplied, e.g. by sharpening algorithms. Of course a decoder should in no way intensify artifacts.
I like players that give us options. Some (at least for DVD) seem to filter everything, while others don't seem to filter anything, and some have settings to allow us to choose.

It will be interesting when more with other 1080p 1-to-1 pixel mapped displays check out these time codes. We've already seen it smoothed out on 720p displays, but it's definitely there.

Can the original poster provide examples from other movies?

Per David's suggestion, I reshot (see attached) at 1/25 second f2.8 I didn't see any significant difference between this shot at the one exposed for two seconds in terms of sharpness. Neither shows the streaking shown in the OPs photos.

Getting better. At F2.8 you'd better make sure you're paralel to the screen and that you have a quality lens. If the frame is on pause the exposure time shouldn't make as much difference as the aperture and the actual exposure itself. FWIW, the tripod and mirror lock up are only for preserving fine detail lost as a result of motion blur. All of the shots I posted are nearly full size images reduced in resolution only, not crops of the whole scree, so motion blur isn't even a slight factor.

Also, motion blur would actually hide artifacts by softening crisp edges. 1/25th is kind of hard to handhold, but it surely be done if the lens is wide enough, and you hold really still.

Can the original poster provide examples from other movies?


I will when I get some more time. My wife is starting to question my sanity. ;)

I would really like to do a direct VC1 to MPEG2 comparison in this fashion with a single title so that we can truly identify what is related to codec and what is related to display, source, etc. I wouldn't even be surprised to find that some of these releases are recodes of extreme bit depth lossy formats, introducing new artifacts to old ones.

VC1 has it's own isses (i know) as evidenced by artifacts in the 1080p trailers that I downloaded which are only 8mbps, but other than a few MISERABLE encodes I have yet to find glaring fault with the VC1 titles I have seen. Now Perfect Storm (bobbed mess) should have never been released on HD-DVD like that, and Traffic looks like it was encoded from an analog SD source...but I still didn't find compression squares on either.

BB Online tells me Aeon Flux is "long wait". Can anyone give me some ideas of a title that looks good, and is VC1 on HD-DVD and MPEG2 on Blu-Ray? i think that comparison could turn out to be pretty revealing.

M:I 3. IIRC is MPEG2 on BD and VC1 on HD DVD.

M:I 3. IIRC is MPEG2 on BD and VC1 on HD DVD.

Yes, almost all paramount encodes are MPEG-2 on BD and VC-1 on HD DVD. If you want a control, Corpse Bride is a VC-1/VC-1 encode that many have compared extensively (posturization issues).

just a note, 2 sec exposure is too long and will diminished the block noise. For video, ideally 1/60 sec is recommended. Tripod and cable shutter release is a must, MLU is optional. f/8 is definitely going to smear the block edges too. Since the object is flat, f/2.8 or wider is recommended.

surely any lens will be sharper at f8.0 than wide open (though as Bill0711 implies that he's using a 400mm f2.8L the distinction may be moot - fabulous lens)

Cheers

Eric

surely any lens will be sharper at f8.0 than wide open (though as Bill0711 implies that he's using a 400mm f2.8L the distinction may be moot - fabulous lens)

Cheers

Eric

Depth of field is 8 times narrower at 2.8 with a 400mm lens than it is with a 50mm lens. If it is indeed 400mm it almost certainly needs to be stopped down. Use a wider lens if possible, though use cautin wider than 28-30mm since barrel distortion gets more extreme.

Maybe we can get overclkr to take some closeup screenshots for us too.

http://www.avsforum.com/avs-vb/showthread.php?p=9387853&&#post9387853
http://www.avsforum.com/avs-vb/showthread.php?p=9387869&&#post9387869

CRT based display. Too good to show this artifact in all of its glory. CRT doesn't show harshness well. :)

Man, I love CRTs...

surely any lens will be sharper at f8.0 than wide open (though as Bill0711 implies that he's using a 400mm f2.8L the distinction may be moot - fabulous lens)

Cheers

Eric


Yes and no, depending on the lens. Regardless, I'm not talking about bluriness due to the f/8 aperture but due to the longer shutter opening, the jitter caused by the TV, floor vibration etc will lessen the edge clarity of each individual block.

peace out.

Benwaggoner, I hear what you are saying, but if one player has an excellent decoder that hides errors (without loosing detail) and another player has a more basic decoder that doesn't, I still blame the decoder.
Well, lots of folks argue about how much postprocessing should take place in a player. Remember all the compaints about "noise reduction" in the Samsung player? That's basically what we're talking about. The techniques that hide errors inevitably reduce detail some as well.

Most DVD players haven't done postprocessing with MPEG-2, for example.

We had this with early software decoders in PCs, they didn't create compression artefacts, they were just crap at dealing with them.
Computer screens are typically MUCH higher quality than televisions, so a level of decode sufficent for SD 60i S-video can look terrible on a nice big RGB display fed by VGA.

Yes and no, depending on the lens. Regardless, I'm not talking about bluriness due to the f/8 aperture but due to the longer shutter opening, the jitter caused by the TV, floor vibration etc will lessen the edge clarity of each individual block.

I've never used a lens that wasn't sharper stopped down than wide open (though I've spent a lot of money on L series glass to minimise the difference). However you make a good point on shutter speed. Certainly if I was using a big prime lens for photographing a projected image I would use it wide open to get as fast a shutter speed as possible even if tripod mounted.

Eric

Anyway the above is off topic, sorry for the diversion.

e

Sharp Aquos are among the best if not the best consumer LCDs, just check any reveiw site like CNET

Sorry, but it's no longer 2004.

Before I got the Sharp I had a Samsung LN4095 (which didn't meet my expectations), returned that for a Sony Bravia (v2500) which also didn't make the cut. If you think the Sharp (with NR off of course, Sharpness at 0) is showing artifacts that the Samsung and Sony won't show you must not have as discriminating an eye as I. Both the Samsung and Sony show pretty obvious Black Crush on all sources, and the Samsung can't even display 1/1 1080p.

At least those sets have something of a black level, not a GREY level, which is what the Sharp displays.

Why can't anyone on the AVS spell "lose" correctly?

Do they watch the TV show "LOOST"?

We should be preparing our flame suits since it is normally considered bad manners to criticize spelling on an internet forum, but the one that has really been bothering me a lot lately is "noone". What the hell is a "noone"?

Two words people: no one!

:p

add for advertisement:D

add for advertisement:D

"Comming soon"

This is exactly the reason i want to have a HD title PQ thread that is format neutral...meaning we judge on PQ only and list what codec was used.....i think a lot of us dual format adopters would be very interested in this. All i want is the best PQ available........My opinion is that Sony does not want to use VC-1 because it is a MS codec and this would be seen as working with the enemy....there is a lot of money to be made why not put games(both xbox and ps3)on BD and Movies on HD DVD??
If not Sony needs to get on board and start properly encoding these titles.....i have 2 bd movies and PQ compared with Most of my HD DVD are not even close.....Sony is butchering these titles and it is just unneccesary.....especially when these movies cost $25 on the average.....

everybody take it easy....

I don't recall any one wanting to go back to old software decoders on PCs because they show mpeg2 artefacts well. This has nothing to do with filtering, the newer decoders show far more detail.

Also, I doubt D-VHS tapes are getting this kind of scrutiny much. D-VHS was great for the displays of a few years ago, but I don't think people are watching them much with 1:1 on 1080p displays. They'll definitely see artifacts with D-VHS there as well.

I am certainly still using D-Theater in my cinema, at least until the films are replaced, though I am currently concentrating on HD-DVD and Blu-Ray. Having said that the films that I have seen have been clean but suffer from that MPEG artificiality that I dislike. Also, D-Theater is only approx 1440 horizontal so has less resolution than the new formats.

Mark

Of course, HD DVD uses higher VC-1 bitrates where the blocks don't get overly compressed to that degree anyway. But we degrade more smoothly than MPEG-2 under bandwidth stress.


Ben,

I thought that BD uses VC-1 too?

:D

Sig updated. BTW, Dolby Plus and True HD are not irrelevant on my PS3 which has both HDMI 1.3 and decoding for both codecs.

Ben,

I thought that BD uses VC-1 too?

Yes, BD titles using VC-1 will look just as good as HD DVD titles using VC-1.

BTW, I've asked in the insiders thread why those WB discs dropped the lossless (true HD) audio on BD that were included on the HD DVD. Has anyone found out why?

Stull waiting on Aeon Flux Blu-ray and I am plannign to directly compare VC-1 and MPEG2. Anyone have a copy they care to loan me?

Also, I wanted to post my impressions of a couple of more Blu-Ray titles I tried today. I picked up Eight Below and Corpse Bride.

Upon inspection of Eight Below, I noticed that while I was able to find some compression blocks, I had to look particularly hard to find them. I was surprised to find (when I went to check the bit rate) that the title is in fact AVC. I didn't know that any titles had been released in AVC yet. Well, the artifacts can still be dug up, but I must admit I didn't notice them once while the movie was playing. This title is definitely a much better show of Blu-Ray than the others I've tried. Still, it isn't as good as my VC1 favorites...

Which brings me to Corpse Bride. Those that have seen this one likely already know, the encoding on CB is excellent. This is a VC1 title, and a pretty high bit rate one at that. I think this is a perfect example of this threads purpose. Here is a VC1 title playing on the same PS3, on the same display, with all conditions the same and there is not one instance of the artifacts I've found in PH and the other titles.

I am even more anxious now to do a direct MPEG2 to VC1 comparison. In my gut I'm pretty sure I already know what I'm going to find...

I am actually a little disappointed that the AVC title isn't as good as the VC1 titles. I expected those two codecs to be neck and neck. To be honest though, I've only seen one AVC title so I really can't draw a conclusion yet. Also, average bit rate of CB VC1 is actually higher than 8B it seems. Combine that with the fact that CB is a digital source and I would expect it to look better even at identical bitrates.

So far i've discovered major blocking in MPEG2 encodes, pretty minor blocking in AVC encodes, and none so far in VC1. There is hope for HD formats!

We should be preparing our flame suits since it is normally considered bad manners to criticize spelling on an internet forum...

I'm not sure which is more pathetic, that so many people can't spell, or that it is so accepted that it is now considered rude to point it out.

So now you're finding blocking in AVC. You're story is just getting better by the minute. Is there a single person who has seen blocks in Eight Below besides you? There are many blu-ray owners out there with high-end displays far beyond your little Sharp LCD. If there were defects this big I think we would have heard about them by now. I will say once again that there is something wrong with your setup. Most likely related to the PS3. I'll just remind everybody of Occam's Razor (http://en.wikipedia.org/wiki/Occam's_Razor)

BTW the average video bitrate of Corpse Bride is ~19Mbps and Eight Below is ~16Mbps.

One thing for sure, as previously mentioned, I also find blocking with VC1 on many of my HD DVD (namely Superman Returns, M:I 3, Batman Begins). They are just a lot more difficult to be detected compared to the MPEG2 compression.

So now you're finding blocking in AVC. You're story is just getting better by the minute. Is there a single person who has seen blocks in Eight Below besides you? There are many blu-ray owners out there with high-end displays far beyond your little Sharp LCD. If there were defects this big I think we would have heard about them by now. I will say once again that there is something wrong with your setup. Most likely related to the PS3. I'll just remind everybody of Occam's Razor (http://en.wikipedia.org/wiki/Occam's_Razor)

BTW the average video bitrate of Corpse Bride is ~19Mbps and Eight Below is ~16Mbps.

dont be a smartass.

i think a 52" 1080P LCD is a reference monitor, not neccesarily good for entertainment, but it will give you solid accurate pixels to analyse.

secondly go watch some 1080P trailers off apple. they are all AVC and riddled with artifacts. this is due to low bit-rate, but it shows that AVC is not immune to quality issues.

I'm not sure which is more pathetic, that so many people can't spell, or that it is so accepted that it is now considered rude to point it out.

it is rude to point out, becuase (no im not going to correct it) this is an informal communication medium.

let it slide. if however its a job application, Hammer The Biatch.

And do you think VC1 is immune as the OP claims?


It uses smaller block sizes so much harder to see.

I have Black Hawk Down and Kingdom of Heaven on the way for my retail PS3 in addition to Pearl Harbor and The Fifth Element which I managed to pick up cheap from eBay. I'll have digital 24-bit grabs ready first week of February, so I hope to see the best and the worst of BD there. As to where Pearl Harbor fits in, who knows? I suspect I'll have more fun finding that out than I will watching the movie :D

Apple Quicktime trailers have SEVERE blocking as well as other artifacts. It is hard for me to believe anyone watching them wouldn't be very aware of them.

I am even more anxious now to do a direct MPEG2 to VC1 comparison. In my gut I'm pretty sure I already know what I'm going to find...

So far i've discovered major blocking in MPEG2 encodes, pretty minor blocking in AVC encodes, and none so far in VC1. There is hope for HD formats!

Could you please (when you get a moment) post pictures of the lack of similar blocking in Corpse Bride? While I have the HD DVD, I have a CRT display and can't easily capture the images that you can (oh, I've tried...and failed). I did get some nice screenshots showing where my guns have drifted slightly since my last convergence though.

benes,

DEAL! :)

So now you're finding blocking in AVC. You're story is just getting better by the minute. Is there a single person who has seen blocks in Eight Below besides you? There are many blu-ray owners out there with high-end displays far beyond your little Sharp LCD. If there were defects this big I think we would have heard about them by now. I will say once again that there is something wrong with your setup. Most likely related to the PS3. I'll just remind everybody of Occam's Razor (http://en.wikipedia.org/wiki/Occam's_Razor)

BTW the average video bitrate of Corpse Bride is ~19Mbps and Eight Below is ~16Mbps.

Benes,

Not picking an argument with you. I can see you don't like the Sharp and want it to be at fault, but let me remind you that both the films that exhibit blocking and those that don't are ALL being evaluated on the same display. My comments were actually not to bring 8B down, but rather to suggest that this is a glimpse of how things can get better when newer codecs get used. Also, I said I could find blocks in the AVC title if i searched (using pause and looking right up on the display) but also mentioned I didn't notice them while watching the film. This was not the case for PH or the other MPEG2 encodes. If you DL some of the AVC trailers on the internet, or Sony store you'll see that AVC at low bitrates becomes susceptible to blocks during high motion scenes too. It's not MPEG2, but you can see that it's roots must come from the same place.

Consequently, the VC1 trailers coded at low bitrates exhibit some artifacts of their own, but not in the form of blocks. It's more like quantization error in fine color gradients.

I can't find any blocks on the Corpse Bride Blu-Ray disc. It's a VC1 title, playing on the PS3, and on the same display (the one you don't like). You have suggested the problem is either the PS3 or the display. At this point I think you yourself should be reminded of Occam's Razor. Remember, these are pretty much "textbook MPEG2 artifacts". (sorry to use that quote from a previous post, but it said it so well)

At any rate, other than offering comparable 1080p 1/1 screenshots that are without the issue your speculation isn't really bringing anything new to the table anymore. I think I've pretty much cleared the display of any guilt, and the last evaluation goes a long way towards clearing the PS3's record too. When i am able to do a direct comparison of the same title it will be pretty telling. I'm sure you'll be one of the first to suggest that the Toshiba must be doing something to improve the picture and that the PS3 is still at fault. ;)

Could you please (when you get a moment) post pictures of the lack of similar blocking in Corpse Bride? While I have the HD DVD, I have a CRT display and can't easily capture the images that you can (oh, I've tried...and failed). I did get some nice screenshots showing where my guns have drifted slightly since my last convergence though.

Forceflow,

My intent is to do this with Aeon Flux when I can get ahold of the Blu-Ray version too. It is not a reference title, but still looks very good and I think it's a good representation of what we can expect from good encodes in the future. Plus, it has garnered a good rating on both platforms so I think the comparison is pretty equal.

I'd love to take the shots of corpse bride for you now, but I am afraid it's not going to be very meaningful without a direct comparison to show that artifacts WILL occur in the same spot when using a different codec. It's like trying to take pictures of something that isn't there. I promise to keep my word and deliver plenty of comparison shots when I get the AF blu-ray disc. I only posted the update because it seemed to help clear the PS3 and the display of some possible blame.

Forceflow,

My intent is to do this with Aeon Flux when I can get ahold of the Blu-Ray version too. It is not a reference title, but still looks very good and I think it's a good representation of what we can expect from good encodes in the future. Plus, it has garnered a good rating on both platforms so I think the comparison is pretty equal.

I'd love to take the shots of corpse bride for you now, but I am afraid it's not going to be very meaningful without a direct comparison to show that artifacts WILL occur in the same spot when using a different codec. It's like trying to take pictures of something that isn't there. I promise to keep my word and deliver plenty of comparison shots when I get the AF blu-ray disc. I only posted the update because it seemed to help clear the PS3 and the display of some possible blame.

Fair enough. I'm sure everyone who has the best interests of video PQ in general will thank you for your efforts.

I don't care what pictures you post I'll stand by what I've been saying all along: There is something wrong on YOUR end.

Damn the facts/evidence, damn the man!

...do you really think anyone in their right mind would release something with the level of blocking you show in your pictures? I've never seen anything like that even on DVD much less BD or HD DVD. Broadcast is the only place you will see that sort of thing.


1. You haven't seen The Fifth Element, Total Recall, The House of Flying Daggers?
2. Broadcast is NOT the only place you'll see that sort of thing. Apparently you don't have that kind of visual acuity. Which to me, is a blessing. I see similar things on both VC1, AVC and MPEG2 encodes (much lesser degree on VC1 encode) but they are there... and I'm only using CRT HD RPTV. I can only imagine how bad it's going to look on 1:1 display.

I own them all. And Fifth Element and Daggers are nowhere near as bad as everyone says. Total Recall on the other hand is actually worse than most people say.

If you're okay with those titles, then you're blessed with lower expectations (no sarcasm here). I, on the other hand, refused to even keep those titles. Heck, the HD DVD version of Total Recall is night and day compared to the BR version.

to me, it is as bad if not worse than what everybody else said. T5E and HoFD are the movies that made me return my 1st BD player (I'm on my 2nd player now).

How about this? Can anyone ELSE take pictures that show the level of blocking seen here? And I mean the SAME level of blocking. I'm sure there are plenty of other people out there with a high-end reference Sharp LCD display. If no one else can duplicate it then what does that tell us?

That would be welcome. I wanted to see how his display handled the VC-1 Corpse Bride, but he is correct, a direct comparison of a MPEG-2 vs. VC-1 title would be even more helpful.

Another source confirming this would be great. That 720p projector image did sorta confirm that this problem is severe enough to transcend 1:1 pixel mapped displays...

Anyone seeing rampant (or even existant) macroblocking on PH has setup issues, period. PH looks great, there is no black crush and there is no evident macroblocking on my 50" SXRD fed by the Panny BD player. Unless you want to argue that my setup is magically "removing" the artifacts...

People need to stop shooting their mouths off with "PQ is falling" FUD until they get a properly callibrated setup.

Anyone seeing rampant (or even existant) macroblocking on PH has setup issues, period. PH looks great, there is no black crush and there is no evident macroblocking on my 50" SXRD fed by the Panny BD player. Unless you want to argue that my setup is magically "removing" the artifacts...

People need to stop shooting their mouths off with "PQ is falling" FUD until they get a properly callibrated setup.

how do you explain the pics? Simply saying his display is introducing artifacts isn't correct. How do VC-1 encodes have less macroblocking (or none)?

I'm not knocking VC-1, I think it's a great codec (although no guarantee of a good encode...witness the awful banding in Superman). I need to find my Nikon battery charger so I can take photos of the image on my tv, but on my end I have no need to "explain" the pics the OP posted; PH doesn't look like that on my screen, so obviously his setup is introducing them (again, unless you're implying that my setup magically removes artifacts). I don't know enough about the OP's setup to say where they are being introduced, but they are not on the software or they would show up on everyone's screen.

I'm not knocking VC-1, I think it's a great codec (although no guarantee of a good encode...).

Agreed. No codec can play a larger part than its master. :)

Frankly, I just think we should hold off on saying that there isn't a problem until Cat. 5 delivers the goods on his Aeon Flux comparison. Even he may be surprised by what he finds...

The 1:1 pixel mapping is why I trust his rig is revealing what is on the disc. I can't achieve that with a 1080i CRT display that resolves <1920 horizontal lines. Not many have displays that are as accurate as an LCD can be (altho, I'm no fan of LCD period).

If you're okay with those titles, then you're blessed with lower expectations (no sarcasm here). I, on the other hand, refused to even keep those titles. Heck, the HD DVD version of Total Recall is night and day compared to the BR version.

I'm assuming the total recall was released with a different video encode for HD DVD versus BD?

Yup. Most likely VC1 vs. the MPEG-2 on the blu-ray. Released in another country from another distributor probably from a completely different master...etc. etc.

I am probably going to pick it up myself along with the Japanese version of Equilibrium on HD DVD.

Yes, different codec. I don't know about the different master, however, because I no longer have the BD version to do scene-by-scene comparison.

Stull waiting on Aeon Flux Blu-ray and I am plannign to directly compare VC-1 and MPEG2. Anyone have a copy they care to loan me?

Also, I wanted to post my impressions of a couple of more Blu-Ray titles I tried today. I picked up Eight Below and Corpse Bride.

Upon inspection of Eight Below, I noticed that while I was able to find some compression blocks, I had to look particularly hard to find them. I was surprised to find (when I went to check the bit rate) that the title is in fact AVC. I didn't know that any titles had been released in AVC yet. Well, the artifacts can still be dug up, but I must admit I didn't notice them once while the movie was playing. This title is definitely a much better show of Blu-Ray than the others I've tried. Still, it isn't as good as my VC1 favorites...

Which brings me to Corpse Bride. Those that have seen this one likely already know, the encoding on CB is excellent. This is a VC1 title, and a pretty high bit rate one at that. I think this is a perfect example of this threads purpose. Here is a VC1 title playing on the same PS3, on the same display, with all conditions the same and there is not one instance of the artifacts I've found in PH and the other titles.

I am even more anxious now to do a direct MPEG2 to VC1 comparison. In my gut I'm pretty sure I already know what I'm going to find...

I am actually a little disappointed that the AVC title isn't as good as the VC1 titles. I expected those two codecs to be neck and neck. To be honest though, I've only seen one AVC title so I really can't draw a conclusion yet. Also, average bit rate of CB VC1 is actually higher than 8B it seems. Combine that with the fact that CB is a digital source and I would expect it to look better even at identical bitrates.

So far i've discovered major blocking in MPEG2 encodes, pretty minor blocking in AVC encodes, and none so far in VC1. There is hope for HD formats!


please come back to this thread and post your pics when you do get AE, thanks

So now you're finding blocking in AVC. You're story is just getting better by the minute. Is there a single person who has seen blocks in Eight Below besides you? There are many blu-ray owners out there with high-end displays far beyond your little Sharp LCD. If there were defects this big I think we would have heard about them by now. I will say once again that there is something wrong with your setup. Most likely related to the PS3. I'll just remind everybody of Occam's Razor (http://en.wikipedia.org/wiki/Occam's_Razor)

BTW the average video bitrate of Corpse Bride is ~19Mbps and Eight Below is ~16Mbps.

and many of those own 720p PJs, which can't hold a candle to 1:1, no overscan 1080p when trying to detect artifacts. Until someone w/ a 1:1 1080p display which disprove the OP's claims, this thread will remain pertinent

I said "high-end". That excludes 720p. At least it does for me.

We've already had more than a few folks with 1080p displays disprove his claims. The burden of proof is on the OP and anyone else making similar claims. And a bunch of blurry pictures don't prove anything. Look in the software section and you will see pristine pictures from the same movie. Obviously they can't both be right. The thread is not pertinent at all until someone else can duplicate his findings. Until then its just another voice in the wilderness.

I've said it before and I'll say it again: One person doesn't take precedence over 1000. Lets get even one other person to find the same level of blocking shown here.

I'd like to see his AE or any other Paramount comparison shots until we declare this thread an anomaly

benes,

no need to question the OP's picture. Another member took a picture with his L-series lens, albeit on his 720p projector, still shows the artifacts (albeit much less so due to the downconversion from 1080 to 720. None of the beautiful shots on this forum (or any other forum) shows 1:1 ratio or close to that, it's always been a smaller sized version of the entire screen. I honestly doubt that if any of those pictures were taken properly and seen at 1:1 crop (100% crop) they're going to be that nice looking.

Put it simply, if the entire picture plane is less than (approx) 2.07 MP, then it's not a true representation of the real picture quality. I don't need to explain to you the reason since you're obviously know enough about imaging (and seems to know at least as much as the people who deals with imaging every waking moment of their lives).

the pictures in this thread is closer to 1:1 ratio than any other picture in this (or any other) forums. In fact the pictures here are LARGER than 1:1 pixel ratio. Hence the reason one can see the blocking clearer in these pictures.

We already know he claims to see these blocks on EVERY blu-ray release.

He said he saw blocks on every MPEG-2 BD release...big difference.

I myself like this thread, the posts, and the informative discussion....cat5 would you be interested in working with me to have a format neutral hd disc pq thread in where we would rate the BEST PQ hd disc's regardless of format??
In a spirited way that was an honest assessment of PQ across the board??
We could also post codecs used for comparison....like the jap import of total recall on hd dvd and the usa version of it on br??....this is just an example....

Deez - More than happy to offer my opinion, but I haven't time to watch every title in both formats so I can't promise to offer anything completely conclusive.

I got ahold of Aeon Flux BR today - A comparison is going to be a lot more work than I thought! What am I getting myself into. Anyway, here's a quick comparison just to get an idea of the differences. The BD version has plenty of blocks throughout, as expected. It also appears like it might be a hair sharper so I wonder if the PS3 is adding some sharpening (a no no in my book). you can guess which is which.

http://www.pbase.com/sbushman/image/72904902/original.jpg

Category 5, thanks a lot for making such an effort!

I have a Sharp (LD-37D6U), and on this display, sharpness setting of 0 doesn't mean no sharpness. While calibrating with AVIA testpattern with HDMI, I had to turn it down all the way to -10 to actually disable sharpness algorithm kicking in (it seems -10 for sharpness and -30 for color are settings to disable sharpness and color decoder respectively; someone can confirm or correct me). The most I could tolerate was -9 setting with PS3. Your set may be different (especially since it is 1080p), but have you checked your sharpness setting with a test pattern? Can you please try this? You don't have to take a picture or anything, but please, just check it out.

MPEG2 doesn't have advanced filtering to smooth out the edges of the macroblocks (unlike advanced codecs; combined with smaller macroblocks and not to mention advanced ME), so the blocks will always be visible to a certain degree. What I am getting at is that under normal circumstances, it is always going to be hard to find block edges in VC-1 (or AVC), but on MPEG2 it will always be fairly easy. I other words, VC-1 may tolerate more sharpening than MPEG2. At least I have noticed that I prefer higher sharpening with VC-1 encodes compared to MPEG2 encoded ones. Just a few cents from me.

Top image is BD

Top image is BD

yes, you can see the blocks. The issue of sharpness is interesting. Does increased sharpness affect the edges of the blocks making them more "rough" and visible? I would assume that it does affect this as sharpness increases the relief of edges of objects in the image, but does that include blocks?

yes, you can see the blocks. The issue of sharpness is interesting. Does increased sharpness affect the edges of the blocks making them more "rough" and visible? I would assume that it does affect this as sharpness increases the relief of edges of objects in the image, but does that include blocks?

it looks more like chroma blocking, rather than ee. standard mpeg stuff.

anyway here are some AVC blocking from Apple trailers

the first one is less bad, it could be grain i guess.

http://www.watertrack.com.au/file/blocks2.jpg

the second pic is good old sky blocking. im pretty sure this is chroma (macro) blocking. what do u think?

http://www.watertrack.com.au/file/blocks.jpg

Forceflow,
I am not sure about that. It would depend on sharpness algorithm. I don't think Sharp's sharpness algorithm is particularly good (at least on my set)! <- Ironic, isn't it? :)

IMO, worth trying out. I can't, right now.

Deez - More than happy to offer my opinion, but I haven't time to watch every title in both formats so I can't promise to offer anything completely conclusive.

I got ahold of Aeon Flux BR today - A comparison is going to be a lot more work than I thought! What am I getting myself into. Anyway, here's a quick comparison just to get an idea of the differences. The BD version has plenty of blocks throughout, as expected. It also appears like it might be a hair sharper so I wonder if the PS3 is adding some sharpening (a no no in my book). you can guess which is which.

http://www.pbase.com/sbushman/image/72904902/original.jpg


Well lets do it then......i will have to rent KOH BR to compare against kong as these are considered top tier in both formats...also, i think adding a separate audio section would be good as well.....i think this is fair as BR has uncompressed audio...which begs the question does using uncompressed audio have any affect on which codec to use and does it adversely affect PQ?? :confused:

using LPCM does not affect PQ (unless if it's taking the necessary space and bandwidth required to achieve kick-ass PQ). Theoretically, LPCM and (say) VC1 should be doable. Technically, however, I don't know.

Deez - More than happy to offer my opinion, but I haven't time to watch every title in both formats so I can't promise to offer anything completely conclusive.

I got ahold of Aeon Flux BR today - A comparison is going to be a lot more work than I thought! What am I getting myself into. Anyway, here's a quick comparison just to get an idea of the differences. The BD version has plenty of blocks throughout, as expected. It also appears like it might be a hair sharper so I wonder if the PS3 is adding some sharpening (a no no in my book). you can guess which is which.

http://www.pbase.com/sbushman/image/72904902/original.jpg
Run a deblocking filter over the first and it looks like the second. More interesting: How do they look in real time? Any difference visible? If yes, how would a MPEG deblocker (Algolith) change the look? Make them the same?

I still see the blocks in the 1st pic.

using deblocking filter doesn't change the fact that it only hides the problem.

The OP says he sees major blocking in Pearl Harbor and everyone believes him. I say I don't see major blocking in Quicktime trailers and no one believes me.
Blocking is definitely there in the QT H.264 1080p trailers (which is not surprising given the low bitrates). It's also there in WMV 1080p trailers.

using deblocking filter doesn't change the fact that it only hides the problem.

thats ok isnt it? if its hiden then there is no problem.

using deblocking filter doesn't change the fact that it only hides the problem.

I wouldn't paint deblocking filters in such a bad light, I mean VC-1 makes good use of one to almost eliminate blocking.

http://www.microsoft.com/windows/windowsmedia/howto/articles/vc1techoverview.aspx
Loop Filtering
VC-1 uses an in-loop deblocking filter that attempts to remove block-boundary discontinuities introduced by quantization errors in interpolated frames. These discontinuities can cause visible artifacts in the decompressed video frames and can impact the quality of the frame as a predictor for future interpolated frames.

The loop filter takes into account the adaptive block size transforms. The filter is also optimized to reduce the number of operations required.

What I have seen in every picture here is quantization errors on block boundary's, the only way to "solve" this problem would be to throw more bits into the encode to lessen the error OR use a deblocking filter.

If you or anyone else can come up with a better solution please let me know. ;)

de-blocking filter, regardless of codec, is bad. It hides imperfection... and to a certain degree, details. Be it VC1 or MPEG2, it doesn't matter. The fact of the matter, going back to the original topic posted by OP, is that MPEG2 block noise is more severe than any other codecs used for HD content, and that's been proven theoretically (based on technical papers/spec) and visually (from pictures and viewing experiences).

de-blocking filter, regardless of codec, is bad. It hides imperfection... and to a certain degree, details. Be it VC1 or MPEG2, it doesn't matter.
I don't think this is a problem with VC-1 as it is part of the spec, and is depended on for future motion compensation, disabling it would give unintended output.

I agree that for MPEG2 de-blocking does hide details, but I would prefer watch a slightly softer picture with no blocks.


The fact of the matter, going back to the original topic posted by OP, is that MPEG2 block noise is more severe than any other codecs used for HD content, and that's been proven theoretically (based on technical papers/spec) and visually (from pictures and viewing experiences).

I totally agree, and the only way for this to be solved is to dump MPEG2, and use VC-1 (or AVC if you hate Microsoft) for future releases.

thats ok isnt it? if its hiden then there is no problem.
Well, in the end the only thing that counts is how good it looks.

The problem with post-processing is that it can also reduce detail in parts of the image that wasn't an artifact. Now, done well, this shouldn't be too bad, but it's always there to some degree.

Also, because there are so many different ways to preprocess, it makes it harder to QA, since you don't know any particular player would decode the video.

de-blocking filter, regardless of codec, is bad. It hides imperfection... and to a certain degree, details. Be it VC1 or MPEG2, it doesn't matter. The fact of the matter, going back to the original topic posted by OP, is that MPEG2 block noise is more severe than any other codecs used for HD content, and that's been proven theoretically (based on technical papers/spec) and visually (from pictures and viewing experiences).
You need to discriminate between out-of-spec postprocessing filters, and in-spec in-loop deblocking filters like in VC-1 and H.264.

The in-loop filter is used for predicting future frames - there's only one right way to do it, so there's no ambiguity. And it's required to get accurate decoding, since future frames are based on deblocked previous frames.

Agreed. My apology for being unclear. When I'm "dissing" the de-blocking filters, I'm referring the out-of-spec and/or add-on filters beyond what the specifications have specified.

Agreed. My apology for being unclear. When I'm "dissing" the de-blocking filters, I'm referring the out-of-spec and/or add-on filters beyond what the specifications have specified.
Very true. The in-loop filter in VC-1 is adaptive based on what the encoder thinks is happening in the picture. The decoder deblocking filter in MPEG-2 on the other hand, does not have the luxury of having this data, so happily softens frames.

Very true. The in-loop filter in VC-1 is adaptive based on what the encoder thinks is happening in the picture. The decoder deblocking filter in MPEG-2 on the other hand, does not have the luxury of having this data, so happily softens frames.This is an excellent point. Many people ignore this when talking about filtering in advanced codecs.

Amir/Ben, do current VC-1 encodes use variable block sizes? Just curious. This may further add in visually hiding the blocking.

In MPEG2, even a person unfamiliar with this artifact can identify macroblocking once you show him one frame with macroblocking: because there is a familiar grid to look for next time!

Amir/Ben, does current VC-1 encodes use variable block sizes? Just curious.
Most definitely. We support 8x8, 4x8, 8x4 and 4x4.

This may further add in visually hiding the blocking
Very true. You are much more liable to see a 64 pixel block (8x8) in MPEG-2 than a 16-pixel VC-1 block (4x4). Also, the smaller blocks can contain grain and code it seperately from black space around it. Not so with a 64-pixel block which is liable to smear a lot of stuff together. And rectangular blocks can better track object edges, enabling us to better preserve their high frequencies (i.e. sharpness).

^^^ Thanks, makes a lot of sense.

Its possible to write a runtime mpeg filter to read ahead of the current frame and intelligently remove blocking. Why do you assume you can only do this at encoding time?

the only downside is that it requires more processing power, which there is more than enough these days.

I am no expert on this so Amir or Ben should answer this, but logically speaking, such a filter will have no real way to differentiate between real edge (real data) and the edge created due to blocking. You may end up softening the real detail. Also, such a filter won't know how much filtering to apply on different frames, and different parts of a frame. The filters inside the encoders have such information, which can be judiciously used by the operators (I believe) while encoding in addition to increasing instantaneous bit-rate.

IMO, an MPEG2 post-processing will not give the same result as achieved by internal filters in the encoders for advanced codecs. We all know about the stand alone video processors that try this, so it is not that nobody has tried. I heard very good things about Algolith Mosquito (http://www.algolith.com/index.php?id=mosquito_hdmi) (among others) which (claims to) "virtually eliminates blocking artifacts" in addition to mosquito artifact.

kdragon is right. In addition to the point he makes, if you do this in the encoder, you can vary the filter strength based on how much blockiness is generated. If you do it at decode time, you can not.

kdragon is right. In addition to the point he makes, if you do this in the encoder, you can vary the filter strength based on how much blockiness is generated. If you do it at decode time, you can not.


This is absolutely correct.

A simple test is to use a non-blocking h.264 high-resolution/detail encode and play it with CoreAVC.

First play it with deblocking disabled. You will get pristine PQ, no blocking, razor-sharp images.

Now play it with deblocking enabled (you might have to restart your player for the settings change to takle effect):
You will get an awefully soft picture. Absolutely inacceptable.

kdragon is right. In addition to the point he makes, if you do this in the encoder, you can vary the filter strength based on how much blockiness is generated. If you do it at decode time, you can not.

Amir, is the VC1 block filter per-video, per-frame, per-block or per-pixel?

What sort of filter is it? bilinear, cosine, bicubic etc? is it anisotropic?

Is there a white-sheet available?

The filtering is done on a per-block basis. However, even within a block, the algorithm makes secondary decisions on a per-pixel basis as to whether it should or should not filter (i.e. if it detects a hard edge or not).

As to filter type, it is again, rather procedural and not a straight filter. At its core, it does use a linear filter to keep complexity low but then compensates by using secondary algorithms to make sure the resulting effect does not cause too much blurring. In other words, there are a lot of “if” statements in there :).

The blocking is very hard to find on the first picture. Some blocking on the 2nd picture. Severe blocking on the 3rd picture to the point that you have to be blind if you can't see them.

I tried adjusting the contrast, brightness and sharpness of your pictures in Photoshop CS2 and the result are about the same. So my guess, both pictures are about as close to the original but the in-camera JPEG processing makes it worse (some cameras pushes contrast, brightness and sharpness like there's no tomorrow, especially point and shoot camera). However, since you're bypassing the display aspect completely, logically yours is the closer one to the source.

The above is just my 2 cents.

here's a simulation based on your picture + added sharpness/contrast/saturation a typical P/S camera adds:

http://i15.photobucket.com/albums/a372/sapiendut/93aa80ab.jpg

and here's the original picture taken by category 5

http://i15.photobucket.com/albums/a372/sapiendut/471d2766.jpg

a funny tid bit I just realized:

Pan and Scan, Point and Shoot (cameras), Piece of $h1t. They all share the same two letters: "P" and "S" :p

Blocking is easy to see on all shots at 1:1 when the image is paused, I doubt it'll be problematic when viewing in realtime at a suitable viewing distance. Except the third one though, that kind of fast motion macroblocking is usually easier to pick up.

The dude picture shows macroblocking mostly in the ground, the "dude's" pants. With Kate it's easier to see in her face, but there is some in the background as well as possible instances in the dark sections of the top she is wearing and less intrusively in her hair as well.

I'm sure if I were watching it regularly I'd not be too bothered.

Cheers...
Duy-Khang Hoang

The blocking is very hard to find on the first picture. Some blocking on the 2nd picture. Severe blocking on the 3rd picture to the point that you have to be blind if you can't see them.

It is interesting is that I have the third image being mirrorred to two different displays running with two different Native resolutions. The Blocks in the third image stand out like a sore thumb on the third one on the display that is running at its native resolution and are much less noticeable on the screen that is not running at it's native resolution. Try this, Look at the third image fully resolved running your fixed pixel computer screen at its native Resolution and then again feeding your screen some other resolution. It pretty significant how much of the blocking is hidden when you are scaling the image.

Erik

Also zooming in 4x to about the same degree he captured does show that the amount of blocking is fairly consistant with what he was able to capture with his camera.

The additional jpeg compression has two effects:
- removal of noise which is dithering the block structure -> block structure better visible
- reinforcing the block structure by running DCT type compression with the same block size -> block structure better visible
The pictures as on disc have all visible blocks if you apply a sharpening filter, but in real time and without sharpening applied the blocking should not be directly visible. It will be part of the general 'noise' in the image and can create a certain look, probably perceived as 'sharp'.
With AVC and VC-1 the blocks are of different shape and size so this helps hiding them better if they are not filtered out.

The first picture although it looks the best still has plenty of block noise.

I have attached a crop that has been sharpened to bring out the high frequencies, and resized to 2x.

Some technical comparisons between MPEG2 and VC-1

http://img343.imageshack.us/img343/8192/interpredictionmodesr5.jpg
http://img348.imageshack.us/img348/7822/comparisonlowjx9.jpg
http://img260.imageshack.us/img260/262/loopfilterlowzy0.jpg

source: http://video.ldc.lu.se/pict/WM9V-MP4AVC-MP2V_comparison-Goldman.pdf

It has been suggested in this thread that certain displays and players are smoothing out blocks which is why some people don't see them. I have removed the display from the equation and you can decide for yourself whether the player is doing any smoothing.

Not that I don't appreciate your effort, I just wondered how you could claim that "you can decide for yourself whether the player is doing any smoothing". This can only be claimed if the software player is not doing anything extra. I don't think we can find the answer to that that easily.

Theoretically there is no law against a software player adding extra image processing between the decoding and the rendering...

Hong.

benes - what software player are you using?

I've been doing full 1920x1080p screen grabs as well - but only for HDDVD (no BD-Rom drive available), and have found some software players produce a less than ideal image.

bones: From the looks of it I dont think your software is doing any smoothing or deblocking.

As the OP's pictures look similar to what you get when you run your PNG's through a sharpening filter, issue seems to be that the OP's display did a bit of sharpening that made the block noise more noticeable then normal.

But the point is still valid that the blocking is there, and it is objectionable to some people.

It is simply a weakness of MPEG2. I would challenge anyone to post a picture of VC-1 or AVC (from a HD or BR disk) showing this level of block noise.

Not that I don't appreciate your effort, I just wondered how you could claim that "you can decide for yourself whether the player is doing any smoothing". This can only be claimed if the software player is not doing anything extra. I don't think we can find the answer to that that easily.

I'm not quite sure I understand what you are saying.

I was just being picky. :p I was saying that just because it is a software-based player, it does not necessarily mean that it won't do any other post processing (e.g., smoothing. etc.).

Hong.

I never thought of checking this thread before but the noise seen in these pictures is what I noticed right away when I had a BR player (PS3) and bought all the best titles BR has to offer (according to AVS members). Of course this is always dismissed as film grain by the BR supporters, but I have done enough movie viewing to see the difference between noise and film grain, and this was definitely not film grain.

I have tried my PS3 with 4 different setups, and each and every one of them showed that noise (some more than others though). I hooked up my PS3 to a Mitsubishi projector (LCD), a Sony KDS-70XBR2, a 46' Sharp Aquos LCD and a 42' Philips Plasma. The first 3 are 1080p, the last one is that weird 1024x1024 resolution. Of course, the bigger the set the more I could notice it. It was very obvious on the projected 144' picture put out by the Mitsubishi projector as well as the 70' Sony.

Another thing : to my surprise, I didn't notice this with the Fantastic Four BR, so whatever they did, it did improve the picture. I'm not an encoding expert, but when a next-gen format commands a premium in price, I expect a premium in picture. I was not blown away by most BR titles. The only two titles that rivaled HD-DVD PQ for me were Fantatic Four and The Covenant, the others (BHD, KOH, Pearl Harbor, Talladega Nights, League of extraordinary gentlemen, ...) did not impress me at all.

I never thought of checking this thread before but the noise seen in these pictures is what I noticed right away when I had a BR player (PS3) and bought all the best titles BR has to offer (according to AVS members). Of course this is always dismissed as film grain by the BR supporters, but I have done enough movie viewing to see the difference between noise and film grain, and this was definitely not film grain.

I have tried my PS3 with 4 different setups, and each and every one of them showed that noise (some more than others though). I hooked up my PS3 to a Mitsubishi projector (LCD), a Sony KDS-70XBR2, a 46' Sharp Aquos LCD and a 42' Philips Plasma. The first 3 are 1080p, the last one is that weird 1024x1024 resolution. Of course, the bigger the set the more I could notice it. It was very obvious on the projected 144' picture put out by the Mitsubishi projector as well as the 70' Sony.

Another thing : to my surprise, I didn't notice this with the Fantastic Four BR, so whatever they did, it did improve the picture. I'm not an encoding expert, but when a next-gen format commands a premium in price, I expect a premium in picture. I was not blown away by most BR titles. The only two titles that rivaled HD-DVD PQ for me were Fantatic Four and The Covenant, the others (BHD, KOH, Pearl Harbor, Talladega Nights, League of extraordinary gentlemen, ...) did not impress me at all.

This thread would be grateful for any pictures that you could post documenting the block noise, as there are some people that still believe it does not exist.

Yes and we would also be grateful for any pictures documenting banding in VC1 as there are some people that still believe it does not exist either.
I would be interested to see that too, I never bothered to watch Superman Returns on HD DVD, as I had seen it recently on DVD.

You could say I'm curious :)

Of course then we have to argue if it was in the source :eek:

This thread would be grateful for any pictures that you could post documenting the block noise, as there are some people that still believe it does not exist.

Like I mentioned, I no longer have my PS3.

As for banding, I haven't noticed any on HD-DVD, but I haven't watched my copy of Superman Returns yet.

Like I mentioned, I no longer have my PS3.

As for banding, I haven't noticed any on HD-DVD, but I haven't watched my copy of Superman Returns yet.

Sorry I missed that part.

If you do watch superman returns the banding is supposed to be in the underwater scenes, let us know what you see.

So which pictures do you think are more accurate to the source? Mine or the earlier ones?


The OP did say anywhere in the film, but the first shots were from the attack sequence IIRC(?). If so, why don't you take a few shots from that part of the film.

Another thing : to my surprise, I didn't notice this with the Fantastic
Four BR, so whatever they did, it did improve the picture. I'm not an
encoding expert, but when a next-gen format commands a premium in price,
I expect a premium in picture. I was not blown away by most BR titles.
The only two titles that rivaled HD-DVD PQ for me were Fantatic Four and
The Covenant, the others (BHD, KOH, Pearl Harbor, Talladega Nights,
League of extraordinary gentlemen, ...) did not impress me at all.


While I agree with your assessment of those titles, posts like this show a misunderstanding about technology. You're not seeing differences between formats. You're seeing differences between *** STUDIOS ***.

Were Sony to release Talladega Nights on HD DVD, it would look the same.

When WB releases Blu-ray discs, they look just as sharp, just as clear, as they do on HD DVD.

Since both formats utilize the same video codecs, there are no limitations imposed on BD that restrict image quality (to the contrary, BD allows for a significantly higher bit-rate and data capacity should a studio choose to take advantage of it). Any trend with various titles looking bettter/worse on one format or the other is merely a function of studio mastering practices.

Theres no point in posting some random pics. If I knew where the first pics came from then I would post them. But I'm not going to search through a 3 hour movie looking for those shots. I already posted the same frames as the later pics and they don't show the same level of blocking.


Understood, but I thought that one of the points the OP is making is that the effects seem to occur moreso in fast motion scenes. If that's the case then only one of your pictures depicts fast motion (the car moving). At least that's my take on it. Is that unfair to say?

I can see this thread going on for 10-12 pages :(

OMG!!!

Nice call. And you made it way back on pg. 1!

This thread started late last night and it's up to pg. 8 already?

I got about halfway through pg. 1 and had enough. How's it goin' fellas? Did you iron this out yet?

EDIT: I guess I made it 9 pgs.

You're welcome.

:)

This thread started late last night and it's up to pg. 8 already?


Check the date again, although this thread was recently resurrected the debate started a long time ago :rolleyes:

Back OT even the pics benes posted on low motion exhibited minor blocking so the debate is definitely NOT over.

Check the date again, although this thread was recently resurrected the debate started a long time ago :rolleyes:

Back OT even the pics benes posted on low motion exhibited minor blocking so the debate is definitely NOT over.
Whoops.

O.K, so it took a month to get 1 pg. away from HD-DVDwonder's prediction fulfillment.

So now that we have a new thread cleverly reversing the title name I was wondering if the final determination on this was whether it was a PS3 issue or a codec issue?

So now that we have a new thread cleverly reversing the title name I was wondering if the final determination on this was whether it was a PS3 issue or a codec issue?
Codec issue. And I'm not a rocket scientist, nor do I play one on TV.

Its the 'signature' of MPEG-2, so yea its the codec ;)

http://img167.imageshack.us/img167/88/dude9pc.jpg
http://img253.imageshack.us/img253/3130/kate7nr.jpg
http://img263.imageshack.us/img263/8317/car8le.jpg
After reading all the comments about how awesome MPEG-2 can look like, I'm slightly disturbed now. The third picture shows fairly obvious blocking which is what this thread has been all about.

But even the first two pictures don't look film-like enough to me. Looking at the first one, I can see quite an amount of compression noise throughout the picture, it's very visible in the sky and on the buildings.
This is NOT what film grain looks like, it's what a mixture of compressed film grain and digital noise looks like.

The compression obviously can't handle these images properly. If you look at where the green fuselage of the plane meets the hangar you can see how noise and blockyness cause an ugly jagged, aliased edge.

Even blocks are visible in this image, for example on the face of the character in the foreground.

And because I know that some of you won't believe me, here's proof.

The first picture shows simulated film grain. This JPEG was saved at a very high quality and reproduces the grain very well.

The second picture shows the same image, but JPEG compression has been cranked up. I hope you can see the difference between film grain and overcompressed film grain.

http://www.digipiction.de/var/grain1.jpg
http://www.digipiction.de/var/grain2.jpg

Note how the second image looks very similar to what you see in the Pearl Harbor screenshot.

the difference between the two is very clear to me. I hope everybody else who claim to be videophiles can actually see the ugly difference between the two.

Benes,
What is your hardware setup? All I see is the software used (albeit I also notice how you are not going into complete detail on how you went from data on disc to your screenshots) but as most who work with computers can tell you video hardware, driver version, registry settings, etc.... can have a profound affect on the display output of any multimedia file.
If you want more here on this forum to understand what you are trying to do, providing us information on your setup (HW, SW, driver version, registry settings for the players) and your methodology of capturing the screenshots would greatly help.

May I also suggest that you post this in the HTPC forum where more PC pro's and hobbyists can look at this and provide more information either agreeing with your finds or disagreeing with it.

And because I know that some of you won't believe me, here's proof.

The first picture shows simulated film grain. This JPEG was saved at a very high quality and reproduces the grain very well.

The second picture shows the same image, but JPEG compression has been cranked up. I hope you can see the difference between film grain and overcompressed film grain.

http://www.digipiction.de/var/grain1.jpg
http://www.digipiction.de/var/grain2.jpg

Note how the second image looks very similar to what you see in the Pearl Harbor screenshot.


Is that grain the same thing that I'm seeing in this shot from Xylon's comparison pics thread; or is it just the scene?

Post #44 2nd pic (http://www.avsforum.com/avs-vb/showthread.php?t=811102&page=2)

Is that grain the same thing that I'm seeing in this shot from Xylon's comparison pics thread; or is it just the scene?

Post #44 2nd pic (http://www.avsforum.com/avs-vb/showthread.php?t=811102&page=2)

The grain in that pic does not look to bad...

But X-men is AVC, which does not suffer as much in this department as MPEG-2 does.

The grain in that pic does not look to bad...

But X-men is AVC, which does not suffer as much in this department as MPEG-2 does.


No, it doesn't look bad, but the grain structure is pretty evident. I haven't seen this film on any medium so I don't how it looks in person. On a side note, some P&S digitals will show this type of noise. Blue seems especially vulnerable on my little Canon A70. Every blue sky pic I've taken shows this noise.

Blue seems especially vulnerable on my little Canon A70. Every blue sky pic I've taken shows this noise.

I had that cam before. Have you tried using JPEG Superfine setting? I think the factory setting was JPEG Fine.

Is that grain the same thing that I'm seeing in this shot from Xylon's comparison pics thread; or is it just the scene?

Post #44 2nd pic (http://www.avsforum.com/avs-vb/showthread.php?t=811102&page=2)

Well, grain is always hard to compress, but I think the X-Men 3 pictures in that thread look much, much better than the Pearl Harbor shots.
I'm actually quite happy with the image quality that's shown of X-Men 3.

I had that cam before. Have you tried using JPEG Superfine setting? I think the factory setting was JPEG Fine.


Yep, I take all of my pics in Superfine. Boy, I miss my old Nikon CP 800. It was only a 2MP camera, but with TIFF capture mode it honestly took some pics better than this Canon.

With the supercrazy MP count, uncompressed TIFF is no longer feasible for point and shoot cameras. You may want to try, however, Canon's (or other companies') newer cameras, the JPEG compression is now better than before, especially with Canon (IMO).

In the norwegian avforum, one person posted this from World Trade Center HD-DVD vs. BR.

I dont own any of them, so I cant verify. But it dosent look good. Can anyone take a look at their BluRay copy of this movie?

neither can look that bad. The BD MPEG-2 pic does NOT show the tell-tale MPEG-2 compression artifacts. There are lots of artifacts, but that's definitely not MPEG-2 artifact.

In the norwegian avforum, one person posted this from World Trade Center HD-DVD vs. BR.

I dont own any of them, so I cant verify. But it dosent look good. Can anyone take a look at their BluRay copy of this movie?
The BD version looks worse than DVD. Until we see that reproduced, I would call that extremely suspect.

The BD version looks worse than DVD. Until we see that reproduced, I would call that extremely suspect.

Agree, would be nice if someone could verify this.

In the norwegian avforum, one person posted this from World Trade Center HD-DVD vs. BR.

I dont own any of them, so I cant verify. But it dosent look good. Can anyone take a look at their BluRay copy of this movie?

Sorry guys, after I linked this to the person who come up with the pictures, he admitted that it was a poorly attempt on a joke.

Can a MOD delete these posts regarding these pictures?

Sorry.

With the supercrazy MP count, uncompressed TIFF is no longer feasible for point and shoot cameras. You may want to try, however, Canon's (or other companies') newer cameras, the JPEG compression is now better than before, especially with Canon (IMO).


I've been scoping out some DSLR's for a while now. I heavily considered upgrading to the Canon Rebel series or the Nikon D50, but with 10MP cameras out now I'm going to wait for the prices to come down a bit more and score one.

I've been scoping out some DSLR's for a while now. I heavily considered upgrading to the Canon Rebel series or the Nikon D50, but with 10MP cameras out now I'm going to wait for the prices to come down a bit more and score one.

This will be my final out-of-topic post on this thread; Mobius feel free to PM me if you'd like to chat about this matter further (I know I'd love to)

MP count will mean absolutely nothing if you don't pair it with a great lens. It's better to use a lower-end body and spend the $$ difference to buy better lens. To put it simply (based on my own experiment), it's far better to get a used dRebel (300D; mint condition runs for US$300 or less) and spend the other US$1000 on a 17-55 f/2.8 IS lens rather than buying a 30D with the 18-55 kit lens (approx same total expenditure).

Sorry guys, after I linked this to the person who come up with the pictures, he admitted that it was a poorly attempt on a joke.

Can a MOD delete these posts regarding these pictures?

Sorry.
Well, of course you can delete yours. But I feel this is a perfect opportunity to illustrate the limitations of the forums. People need to be diligent and demand verification. Fortunately that is what occurred here, but it doesn't always work out so well.

It is clear from the post in question that your agenda was not to mislead anyone. However, if you want, I will delete my response. And if I forget, shoot me a PM.

I'd delete mine too (provided you've deleted yours, in order to keep the response flow in context)