High Def Digest Commentary: Specs Vs. Reality

[Petra_Kalbrain]

A very eloquently written piece.

Specs Vs. Reality

Before I respond I will let the masses voice their opinions... so... leave your message after the tone.

(beeeeeeeep)

[WickyWoo]

WEll, never doubt his motivation for writing this, he's an HD fanboy through and through

That being said, he does make some good points (the ludicrous "it's MPEG-2 it can't be good" or that "30GB is NEVER enough"

Notice the strategic playdown of the audio issue though

should mention at this point that at least one working Hollywood sound mixer has voiced his opinion that, when played back on his professional dubbing stage, well-mastered Dolby Digital Plus soundtracks encoded at the high 1509 kb/s bit rate that Paramount uses can be audibly transparent to the studio masters, when tested on movies that he mixed himself and would presumably know better than anyone else. But what use is the informed opinion of an expert in the field when it's easier to just point to the specs list on the back of a disc's packaging to make conclusive statements about matters of quality? In the forum on this site, a number of readers have made proclamations such as, "Compressed audio is just not acceptable these days" and "Whether you can tell the difference or not is irrelevant

here he cites FilmMixer from AVS, who shall we say does not work in a format neutral environment.

The disc's audio being indistinguishable from its studio master is "irrelevant"? Even with just a Dolby Digital Plus track, the 'Transformers' disc rated the highest score for audio quality that we can give. What more could we demand from it? It's absolutely terrific, but it's just not absolutely terrific enough if the packaging doesn't have a listing for TrueHD or PCM, even when it's likely impossible for human ears to tell the difference? What kind of argument is that?

By the same guy who says "it's great, but we don't know if it could have been better", pure spin

quality of the work is more important than the codec used to get there.

It's also more important than the bit rate. As far as I'm concerned, Sony's decision to incorporate a bit rate meter in their PS3 Blu-ray player is one of the worst things to have ever happened to the home theater hobby. Because of that one seemingly-innocuous and frequently-inaccurate data display, now just about anyone, no matter how technologically ignorant, can believe themselves to be experts in the field of video reproduction, based on nothing more than whether their bit rate meters read a high number or a low one -- as if that number were even relevant

That one I can agree with, however the people who run it through a measuring program on a computer to track actual bitrate should not be discounted

I found it extremely amusing to read complaints about the low bit rate used on 'TMNT', a disc with a razor sharp and amazingly detailed picture that some owners nonetheless decried as "soft" against the evidence their own eyes gave them, for no reason other than an ill-founded assumption that the picture would have been even sharper if the bit rate meter spiked a little higher. How would they know? Have they compared it against the studio master?

He's right there, the disc from the 10-15min I've seen is pure reference. It helps that it's short, and uses fairly simple animation.

As time goes on, compression tools and techniques become more efficient, requiring even less space to achieve visual transparency to the original master. Warner Bros. has many times over demonstrated outstanding results within the 30 gb limit of HD DVD, even on very long films such as the 'Troy: Director's Cut', a movie that runs 3 1/2 hours and yet fits comfortably on a 30 gb disc with beautiful picture quality

Cause you forget about the bit-starved first hour after all that blood and action

[SamInNorthCakalakey]

You cannot say what someone else sees or what they're feeling. It's impossible. What's good or bad for you could be just the opposite for someone else (and likely is).

All we can judge are the numbers. Big Blu outsells by a hefty margin. And since most of those consumers are the early adopters who crave better content than the masses, we have to assume they think Blu has better content.

Numbers don't lie. Blu players are more expensive and yet they outsell the Wal-Mart brand competition.

[KingDeezie]

All of the statements in that article are completely and utterly both unprofessionally written and straight up propaganda....

I thought that this person was supposed to be working for a neutral site?

Also, is anyone getting tired of listening to people refer to FilmMixer's statements about audio? Who knows who this guy works for, what his agenda is, and who is fluffing his pillow? Of course he is going to say that his work on "We Were Soldiers" for the HD-DVD release at 1.5 MB/Sec was equal (transparent) to the master, because he WORKED ON IT.

We as intelligent people understand that Halloween with an uncompressed audio track won't sound as good a Transformers with a DD+ track; but that's comparing apples to pears....

What we as intelligent people understand is that the same movie can be improved upon with the availability of more bitrate and space...

Do I feel that Transformers could look better on Bluray? No. AVC was used and the bitrate was high enough that I believe increasing the bitrate would just diminsh return....

Do I think Transformers could sound better on Bluray? HELLS YES. The inclusion of a TrueHD track, or better yet a 24 bit PCM track, would really improve the sound quality of the film...

And, right there is the beauty of Bluray. Sure HD-DVD can do some things great, but Bluray really has the capability to do everything great as far as todays tech standards are concerned...

Transformers with a high quality picture and higher quality sound makes it that much better and truly a great step ahead, and this is what Bluray would have provided.

I'll end this little rant with some more ranting in the fact that reading all of the fallout from this article has made me crazy and even less sure why HD-DVD supporters even exist. Bascially they are backing a format that has less space, less bitrate head room, smaller studio support, and smaller CE support. Still they fervently defend it, and go as far as to say that "On average HD-DVD has a lower PQ score because their studios have put out more catalog titles." Am I missing something? What the hell is wrong with these people?

[sj001]

I read this article a couple days ago, the guy is full of BS, plain and simple, and everything just seems to be defending HD-DVD. But using their standard tactic... "I'm neutral..." Yeah, sure you are.

Um, sorry but DD+ IS NOT the same quality as PCM. It isn't the same on paper, or listening to it, it is night and day for most of the tracks I have heard, and I am definitely not an audio expert.

[Canada]

I loved his comparing of Transformers and Dinosaures. I have never seen Dinosaures but I don't think it is too heavy on action. Example Volver has a PCM track, the plain Dolby digital on the DVD of Black Hawk Down probably sound better than PCM 5.1 of Volver. You need to compare the DD True HD and the PCM on a Blu Ray, 300 for example. I just got a new reciever and in the 2 min or so I noticed that PCM was better. It was louder and more crisp I got the feeling in that artical he was saying DD + was better this is bull s*** and it reaks. Next he is goin to say the DD on black Hawk Down DVD riveled that of the DTS track give me a break, as I said earlier bull s***.

[werewuf]

The "Reality" is that this was an incredibly irresponsible piece. I wrote into the blog and said so. Uncompressed sound has full dynamic range whereas compressed audio of any kind does not. For the different formats of DD and DTS there have been different levels of compression depending on whether your talking laserdisc or DVD or HD discs. It varies. DTS on laserdisc has less compression than DVD and if you compare the same soundtrack it is easily noticeable if you listen to the details. I have the Fifth Element on DVD and Blu and the difference is stunning between DD and True HD, and it does not take an educated ear to notice. Audiophiles for years have complained about the need to improve the quality of sound available in both CD's and DVD's. Currently there is an upsurge of interest in vinyl recordings because the new generations are finding out there is an audible difference between digital and analog and are seeking it out.
Personally I have never liked the sound of mp3's because there is a radical difference in the quality level of Cd's and way lower than vinyl. For me the argument spills over into the soundtrack arena as well. I still have my Pioneer CD-95 laserdisc player and have done recent comparisons on various soundtracks with DVD and lower compressed laserdisc (or uncompressed Dolby surround) DD and DTS are far superior. Now there are recent advancements in mixing and sound special effects trickery that didn't exist a few years back, however wider dynamic range is wider dynamic range.
Some of the yayhoos over there made the argument that 90% of people can't tell the difference or don't care. Maybe so, however there would be no audiophile industry if a decent enough percentage did not care. Dolby labs and DTS obviously think it's worth pursuing otherwise there would be no True HD and HDMA.

[Canada]

Originally Posted by werewuf

Dolby labs and DTS obviously think it's worth pursuing otherewise there would be no True HD and HDMA.

Exactly thank you

[Nicadian]

What really really angers me about this article is that just because he has a buddy that he is a proffesional, he thinks that we folks at home we don't know squat and we are bunch of idiots.

[Oblivion]

Not a bad little article IMO; I didn't think it was too anti-Blu-ray?

re Video I agree with all he says; MPEG2, VC-1 and AVC can all look great. It's all title and transfer dependant. Audio though I'm not so sure. I'm a big fan a lossless audio. Hand on heart whilst the Transformers soundtrack was one of the best HD soundtracks I've heard, I still think there is a real benefit to including lossless sound.

[Blu-Light]

As soon as I saw the site it was posted on I knew it would be nothing but a pack of FUD presented by your friendly neighborhood DUD lovin' reviewer. He is ANYTHING but neutral in his commentary. He has beat down BD and praised DUD since this war started.

[cajmoyper]

Originally Posted by Oblivion

Not a bad little article IMO; I didn't think it was too anti-Blu-ray?

re Video I agree with all he says; MPEG2, VC-1 and AVC can all look great. It's all title and transfer dependant. Audio though I'm not so sure. I'm a big fan a lossless audio. Hand on heart whilst the Transformers soundtrack was one of the best HD soundtracks I've heard, I still think there is a real benefit to including lossless sound.

I agree with your very first part, but Transformers kind of blended in spots and it didn't sound 5-star worthy to me.

[myst420]

This is the reply I posted over there in the land of the damned....

A lot of disinformation has been spreading across the web of late regarding bitrates, lossy compression, and whats “good enough”. More recently there was a lengthy opinion piece which portrayed opinion as fact. Its time that some of this opinion be refuted with scientific fact.

First lets talk video. All high def disks are currently compressed with one of three codecs. Those are mpeg-2, vc-1, and h.264 (avc). Mpeg-2 is an older codec while vc-1 and avc are newer codecs which achieve greater efficiency at the expense of computational complexity. It takes much, much more cpu time to compress using the newer codecs. As a general rule the newer codecs achieve about twice the compression as the older mpeg codec.

The first step in producing a video encode is to produce a digital master of the video to be encoded. This can happen in a number of ways but thats outside the scope of this writeup. Basically no matter how good of job done with compression or even if lossless is used the video will never look better than the master. Its critical that the studios do a good job of this step. Some of the early BD transfers are good examples of what happens when this goes wrong. (T5E and HOFD)

Now for the encoding step. The studios have a variety of tools available to them for doing the encoding. There are a number of tradeoffs that they must make. One is average bitrate, another peak bitrate, and the final is codec selection.

Codec selection: Basically at this stage there is no good reason to use mpeg-2. That leaves a choice between h.264 and VC-1. Each has its strengths and weaknesses. VC-1 appears to perform better when run at low bitrates which AVC appears to work better at higher bitrates. Disney is the one studio that appears to select between them on a case by case basis. Fox and Sony going the high bitrate AVC path while Warner and Universal prefer VC-1.

Peak Bitrate: This is a compromize that is determined by the choice of content on the disk. Each format dictates a maximum that can be used for all streams that need to be present on a disk. For example HD-DVD starts with a 30mbps budget and from that one or more audio streams need to be allocated. Audio is often a constant bitrate, for example the better lossy codecs use 1.5mpbs. So if a disk had just one of these that would leave 28.5 for the video. Typically a disk will have multiple audio streams but only one high bitrate one. If a disk supports picture in picture via a secondary video stream then this must also be subtracted from the total bandwidth. For the Warner disks that have PIP and multiple audio streams its not uncommon to see the peak video set down to the low 20s.

Peak bitrate has a great impact on the video quality of a movie during stressful scenes. From the perspective of the codec a scene is stressful if it has a lot of information to encode. Examples would be scenes with a lot of film grain or scenes with a lot of fast movement. Basically the less each pixel has to do with its neighbors and the more it changes from frame to frame the harder it is to encode. Example stressful scenes are the opening sequence in Casino Royale and the explosion in the opening of Swordfish.

To help with high stress peaks each format specifies a buffer which can be used to exceede the normal bitrate peak for a short period of time. However, if the stressful scene lasts for a substantial amount of time this buffer can be exhausted. An example of this is the explosion in Swordfish which becomes quite blurry and shows large compression artifacts as the available bitrate is not sufficient for the video sequence to be encoded.

Average bitrate: This is a disk capacity issue. Most movies are primarily made up of low stress scenes. Exceptions to this are movies with lots of film grain. As a result in the low stress scenes a studio can choose to increase the compression ratio in an effort to conserve space on the disk. The longer a movie is, the more likely this is to be necessary. If done well, its unlikely to be noticed.

What does this mean?

All lossy codecs, including all the codecs currently used for high def video make a trade off. In exchange for making the video fit within the limits of a disk they remove some elements of the master. This is not a debatable. It can easily be proven scientifically by comparing the master to the decode of the compressed video.

Given the same level of effort by the tech, the same codec, and the same tools, less compression will always result in a result that is closer to the original master. The codecs work by removing information from the master until they get something that will fit within the average and peak bandwidth limitations. The lower you set those limitations the more information is removed. This is a scientific fact and is not debatable.

The question is “will people notice” and “whats good enough”. A while back I did a test where I generated a video sequence then compressed it at various bitrates and posted screen caps from selected scenes in each. Most everyone that looked at the images was able to identify differences between the master and the decompressed versions. So yes, at the current bitrates we are not transparent to the master.

As to whats good enough, this is open to debate and is entirely up to each person to make up their own mind. Unlike some I will not tell anyone to be happy with reduced quality or that they will not notice. Personally if I am going to spend large sums of money on equipment, I want the best I can get, especially when the cost difference is minor.

Lets put an end to the argument of “good enough” Don't tell others whats good enough for them. Give them the facts and let them chose. Quality does cost more, right now its about $200. That will go down over time. If you want to say, "I saved $200 because I couldn't see the difference" thats fine, but don't pretend to know what others will and will not see.

[T-Wrecks]

by direct comparison on the same displays and sound systems, I suppose. I think he has a point about picture quality being similar, but sound and storage are legitimate deficiencies that will cause HD DVD to suffer in the long run.

Has this guy run an article on why 1080i and 1080p look exactly the same and therefore cheaper HD DVDs are a bargain? Just curious, because his line of thinking extends directly into this sort of argument.

[Seretur]

myst420 -- I saw your post over at HDD, and it's a much better opinion piece than the one Zyber posted; it's clearly written, understandable by laymen, and also happens to be factually correct.

Unlike his piece, I'm afraid. Zyber has written several truly informative articles there over the past few months, clearing up the usual issues anyone coming to HD might have (from aspect ratios to HDMI), but this one editorial isn't clearing anything up at all: it muddies the waters.

Basically, it's the equivalent of an article from circa 1998, positing a theory that non-anamorphic transfers are as good or perhaps better than anamorphic ones, as 98% of the users don't own widescreen TV sets.

I've read lots of those in my time, and have wholeheartedly supported Bill Hunt and the other aficionados who went on a crusade to demand anamorphic transfers as a DVD standard. Even though I bought my first widescreen TV in 2003.

Having lost the battle on most spec counts, HD DVD is now pressing up their advocating of the lossy audio -- not only in Zyber's post, but across the pertinent forums. It seems strangely similar to the "BD yield issues" they ranted about two months ago. And it's probably going to go away in time, as facts will prove them to be brutally wrong in any case.

In any case, I'm glad to see you on this forum, which is getting better daily.

[Clark Kent]

It's clear that Zyber is having problems with the fact that his chosen format is losing badly. This "article" was all about trying to minimize HD DVD's weaknesses and trying to convert fence-sitters into HD DVD buyers. This is just a more complex version of FUD and misinformation. I really hope his misinformation doesn't mislead unsuspecting people into buying into a D.O.A. format like HD DVD.

[jd213]

Originally Posted by WickyWoo

He's right there, the disc from the 10-15min I've seen is pure reference. It helps that it's short, and uses fairly simple animation.

I was watching TMNT again on Saturday, and I actually noticed posterization, blocking, and noise in some of the night sky shots when viewing on a new Sony VPL-VW60 projector on a 100" screen. I guess it's possible the posterization was in the original master, but I doubt if the other two artifacts were.

[thegland]

Originally Posted by cajmoyper

I agree with your very first part, but Transformers kind of blended in spots and it didn't sound 5-star worthy to me.

i would agree with you,transformers is not a 5 for audio,but i would put it at maybe 4.9

[darinp2]

Overall I thought it was pretty good, but left some things out. While the final result is what matters, often times looking for the reasons that things happen can help see where problems will exist in the future. I was watching the World Series tonight and thinking about how baseball has gotten to the point where teams analyze things they didn't long ago, like mapping where a player is placing the ball as he gets older and maybe predicting a drop-off in performance before it is obvious to others (and trading the player) or even analyzing things within a single game (like a pitcher's speed). I'm sure there are some who would argue that all you should look at is a batter's batting average or how a pitcher is doing for balls, strikes, and runs. Or with football, that they shouldn't put players through camps with they time them, check their strength, give them mental tests, etc. before deciding who to draft. Kind of like when Amir said to just look at what they had released, when a truly knowledgeable person in his position would have known that hadn't delivered some of the toughest stuff. Like seamless branching on a tough title, or even a 1.85:1 difficult movie like Saving Private Ryan with lossless audio and PiP.

And there is the factor that just because something can be done with a codec at a particular bitrate, that doesn't mean it always will be. Especially as things get beyond the more early adopter stage and the reality hits that this is a business and people doing encodings will be given limited timeframes to do things as we move forward. If I was a builder looking for somebody to paint rooms in houses I was going to sell I wouldn't give them 2 weeks to paint a room and then check the job, I would give them the amount of time I know I'm going to allocating for each room. As it is, studios are more likely to give more encoding time to titles they deem important (usually the biggest money makers, but not always), but most of us have titles we care about that aren't going to be the biggest sellers and might not be important to the studios. This is one reason I don't want their encoding people to have to do everything perfect to keep from ending up with problems. Even Amir has admitted that encoders aren't paid enough to catch every problem and so some will slip through.

I'm not sure if Josh thinks this way, but one thing I have seen on the AVS forum is people claiming that the test of whether a certain bitrate is enough for a format is to check the best stuff and see if the format can do that as well as with a higher bitrate. If it can then they will deem the bitrate for the format high enough. To me this says that they don't understand compression and these formats. If an engineer was tasked with figuring out what bitrate limitations to build into a format and they decided to base it on the best looking things as if things that won't look as good must require lower bitrates, I would consider them an incompetent engineer for the job. The difficult things can require the highest bitrates to get right (compared to the master), so if you only design in enough bitrate for the best looking stuff, other kinds of things will suffer. And if you only check the best looking (or "reference") things, that won't tell you if the bitrates are high enough for other things. Just look at which HD DVDs have the highest average bitrates and I bet it isn't the cream of the crop as far as the best looking HD DVDs. For instance, the original Full Metal Jacket had an average bitrate for video that was something like 26Mbps. That is peak for video and audio of 30Mbps. It isn't very good looking, but with bad compression could look even worse (and compression could be one reason it looked as bad as it did).

The bitrate requirements for clean things like King Kong and Hot Fuzz are likely to be lower than some things that won't end up looking as good to most people, even if they were watching the master, but if you choose peak bitrate requirements just for those clean things, then some other things are likely to suffer compared to their masters, from that peak bitrate limitation.

I honestly believe that many people arguing for HD DVD's bandwidth and size limitations would have also been arguing for a format with 2/3rds the bandwidth and size of DVD, if DVD had been in a war with such a thing. They could have told us how much cheaper it would be to make DVDs on modified CD machines and how the lower bandwidth and space wouldn't matter going forward, since codecs get better with time. They do if companies invest money in making them better, but studios also come up with more things they want to do. Amir wouldn't say what it was, but he said that at least one studio wanted to use seamless branching for something that users wouldn't even see (not like where users can select between 2 versions of a movie on their own). If Paramount decided to not put lossless audio on Transformers because of bandwidth or size limitations, what would they have done if they had to handle lots of seamless branch points? It isn't impossible, but adds to the difficulty of encoding things (bitrate limitations tend to drop a lot at the branch points), and it seems like they felt they were already limited without those.

--Darin

[LeoneFan]

Paramounted has admitted that they didn't offer PCM or lossless because of disc size limitation.