Industry Insiders Master Q&A thread III: ONLY Questions to Insiders

[trbarry]

To whomever ...

Are all interim AACS licenses now expired? It seemed the last date was somewhere around now.

- Tom

[mentho]

Originally Posted by amirm

Hana (or ANA) is not used for scaling in 360. Instead, the graphics processor (GPU) is used to do the scaling. I won't disclose the internals of the filter but it is not bilinear at all. It is much more sophisticated than that with far more taps. Given where the hardware is, the scaling logic is the same in both products.

It's not used for scaling, really? I was under the impression it was ever since I read this article: http://arstechnica.com/articles/headstart.ars/2 Scott Henson is quoted as saying that Ana is the scaling chip.

So would you mind telling us what it does, in general terms if you must?

And more kudos from me for all your great contributions here -- I find myself checking this thread more than is probably healthy.

[rwestley]

Paidgeek, any news when and If Sony plans to put a forced 1080/24 option on the PS3.
I just updated to 1.81 firmware this morning and there is no fix on the latest update.

[BenDover]

amir or xbox folks, when will the elite become available again...there hasn't been any availability in my area since the original debut...are they waiting until they incorporate a new cpu into the console before manufacturing/distributing more elites?

tia

[amirm]

Originally Posted by mentho

It's not used for scaling, really? I was under the impression it was ever since I read this article: http://arstechnica.com/articles/headstart.ars/2 Scott Henson is quoted as saying that Ana is the scaling chip.

The article is wrong unfortunately. HANA/ANA are video encoders, not scalars.

So would you mind telling us what it does, in general terms if you must?

You basically have a bunch of pixels in memory ready to be displayed and you must convert it to the appropriate standard, whether it is composite, component, etc. You need to clock the samples are the right rate and format the signal (including modulating it for some of the output formats) before you can hook it up to you TV/monitor.

And more kudos from me for all your great contributions here -- I find myself checking this thread more than is probably healthy.

That's great. I wonder if at some point we should run a satisfaction poll to get a sense of how much people get out of this thread.

[RealEstateWagon]

Amir

If I in the near future have various "my-own-cuts" and bookmarks saved on the 360 add-on, will it be possible to via a USB memory stick share this with another HD DVD player, e.g. laptop? I know this is to be possible via Internet for HD DVD players, but just via the USB port I don't think is possible on the 360 at this moment. What you reckon?

[stpat]

Hi Amir,

The latest batch of Universal releases has been questionable in terms of video quality (40 Year Old Virgin shows lots of edge enhancement etc) and word from people who have gotten some upcoming releases early (Mystery Men etc) say they're just as bad.

Any idea what's going on? It seems like they're sacrificing quality for quantity, and while I can appreciate their efforts to get titles out, poor video quality is not the way to promote an HD video format.

Thoughts? Can you talk to anybody of influence and point out that this is detrimental to the format?

Cheers,
Stuart

[paidgeek]

Originally Posted by Jason One

Paidgeek, the PS3 is already upscaling interlaced material when it upscales DVDs, which are native 480i. Surely this can't be the reason the PS3 doesn't upscale standard-def extras on BDs.

You are correct that 480i can be upscaled from DVD. I don't know why it is not supported from BD added value so I have asked.

For the record though, DVD movies don't require interlaced upscaling as they are encoded as progressive frames. Exactly the same as HD-DVD in that respect.

[paidgeek]

Originally Posted by rwestley

Paidgeek, any news when and If Sony plans to put a forced 1080/24 option on the PS3.
I just updated to 1.81 firmware this morning and there is no fix on the latest update.

I was told they could not work it into this release. Hopefully it will be there in the next one.

[amirm]

Originally Posted by paidgeek

For the record though, DVD is movies don't require interlaced upscaling as they are encoded as progressive frames. Exactly the same as HD-DVD in that respect.

Say what? Both of these statements seem incorrect to me. Would you like to read them once more and see if you made a typo?

[Phloyd]

Originally Posted by amirm

Say what? Both of these statements seem incorrect to me. Would you like to read them once more and see if you made a typo?

Are you saying that HD DVD is not progressive? Because most DVDs these days certainly are flagged progressive.

Please clarify what you think is in error here.

[paidgeek]

Originally Posted by amirm

Say what? Both of these statements seem incorrect to me. Would you like to read them once more and see if you made a typo?

Bad grammar for sure..., but otherwise correct...

[UxiSXRD]

Maybe there's more to the required 2:3 cadence flags mandated by the HDDVD spec (ala the discussion with RB Films awhile back and rating the HDDVD version as 1080i60)? Hopefully Amir (and/or any other MS/HDDVD insider) could clarify.

[rwestley]

Paidgeek, thanks for the information about 1080/24 on the PS3. Did you get a chace to viw the Springsteen Sessions Blu-Ray disk. There are many posts on the Blu-Ray software thread about the poor dim quality of this disk. Do you know what the problem is?

[Josh Z]

Originally Posted by stpat

The latest batch of Universal releases has been questionable in terms of video quality (40 Year Old Virgin shows lots of edge enhancement etc) and word from people who have gotten some upcoming releases early (Mystery Men etc) say they're just as bad.

Any idea what's going on? It seems like they're sacrificing quality for quantity, and while I can appreciate their efforts to get titles out, poor video quality is not the way to promote an HD video format.

Thoughts? Can you talk to anybody of influence and point out that this is detrimental to the format?

And to balance things out, the Blu-ray insiders should be asked the same question about why, for example, recent Lionsgate releases such as Dirty Dancing and Basic Instinct have been so "questionable in terms of video quality". It seems like they're sacrificing quality for quantity, and while I can appreciate their efforts to get titles out, poor video quality is not the way to promote an HD video format.

[Robert George]

Are you saying that HD DVD is not progressive? Because most DVDs these days certainly are flagged progressive.

Really? Since when?

[9158]

paidgeek, is Sony considering adding these features in future firmware updates?

- ability to perform system updates WITHOUT having to plug in the controller with a USB cable

- ability to send game audio as well as players' voices to bluetooth headset/headphones, in stereo for headphones. Also, ability to send music and movies audio in stereo to bluetooth headphones. (see also this thread: http://avsforum.com/avs-vb/showthread.php?t=861481)

[amirm]

Originally Posted by Robert George

Really? Since when?

Since when indeed . Let me explain it for the rest of the people who seem to be under misunderstanding.

The source for DVD is usually a 480i tape. As such, what is being fed to the MEPG-2 encoder is 60i interlaced "video." The MPEG-2 encoder in real-time analyzes the source field by field, to determine if is progressive or not. Now, this test has to be "conservative" in that you don't want to throw out interlace fields by mistake as otherwise, the artifacts will be nasty. So the encoder errs on the side of flagging things that are in doubt as interlace. Result is that what is output, even for “movie” content, has mixed set of flags, indicating both progressive and interlace frames/fields. This makes the job of the processor in the player much more difficult as it can’t trust the flags to be correct (hence the reason good video processors such as HQV, DVDO, etc. ignore these flags and do their own image analysis to determine if the source is progressive or not).

Before anyone says the above is a corner case and doesn’t happen in reality, let me give you some examples. Since it was Paid who claimed DVD is progressive, why don’t we pick one of the movies out of his shop, encoded with their Sony MPEG-2 encoder . I happen to have files with complete flag list of the frames in movies The Fifth Element, Spiderman, an MIB 2. If you look in there, you see a ton of fields flagged as interlaced. TFE for example, starts completely in interlace mode! It then switches to progressive but at chapter points, jumps back to interlace. It also does this at other points in the movie. So if a reference quality title like this has mixed flag, not much hope is left for others which are not.

Now, HD DVD is a completely different animal. Here, we are being fed 24p material, not 60i. The encoder gets told the frame rate is 24p and it encodes at precisely that rate. As such, it will not perform any image analysis to determine interlace or progressive. As the movie is being encoded, the stream is flagged as being progressive with this additional metadata. These flags are guaranteed to be correct by definition since the encoder is locked into progressive 24p mode and never changes randomly as it does with DVD. So a decoder can easily trust them to produce proper 24p output.

Now you see why I was surprised when Paid made that statement. There is flagged encoding, and then there is proper flagged encoding . In some sense, HD DVD encoding is BD encoding with additional metadata for conversion to interlace. The two are much closer to each other than DVD is to either one of them.

[Richard Paul]

Amir, so there is no difference in how you encode 24p video between DVD and HD DVD except that proper encoding is more likely to be done with HD DVD? Also I notice that the DVDs you mentioned are several years old and isn't it much more likely that recent DVDs are now encoded using a 24p video stream?

[amirm]

Originally Posted by Richard Paul

Amir, so there is no difference in how you encode 24p video between DVD and HD DVD except that proper encoding is more likely to be done with HD DVD?

No, that is not correct as I clearly stated in my post. HD DVD is not just "more likely" to be proper. It is always proper. Again, HD DVD is like BD except we also add the flags for proper conversion to interlace. DVD on the hand can be random in its encoding. If you buy a DVD, you have no idea if it is encoded in pure 24p whereas you are assured of that with movie content in HD DVD. That is an important difference.

Also I notice that the DVDs you mentioned are several years old and isn't it much more likely that recent DVDs are now encoded using a 24p video stream?

First, this is immaterial to question at hand. Comment was made that DVD and HD DVDs were identical in how they are encoded where they are not. I even asked a second time and comment was confirmed.

If you want to say that DVD sometimes can be encoded with pure 24p, well, it can and I am sure we all agree that is a “good thing.” In which case, HD DVD’s similarity to DVD in that scenario is a good thing just the same. So no point is won by the other side in stating it .

Keep in mind that many DVDs are still produced with real-time encoders so the above problem remains. This is the reason most magazine reviews of DVD players test for this situation. But I am not here to put down or defend the DVD format, just saying how HD DVD production is much improved in this manner. So putting them in the same boat is very much incorrect.

[amirm]

Originally Posted by RobertR1

Amir,

Are you planning to address any of Universal's quality issues on lastest batch of releases or do we have to pray for a Universal insider?

Your question didn't lend itself to a quick answer Robert. You are asking me to make an assessment about a bunch of Universal titles I have not seen as I have been out of the country and very busy with other business. But I will take a shot tomorrow at making a few remarks, time permitting (trying to spend a bit of time with the family ).

[cybersoga]

There is no HD-DVD player other than software playback on a PC, that bypasses all interlaced processing. HD-DVD players ignore the film flags. The problem with interlaced processing is that it is not 100% perfect at reconstructing film frames every single time and there is the matter of the chroma bug rearing it's ugly head due to interlaced chroma up-sampling algorithm being used. Look at the jagged edges on the new Warner anti-piracy logo on Toshiba's best HD-DVD player vs Sony's best BD player. Sony's BD players don't use any intermediate interlaced processing for 24p mastered discs, they are progressive all the way.

any insiders care to comment?

[Tomeo]

Question for AMIR :

I really enjoy my 360 HD-DVD drive. Very good job

Just one question : does the MS HD-DVD team plan to develop a firmware that will add the 24p output feature like the PS3 ?

Thanks a lot !

[9158]

So the reason why BD extras in sd are not upscaled by the ps3 is that that content is stored as 60i with no flags for progressive output?
Have i understood correctly?

[Frode]

Amir, per our earlier conversation in this thread I was led to believe that the forced 24P stream was a feature of VC1, and not something specific to HD-DVD. You could still use a mixed stream with mpeg2 - is this correct?

[amirm]

Originally Posted by Frode

Amir, per our earlier conversation in this thread I was led to believe that the forced 24P stream was a feature of VC1, and not something specific to HD-DVD. You could still use a mixed stream with mpeg2 - is this correct?

You could but again, when anyone encodes movies for HD DVD, they start with 24p masters. They do not start with 1080i60 and ask the MPEG-2 encoder to attempt to convert it back to progressive. Whereas for SD, 60i is the common format where this process is occurring.

What you are saying is that if the source is a mix of interlaced and progressive content, can I create an HD DVD compliant stream that also looks that way. The answer is yes, the stream syntax allows for that. This extra flexibility in the syntax is a good thing. But that doesn't make the process the same as DVD. Here is something you desire to have (assuming you don't want to process the master and convert it properly to progressive). In case of imprecise inverse telecine in an MPEG-2 SD encode, it is not.

I would have been fine if Paid made all of these qualifications in his posts. But a blanket statement was made and hence my corrections.

[rover2002]

Originally Posted by amirm

Your question didn't lend itself to a quick answer Robert. You are asking me to make an assessment about a bunch of Universal titles I have not seen as I have been out of the country and very busy with other business. But I will take a shot tomorrow at making a few remarks, time permitting (trying to spend a bit of time with the family ).

Even the site administrator Alan Gouger is pee'ed of to the point he won't buy enymore Uni titles blindly ! he'll now wait for reviews.
Thanks for taking the time to view a few

[Phloyd]

Originally Posted by amirm

The MPEG-2 encoder in real-time analyzes the source field by field, to determine if is progressive or not. Now, this test has to be "conservative" in that you don't want to throw out interlace fields by mistake as otherwise, the artifacts will be nasty.

Can you please explain how this process is any different at all to the way the output from any HD DVD player is converted to 24p?

Since the data on HD DVD is flagged progressive, the output from the decoder is inherently 29.97 fps interlaced and any conversion to 24 fps has to essentially do the same process you describe above, regardless of the encoding.

Or am I mistaken?

Is there an HD DVD player that ignores the flags and outputs 24 fps without any intermediate interlacing?

[RobertR1]

Originally Posted by amirm

Your question didn't lend itself to a quick answer Robert. You are asking me to make an assessment about a bunch of Universal titles I have not seen as I have been out of the country and very busy with other business. But I will take a shot tomorrow at making a few remarks, time permitting (trying to spend a bit of time with the family ).

Thank you Amir. I look forward to your feedback. I'm generally concerned since these titles in question were really towards the top of my purchase list.

[amirm]

Originally Posted by cybersoga

There is no HD-DVD player other than software playback on a PC, that bypasses all interlaced processing. HD-DVD players ignore the film flags. The problem with interlaced processing is that it is not 100% perfect at reconstructing film frames every single time and there is the matter of the chroma bug rearing it's ugly head due to interlaced chroma up-sampling algorithm being used. Look at the jagged edges on the new Warner anti-piracy logo on Toshiba's best HD-DVD player vs Sony's best BD player. Sony's BD players don't use any intermediate interlaced processing for 24p mastered discs, they are progressive all the way.

any insiders care to comment?

Happy to comment .

First, I want to make sure we don’t mix topics here. What is encoded in HD DVD is progressive and is not subject to interlace processing. As otherwise, the identical VC-1 stream used on BD from Warner, would suffer which you agree it is not. So what you are asking about is how the progressive stream on disc is output by the player and of course, this is an implementation issue, not a format issue. The picture here, pun intended , is kind of complicated so let me explain with some examples.

First, PC players are not the only ones doing it right. Xbox 360 outputs proper progressive signals and has no chroma bug either . You imply that all BD players output proper 1080p signals because of the encoding method but that is not the case. The Panasonic decoder outputs 1080i and then feeds this signal to a (now obsolete) Marvel AVC2510 to convert it back to progressive. The Marvel part is a low quality de-interlacer as compared to the much nicer Reon part used in Toshiba XA-2. As such, it fails on many sequences like in MI3 which the Reon handles properly. More importantly, the Reon is very programmable so other things can be done to improve its performance in the future. The Marvel part is discontinued so it is hard to imaging Panasonic would get any support to improve its performance in the future.

And there is nothing inherent about HD DVD that makes it difficult for people to build players without a chroma bug. We have done it. I am hopeful that others will do the same (or upgrade their firmware to fix it).

As to Sony and its Pioneer brother, they suffer in reverse. Their do a low quality conversion from interlace to progressive. So if you play any video sourced content, you are not going to see anything as good as Toshiba XA-2. And for that reason, they may not play DVDs as well as the XA-2 either.

BTW, we have heard that the LG combo player outputs 24p with latest firmware with HD DVDs. I don’t own one so can’t tell for sure. But if so, you can add that to the list of players which output progressive properly from HD DVD.

All of this shows that we are talking about implementation factors, not format specifics. If you use a decoder which outputs 1080i, whether the source is BD/HD DVD, then you have an issue on how to output it back to progressive. The content on disc is progressive in both formats so the best course is to decode to progressive, and then output that.

Per my note earlier, for players which are programmable, I hope the manufacturers take the right steps to make them process the original video properly and you can bet that we put constant pressure on folks to do exactly that.


Now I really have to go and clean up the garden or I will be in serious trouble at home and my AVS privileges suspended for good .