Industry Insiders Q&A MASTER THREAD [separate thread for Xbox/Add On & PS3] - Page 84  

Thanks for the pointer. I am asking the team about it.


Yes, we have a ton of people who eat this dogfood (myself included). But most are now testing the new software update so they may not be seeing this kind of problem.

I don't think he does. Based on his location, I suspect he works for Telenor. Folks here probably don't know much about this company but Telenor has been quite active in development of MPEG-4 AVC. For example, their encoder/decoder for AVC won over the Nokia implementation and became the basis for the reference implementation (we did the same for MPEG-4 part 2). I am sure the industry is very appreciative of the role they have played in this regard.

I am sorry but I really can't. The entire reason this thread was created is because people got frustrated with our exchanges in the format thread, having a single goal of who can wrestle the other to the ground, as opposed to unearthing information in clear concise way that people can read and learn about. Your multipart post, with many quotes of my previous answers, looks very much like those argumentative posts of the past so I choose to not answer it.

You are of course welcome to assume I don't know how to answer the questions and assume you are right. My feelings would not be hurt one bit .

BTW, I do the same with some of the other questions asked here so don't feel like I have singled you out (see my response to Nils and lack of response to some other queries).

Note to mods: happy to have this post deleted if you see fit. {Nope. Helps to explain to others that this is supposed to be a question and answer thread, not a court of law. -Doc }

Amir,

I presume you are not referring to Sony Pictures right? Naturally you would know that we have sent samples to Dan and are having a dialog with him on the results.

Your post indicates again that Sony Pictures must have dark or political motives for using something other than VC1 for our titles, we don't.... As I have mentioned before, we author and compress for a number of studios, if they want their titles prepared with VC1, we are happy to oblige.

You are getting further and further out on a limb about the ability of AVC to reproduce the master, grain and all. To date we have done many side by side tests internally and with those who actually make the films. It is safe to say they know what their titles are supposed to look like and their opinion runs counter to yours. Look for confirmation in the next few weeks...

Sorry, but I can't give any specifics on the player release dates. Sony Electronics will want the honor of announcing their product plans..

Sorry for the late response...

I don't have all the specifics on this player. I know it will output 24p, but I don't have confirmation about the audio decoding specs yet...

Paid, since your back could you answer the following:

How much persistent storage is mandated for profile 1.0 BD players?

No persistent storage is required for these players.

Thanks amir.

You are correct. I work for Telenor, although not specifically with encoder/decoder development. I'm in the Broadcast divison (DTH/Cable/IPTV/DTT), but my HD interest is more of a personal quest for the best possible PQ/AQ in my Home Theatre.

Just as a side note, to prove I'm not on a anti-Microsoft campaign, we're upgrading our SMS system and part of that process is switching to BizTalk, .NET Framework 3.0 and ASP.NET 2.0 plus Atlas. I'm very happy with the results so far!

I don't understand why people continue to think that Warner is softening their transfers by using poor compression methods. Just because of a couple of movies that were shot with soft lens filters? (ie. Batman and Vendetta). Even Superman looked exactly as I remember it on dgital projection.

What about all of the sharp transfers like Training Day, Grand Prix, Four Brothers, ATL, The Searchers and The Ant Bully? Just off the top of my head.

Sure, here are a couple of my main questions, the others you can ignore or come back to when you get the time...

Since you claim there is a definite signature on AVC encodings (softness), are you saying that both MPEG2 and AVC cannot reach visual transparency using the bitrates avialable to it on Blu-ray?

When HD DVD's reviewed are rated less than 5/5, is that always the fault of the master?

Thanks.

Hello Amir,

Has it anything to do with the DVD Audio format from the good old DVD?

I wish, since there aren't almost any releases any more.

Thank you.

Hello Amir:

You mentioned previously about how all three codec's were compared against each other and VC-1 won/approved, I'm not sure if it was through the DVD Forum or some other org.

Anyway with the recent posts about having new A/B blind tests between codec's I was wondering if the same type of review that VC-1 initially won was a one time comparison or do they re-review all the codec's as time goes on and the codec's evolve with the technology?

Hello Tom. Good question .

The original DVD Forum tests consisted of two shoot outs. We won the first one. People thought we cheated (used pre-/post-processing to get better results which we had not). So they made the rules much more strict, but we won again. After this, VC-1 was made mandatory in the standard so there was no additional need for such tests.

However, the studios are not obligated at all to believe the above data even though most of them participated in it. As such, one by one, they conducted their own shoot outs, with their own material, data rates, and test conditions. This is how we won the support of Warner, Paramount and Universal. As there is no exclusivity arrangement here, the studios continue to evaluate new revisions of encoders from different companies and as such, we are in constant pressure to do better than our competition. This is why we know so much about the status of competing encoders and have gained in our confidence to stay ahead of them.

Your basic point that codecs, especially the new ones, keep getting better is very much true. But it applies equally to VC-1 as it has also improved (and rather rapidly) since the original DVD Forum tests.

As another meta point, folks should ponder why the codecs which did not win the shoot out, still became mandatory in the standard! I can assure you that if we had not won, there would not be a trace of our technology in either format. Anyone who tells you that business/other motivations don't get in the way of best man getting picked for the job, hasn't spent a minute in the circles where such decisions are made...

In a way yes. By adding these sampling rates, you get the same suite as DVD-A because TrueHD is the same as MLP used in DVD-A compression. But in a way, it is better in that DVD-A had an expensive (to license) navigation system. In the case of HD DVD, one simply makes an audio-only playlist (my preference) and you have a nice CD-type of experience which doesn't cost a penny more in licensing costs. And of course, you can also go crazy with HDi menus and such and blow away anything DVD-A could (not my preference).

Having said this, my personal opinion is that the time has come and gone for any kind of mass market around high-resolution optical formats. I much rather download the same files over the internet when I want it, than to get stuck with a few specialty retailers stocking a couple of discs here and there. And most of the world could care less about any improvements over CD it seems.

Amir,

I would like to hear your comments on what you are seeing in these MPEG-2 vs. VC-1 comparison uncompressed PNGs:

http://www.avsforum.com/avs-vb/showthread.php?t=813105

I am on a slow cellular modem in Hiroshima so can't look at the pictures easily. I have asked the team to look at it and when I hear back, and finish taking pictures at Miyajima (hey, work without play doesn't make sense in my book! , I will comment.

For now, I am surprised that you would say this in that thread, and then expect me to respond to you here:


If you don't believe my answers, not sure why you are asking me more questions here. We are not gladiators who you poke to fight for entertainment. You should at least give us some food and water before doing that .

I feel it's time to post something I've been PM-ing to those who have had their posts deleted:

"This thread is not a court of law. You're not a lawyer trying to catch someone in a lie or misstatement. Or badger them until they come around to your way of thinking. Take the answers the insiders give you and you do with them what you will. Questions. Answers. That's all that goes in that thread. I understand that the answers may come off as slanted. No surprise, there. Most of our insiders are admittedly not independent. You might not agree with the answers you get but this thread is not the place for challenging those answers. There is a whole forum where you can discuss it in an intelligent, non-bashing, non-confrontational way."

Ask your question. Get your answer. Move on.

Thanks

Dear Amir,

while you are at it please ask your team to take a further look at the recent "The Departed" VC-1 transfer.

Despite rather good marks "in the press" this Warner VC-1 transfer contains some of the most questionable sequences (purely from a technical point of view, the movie itself fully deserves it's praise) I have ever come across during my career in this business.

I'm preparing a thread dedicated to show the defects of this transfer. Right now I'm still figuring out a way to take meaningful "screenshots" (digital screen captures).

Before that please allow me to ask you for your professional opinion about this transfer.

Please ask your team for example to take a look at chapter 33 (dialog in hallway). There is some SERIOUS noise reduction going on. This scene is just an example of the worst case of noise reduction I have ever seen applied on any disc based HD transfer so far. The Departed has quite a lot of natural film grain. eg. in Chapter 33 there is no grain at all. It is not just that excessive NR is applied but the used filter is subpar - it smooths and smears the picture significantly, ALL detail and definition is lost, typical noise reduction artifacts are all over the place AND the VC-1 video stream AVERAGES 6.5 MBit/s during this scene (which is not surprising as there is no detail left to encode). Wait, it gets even better. After each cut in this scene the first 2 frames contain excessive amount of (natural) film grain. This film grain "magically disappears" from the third frame of each cut onwards - till the next cut. Cheap NR at it's worst.
As mentioned the VC-1 bitrate during this scene averages ~ 6.5 MBit/s. The very next scene of chapter 33 is not filtered - film grain is still present - and the VC-1 bitrate suddenly jumps to ~20Mbit/s. I don't presume VC-1 encoding itself is at fault regarding this inconsistent and excessive filtering (chapter 33 is just one example of many) but I would like to ask you why studios using VC-1 choose to filter their transfers prior to encoding in the first place. I just finished watching Chicago from Buena Vista - this AVC encoding with regular 40MBit/s peaks conserves the inherent, excessive film grain in a wonderful and natural way (remark for the moderators: suitable for this thread I worked as a projectionist for several years - so I would consider myself a "insider" regarding this matter).

Why do studios supporting your codec choose to apply excessive filtering/NR for many of their releases (Warner's Batman Begins, Charlie and the Chocolate Factory being other prime examples)? Does MS/the VC-1 codec team encourage this? If not, are you discussing this matter with the studios?

Looking forward to your comments! Thank you very much for your time.

Amir-

Any info on what's up with TL-51? I thought Toshiba would have submitted it this week for approval... Are they still planning on submittied it and having it approved by the end of the year?

Thanks,
Chris

Amir,

Sony Pictures participated in both of these events and my recollection of the final results is different. First off, the tests were based primarily on sub 10Mbps encoding, the total amount of encoded content was on the order of minutes and there were no direct view full resolution panels present. If you are claiming that these events proved anything accept that a reasonable picture could be produced as judged on 5 year old display technology, then I beg to differ.

In 5 years you will probably be be storing your movies on a server in Utah and retrieving them via Verizon FIOS at 100mb or better download speed over Fiber Optic lines. Hmmm, I wonder what encoding format they will use?

All sort of kidding aside. The cost will come down.

This will be driven by a combination of price pressure from retailers and pricing / positioning decisions made by Major Studios.

My pleasure, I trimmed my post because I felt it was off topic..trying to keep everyone happy here..

Rich

May I ask what Encode format we are comparing VC1 to?

Speaking for the former (BD/Uni), I'm aware of no formal discussions. One would assume that individual BD company execs use whatever relationships they have with Universal execs to encourage Universal to support Blu-ray.

I can't speak from experience, but I would imagine the same scenario applies for HD DVD and Fox/Sony/Disney. Of course, there are far more companies with far more execs in the Blu-ray camp to lobby a single studio, Universal, than there are HD DVD companies/execs to lobby Fox and Sony and Disney...

Amir, this strikes me as a gross oversimplification. A speaker is a physical construction with a bit of electronic shaping available. You might be able to tweak the sound slightly via changing the crossovers, amplification, etc., but the physics set some pretty hard limits.

A compression spec, on the other hand, can produce wretched results or extraordinary results depending on who does the implementation, efficiency of algorithms used, computational horsepower applied, etc. Joe Kane, an industry expert whose opinion I suspect you value very highly (based on his services frequently being hired by Microsoft) has said that AVC quality will most certainly surpass that of VC-1 as processor improvements allow more complex algorithms to become feasible. This opinion surely doesn't square with your contention that VC-1 is essentially always going to do a better compression job than AVC.