The Matrix Collection official...10/14 - Page 23

Art, I have to say I object to your cropping of my post to try and make me sound like an inflexible idiot. When taken in context with the rest of my (brief) post, my position is nowhere near as unreasonable as you would like to make it sound.

Here is my full post:

Ohh, it doesn't even come close to touching The Godfather, The Omen or POTC box sets.

Clever word-smithing there. The simple fact is, with either advanced codec, they will use as much bitrate as you allow them to have until the image is lossless to the material being fed in (try it: run an AVC encode using x264 (http://x264.nl) feeding it a raw black image, it'll spit out a file running in the low teens for kbps; feed it some high def material however, and it'll go as high as you allow it). These Matrix encodes (the audio/video combined/muxed) used the full space of a 30 GB HD DVD disc. However, their bitrates were limited to those used by HD DVD, and their capacity was strictly limited to that of HD DVD as well. Had these been encoded with Blu-rays capabilities in mind I assure you the software would have used that added space and bitrate.

I've never heard of a scenario where added bitrate didn't improve PQ (even if such improvement was barely noticeable or literally only visible by doing a pixel-by-pixel comparison of the encodes to the original source material). Certainly added bitrate never hurts.

Actually, modern encoders have two options usually.

1. A quality mode (where you give it some arbitrary "quality" number to shoot for and it encodes the material and keeps it at this quality regardless of the file size).
2. A bitrate mode (where you give the encoder the average bitrate to use throughout the encode) which usually requires two "passes"; the first pass to look at the complexity of each frame/scene, the second pass to actually encode the video and prioritize the bitrate based on the complexity info gained from the first pass.

I'm no professional, but my guess is that studios and "engineers" use the bitrate mode as it's simple to feed it in the correct numbers to maximize quality for disc based media. I imagine the workflow goes something like this:

• Encode extras (audio and video)
• Encode lossless audio and other audio assets (alt language tracks, commentaries, menu sounds, etc)
• Author in assets such as subtitles, menus and so forth
• Determine free space left on disc/media with these assets
• Determine video bitrate based on formula: optimal_bitrate = (free_disc_space div length_of_feature_in_seconds)
• Encode video using bitrate determined from previous step

Using that workflow you get the most bang for your buck, so to speak. Lazier studios wishing to reuse the video encode for future "special edition" or "collector's edition" releases may lower the video bitrate to leave space for future extras/bonuses (this is why I suspect some/most Fox titles leave anywhere from 5-10 GB of free space on their releases).

Anyways, I just wanted to correct your assumption on how the authoring process can work.

Actually, since the major studios can now use the same encode worldwide (no more PAL/NTSC mess), they would figure out which region is going to need the most space for extra languages and such before encoding. Extra foreign laguages would also be subtracted from the maximum bitrate available. They probably also leave themselves some overhead above and beyond that to play it safe. Everything these days is about cost effectiveness, which is why we did not get fresh encodes for the Matrix releases on Blu-Ray.

When I think of disk space and compression I think of the Superbit releases.
Sony said the big difference was using the full storage of dual layer DVDs @ 480i.
This is the paradigm that comes to my mind when discussing 50gb.

And something else to consider.
It seems like every DVD release of films from the last 30 years (where the director still lives) has been followed with a D.C. or Extended Cut, etc.
This chicken-shite marketing scheme by the studios has irritated nearly all of us, at one time or another.
Let us put an end to this, and insist on seamless branching and 50gb.

Yep, this is true. I was trying to keep it simple(ish). =) Also necessary for consideration: if you have PIP commentary (BonusView stuff) you need to lower the max bitrate by the max bitrate used by the PIP track (since PIP tracks use the same bandwidth as the main feature). Coincidentally, this is why I'm not a very big fan of PIP features: they lower the max bitrate possible for the entire feature.

That is a very interesting point; in this case, however, I feel the trade off is worth it (assuming the content quality of the pip track is worthwhile of course, such as the one found on Sleeping Beauty). That is, of course, subjective.

Well then i suggest you check out Xylons picture threads....Quite a few of Universals portovers had added bitrate and WORSE image quality.


No i was trying to say if the master isn't up to 2008 standards then they should spend money on a new one and give us the best possible image quality and that goes for all releases from every studio....I was just adding with some releases the master is in such good shape they can actually have a great looking Blu Ray but use less bits as they don't need to throw more bits at it to produce such an image....I don't personally know what shape their Matrix movie masters are in but someone suggested that possibly the artifacts are in the masters and i'm saying if thats the case then they should spend money on making new masters and give us optimum quality.....I don't know if that is the case though.

Found a temporary solution to the encoding problem: 3 frames are missing in the 1:45:46 mark that might cause pixelation or skipping artifact while playback. To avoid this, try switching the video resolution to 1080i in your player.

Bravo Warner for rolling out these low bit rate experimental encodes with DVDs in ugly black cases and sucking $75 out of us. Plenty of space = missed opportunity.


Well said provided they don't introduce other film artifacts in the form of DNR / EE that cannot be mitigated by supplying high bit rates.

And what exactly does playing at 1080i do to fix this? I tried that, same 3 skipped frames.

I wasnt talking about the Matrix movies there, but addressed that very few movies can utilise every space on a BD or HD disc unless you pack them with extras.

Like this. A HD DVD can run about 30mbs at full mux rate. On a HD30 that gives you about 133 minutes of full mux video. And since no movie needs to run at full mux (VBR is more effective then CBR), and most movies are under 2 hours. The space is more then enough. Of course this leads back to if 30mbs is enough muxrate from the beginning. But for 30mbs, 30GB will be more then sufficient.

The same goes for BD.

Full mux is 48mbs that gives us almost 139 minutes of footage. So for a muxrate of 48mbs, 50GB will be enough aswell.

If we now go back to The Matrix, so would a qualified guess be that a reencode would take more space if you bitbudget for it. But this is not the same thing as a bigger sized encode will look anything but marginal better.

Basicly, I use this a define lossy encoding in these different levels.

1. Encoding that hold up for screencap examination to the master.
2. Encoding that hold up for side by side playback with the master.
3. Encoding that hold up for viewing without access to the orginal master.
4. Everything else

So if an encode reaches level 2, no reencode can do much more then improving the number 1 level.

It seems to me that the people that are seemingly arguing that wasted space is a good thing keep falling back on other topics to try and prove their point.

If the Universal discs look worse with a higher bitrate it is because they changed other things during the encode (DNR, EE, etc); it does not look worse BECAUSE the bitrate is higher.

Likewise, if you compare a low bitrate encode of a flawless master to a high bitrate encode of a flawed master it is entirely possible that the low bitrate encode will look "better". This would seem obvious... starting with a bad master will sacrifice picture quality, but again, the fault of the bad picture on the high bitrate encode is not the higher bitrate but the bad master.

How about if we put it this way: if two encodes are done of the same master, utilizing the same exact "other" settings, the encode with the higher bitrate will be more accurate to the master (even if the difference is not appreciable to some users). In addition, the additional space on these discs is absolutely free and there is no reason not to utilize it for better accuracy.

I think this is the only thing the people on the side of NOT wasting the extra space trying to say... I think they feel it's obvious that we're only talking about the bitrate and that of course if other factors are mixed in it will affect the output one way or another.

Its not about if its a good thing. But if a reencode will not improve the image I rather se them send their money on encoding another movie.

Also since we also dont know how many manhours a new encode will get, we cant even say that it will look better, it may even look worse if they dont spend that much time on a reencode as they did on the first encode.

And when some people dont buy this title because the bitrate isnt high enought, then we go back to the question, if the orginal encode is true to the master. Because if its true to the master, there is no reason not to buy this title (other then if you dont want to buy the movie).

My issue with people wanting the max bitrate available is that it's really an arbitrary thing to harp on. Depending on movie length or extra features or soundtrack options, each movie is going to have a different possible max bitrate.

So for a long movie with significant extras, you'll be perfectly happy with a 30 Mbps encode (assuming they do a good job of course), but you will not be happy with a 30 Mbps encode for a shorter film? If the quality is the same, why do you care? Does just knowing that there are extra bits available on the disc ruin your movie-watching experience somehow?

I understand not wanting encodes that are too small. 15 GB is clearly not the upper limit for how good The Matrix Revolutions could look. But I think a quality threshold should be the goal for encodes, not an arbitrary bitrate threshold that will change across each and every disc released.

Would I complain if each disc had encodes that used every last bit of available space? No, of course not. But I'm also not going to complain about beautiful encodes, that clearly use enough bits to maintain high quality, just because it was possible to throw more bits at it.

This is arguably another misunderstanding on your part, in citing that there are people (rather than specific persons) who want max bitrate simply as an arbitrary thing. This is still quite different than persons who are noting issues and shortcomings in the final result (i.e, "quality thresholds") and suspecting that more bitrate would have been helpful. That is not so unreasonable, imo. Even if it is not, who are you to tell them they are wrong for believing so? Have you seen the master? Have they? It's just differing opinions at that point, and neither you or they should feel that theirs takes automatic precedence.

Even if the underwhelming result is entirely due to a poor master, that logically only leads to the notion that a better master would facilitate a better result, and it is highly likely more bitrate will be called for to allow that higher performance master to shine through. It's going to have more detail, more hf noise, more whatever, and that will naturally need to needing more data resources (especially under pk conditions). High performance video and high bitrate are simply complementary aspects in the scope of things.

Let's leave the master talk out of this - I'm fully aware that any encode can only be as good as the master, and that it's not the encodes' fault that more bits can't fix the problem. I didn't bring the master issue issue up, and I don't support it at all.

What am I misunderstanding? People are saying that no space should be "wasted." I disagree with that statement, since I think the underlying logic is faulty. 50 GB is an arbitrary limit imposed by the current standard we are using - different movies will need different amounts of space to achieve the same level of quality, so asking for all movies to fill that arbitrary upper limit is silly. By this line of thinking, we should never have BD-25 releases, ever, because we COULD use a BD-50 with a higher bitrate.

I'm not arguing against specific scenarios where movies could truly use more bits. If you said "a disc with noticeable compression artifacts but 10 GB free should have utilized a higher bitrate", I would agree. If you said "a disc with two lossless tracks, a ton of extras, and a poor encode squeezed on a BD-25 should have used a BD-50 so the bitrate could be higher", again I would completely agree. But a general rule of "always using 50 GBs" is completely arbitrary, and I don't think pointing out exceptions in any way validates that rule.

And if you are going to go the route of "any movie could always look better" (which I agree with in a mathematical sense, but not in a practical sense), then you are saying that all 3 hour movies are essentially compromised no matter how good a job the compressionists do.

On an optical disc, unused space is wasted space, period. A few 100 megabytes or even a whole gigabyte of slush fund space is excusable. Leaving GB's (plural) is getting a bit sloppy, imo...put that space into the movie. Make everybody happy. It's not hard to understand unless you are willfully trying to twist it into an absolutist concept. If we were actually in the domain of "the encoder filling in random bits to pad the bitrate" at optical disc sizes, then there would very well be no need for 1 TB masters and such. A lot of stuff is cut away from the master just getting to consumer sizes. Whether that stuff is stuff the consumer wants or is able to appreciate is another thing, altogether. In the cases where the master is just plain crappy and barely kissing the notion of "hi-def" resolution, it is only natural to question the use of that master to create an hdm product. You can substitute in a better (proper) master, but the result isn't going to be better if you just use the same bitrate that was seemingly fitting (maybe even overkill) for the earlier crappy master scenario.


That's really all you need to take from this, as far as citing what you've heard "people" say. The rest, you may consider re-evaluating if it really is "people" or "a person" that said it, and if that justifies insinuating that a group of people must be thinking that, when the call is made for "more bitrate".


Has anybody actually been pushing that idea? Is it "people" or "a person". If the latter, the same logic applies to not blanket "people" in with "a person".

There is more at work there than just the higher bitrates. Universal is obviously altering the masters to add more DNR (or other filtering) before encoding to BD.

Same here. My display is native 1080i, so I'm using 1080i output anyway. The frames are still missing.

Not true. The Matrix movies occupy between 22 and 26 GB on disc, so none of them uses the full capacity of HD DVD. The bitrates were obviously selected by choice, not due to limitations of the medium.

I'm not trying to actively accuse anyone (besides that one person who isn't even in this thread) of saying this. Please don't take it that way.

I'm saying that, outside of context, having an 'always fill 50 GB' rule is completely arbitrary and meaningless. I think agreeing upon this basic principle is a foundation for more reasonable conversation in regards to specific examples and what constitutes "waste."

Certainly reasonable conversation is desirable. In the same spirit, we should not have calls of "bitrate whores" and "people who watch the bitrate meter instead of the movie" sort of accusations every time the term "bitrate" is brought into discussion. It's an integral part of the entire process, regardless of whatever notions of "codec efficiency" and degrees of "transparency" are prescribed to.

Agreed. I'm not using those terms here, am I?

Agreed, but I think that needs to get out in the open as an inherent ramification for a call for reasonable discussion. Anybody else here who feels otherwise is not signing on to "the plan".

As others have already noted Universal changed things in the master prior to re-encoding them. The bitrate/codec had nothing to do with that.. for a better studio to compare against, look at Paramount (though be careful, they typically abandoned lossless audio on their HD DVD releases to keep the video bitrate from being killed).


AVC/H.264 will use as much bitrate as you give it. So even with variable bitrate (VBR) you'll get 30mbps if you tell it you have 30 GB of space and a movie that's less than two hours in length. Only if there's a ton of static imagery/scenes would you see the bitrate drop below the max (which for movies like The Matrix, is rare, if ever).


IIRC, the actual capacity of an HD30 is 27.9 GB, so it's understandable why they might drop below this. Do you have any proof/sources for your numbers, BTW? And is that the total disc space including extras/etc?

The skipping wasn't noticeable when changing PS3's video output resolution to 1080i.

http://home.earthlink.net/~jmorg/matrix/Matrix.htm

Maybe a bad flag left over from the HD-DVD encode?

It never looks like that on the PS3. The PS3 somehow flags those frames as bad and just skips them. Maybe you aren't familiar enough with The Matrix Revolutions to notice where the frame skipping happens, but I can tell you 100% that there is a quick jump during Neo's kick over Smith's head. I also verified it against the HD DVD copy to make sure my memory wasn't just bad.