New H.265 twice as efficient as H.264!

[5th Of November]

I just saw this posted in the Technology section on Google News.

http://www.slashgear.com/new-h-265-v...ates-27266856/

http://hothardware.com/News/ITU-Appr...eaming-Video-/

If this only needs half the bit rate, then wouldn't it allow 4K enough space to fit on a 50 GB Blu-ray? I realize that the studios would have to quit putting extra stuff on the same disc as the movie, but why wouldn't it work? It would be cheaper for them to put the movie on one disc and any extras on a second disc, than it would for them to start using 100/128 GB BDXL discs. All that would have to be done is making new players that play the codec, just like they did with 3D. The discs can stay the same.

[vargo]

Originally Posted by 5th Of November

I just saw this posted in the Technology section on Google News.

http://www.slashgear.com/new-h-265-v...ates-27266856/

http://hothardware.com/News/ITU-Appr...eaming-Video-/

If this only needs half the bit rate, then wouldn't it allow 4K enough space to fit on a 50 GB Blu-ray? I realize that the studios would have to quit putting extra stuff on the same disc as the movie, but why wouldn't it work? It would be cheaper for them to put the movie on one disc and any extras on a second disc, than it would for them to start using 100/128 GB BDXL discs. All that would have to be done is making new players that play the codec, just like they did with 3D. The discs can stay the same.

It would make more sense to include BDXL compatibilty as part of the spec.

[lobosrul]

Originally Posted by 5th Of November

If this only needs half the bit rate, then wouldn't it allow 4K enough space to fit on a 50 GB Blu-ray? I realize that the studios would have to quit putting extra stuff on the same disc as the movie, but why wouldn't it work? It would be cheaper for them to put the movie on one disc and any extras on a second disc, than it would for them to start using 100/128 GB BDXL discs. All that would have to be done is making new players that play the codec, just like they did with 3D. The discs can stay the same.

4k has FOUR times (2x horizontal, and 2x vertical pixels) the resolution as 1080, not double. So even if h.265 only needs half the bit rate as h.264 (and I've heard its more like 30% than 50) the disc needs to be twice the size, given the same length of content.

[5th Of November]

So if BDXL and H.265 are added to the Blu-ray specs, then 4K should be no problem?

[vargo]

Originally Posted by 5th Of November

So if BDXL and H.265 are added to the Blu-ray specs, then 4K should be no problem?

I don't know if adding to Bluray specs is the correct term. 4K using H.265 is a as different from Bluray as Bluray was from DVD.

[Ruined]

Originally Posted by vargo

I don't know if adding to Bluray specs is the correct term. 4K using H.265 is a as different from Bluray as Bluray was from DVD.

The main issue is backwards compatibility with existing Blu-ray players. Even 3-D streams are backwards compatible with old BD players due to H264 enabling this. H265 discs though would not work.

IMO, seeing a new disc-based format for 4k is unlikely as Blu-ray still has yet to fully overtake DVD making 4k physical media adoption slim. The only way it would happen would be if it was rolled out like laserdisc, with $99 movies.

[Anthony P]

Originally Posted by vargo

I don't know if adding to Bluray specs is the correct term. 4K using H.265 is a as different from Bluray as Bluray was from DVD.

that would be like saying 3D BD is as different from Bluray as Bluray was from DVD

obviously the larrter is not true, it was added to the BD format and now we have 3D BD.

[vargo]

Originally Posted by Anthony P

that would be like saying 3D BD is as different from Bluray as Bluray was from DVD

No, not all all.

My point was that 4K H.265 on a BDXL disc is technically feasible. 128GB is more than enough.

But even if you are still reading a Bluray disc with a blue laser, it is very unlikely such a format would be called 'Bluray something' (ie Bluray Ultra etc) or added as an extension of the Bluray standards. Backwards compatibility is not possible. It is too technically disparate.

You would start fresh with new branding and new specifications.

[space2001]

When sony releases they 4k streaming service, I was told anyone who buy's the blu-ray will be able to unlock the 4k movie from the streaming device. They don't' know how they will implement it yet but that is how it is for now. Also it will only be sony movies for now, unless other studios sign on.

[Anthony P]

Originally Posted by vargo

No, not all all.

My point was that 4K H.265 on a BDXL disc is technically feasible. 128GB is more than enough.

But even if you are still reading a Bluray disc with a blue laser, it is very unlikely such a format would be called 'Bluray something' (ie Bluray Ultra etc) or added as an extension of the Bluray standards. Backwards compatibility is not possible. It is too technically disparate.

You would start fresh with new branding and new specifications.

why, didn't they have BD live, 3D BD..... maybe they will or maybe they won't

[Penton-Man]

Originally Posted by lobosrul

4k has FOUR times (2x horizontal, and 2x vertical pixels) the resolution as 1080, not double. So even if h.265 only needs half the bit rate as h.264 (and I've heard its more like 30% than 50)....

Lobos Rule!

Wolf Howling - YouTube

C’mon, lobos have got better hearing than that!

Let me tell you about dem humans. Perhaps that 30% figure is true if you think humans only see in ‘PSNR’. The reality is that humans see and process imagery ‘subjectively’. Read the pdf listed in this link for science beyond what you’ve heard…
https://forum.blu-ray.com/showthread...ly#post6994343

And/or, from a peer-reviewed article by Jens-Rainer Ohm et.al.,
http://www.hhi.fraunhofer.de/fileadm...erformance.pdf

b.t.w., that be the same Jens who co-chairs the JCT-VC committee that did the work on HEVC, which ITU membership recently approved….https://forum.blu-ray.com/showthread...189830&page=38

For those having an aversion to plodding through the scientific literature, in essence, evidence shows HEVC subjectively provides ~ 50% bit rate reduction on average compared to H.264/AVC.

[lobosrul]

Originally Posted by Penton-Man

Lobos Rule!

Wolf Howling - YouTube

C’mon, lobos have got better hearing than that!

Hehe, my handle actually refers to these Lobos: http://www.golobos.com/ We did not rule at all last Saturday

In short, a subjective increase in video quality is.... subjective.

[Penton-Man]

Originally Posted by lobosrul

...In short, a subjective increase in video quality is.... subjective.

Yes, but it is an entirely accurate parameter to evaluate the HVS (human visual system) as long as the test methodology is not flawed…and editorial review by peers, of either journal article, has found none.

Unlike flawed ‘objective’ measurement which is touted as accurate, like, for instance, that used as a basis for this tech journalist’s *insight*….http://reviews.cnet.com/8301-33199_7...vs-are-stupid/

namely… “but the average person with 20/20 vision can resolve 1 arcminute”,

ergo..“One arcminute of resolution is a best-case scenario."

lah-dee-dah-dee- dah with the math exercise from there.

Which, b.t.w., forms the basis of this chart which is touted on internet forums as a Commandment, rather than just a rough guideline for 4K viewing… http://carltonbale.com/1080p-does-matter/

When the fact of the matter is that this basic ‘objective’ measurement (one arcminute of resolution) is incorrect/limiting in its applicability to real world viewing because other than “resolution”, professional imaging scientists know and hold valid that there are other types of visual acuity which also play an important role in human vision, namely ‘recognition’, ‘detection’ and ‘hyperacuity’.

I don’t have time right now to elaborate further but, perhaps I’ll save that for another post on a rainy day.

[Anthony P]

Originally Posted by Penton-Man

Yes, but it is an entirely accurate parameter to evaluate the HVS (human visual system) as long as the test methodology is not flawed…and editorial review by peers, of either journal article, has found none.

I don't fully agree, there is subjective and objective but there is also ignorance and knowledge. I never trust early subjective results not because there is anything wrong with it but usually when you change something (resolution, compression schemes, encoders....) things will be different and so people are ignorant. What I mean is that my guess with H.265 they found ways to improve issues that existed with H.264 (that is usually the definition of "efficient"). Anyone used to H264 (and good for such a test) will be knowledgeable with PQ issues of h.264 over compression, now if h265 fixes those issues but causes issues else where (can't pay Peter without robbing Paul) the person at this time might not have the experience to judge them as critically.

[Penton-Man]

Originally Posted by Anthony P

I don't fully agree, there is subjective and objective but there is also ignorance and knowledge. I never trust early subjective results not because there is anything wrong with it but usually when you change something (resolution, compression schemes, encoders....) things will be different and so people are ignorant. What I mean is that my guess with H.265 they found ways to improve issues that existed with H.264 (that is usually the definition of "efficient"). Anyone used to H264 (and good for such a test) will be knowledgeable with PQ issues of h.264 over compression, now if h265 fixes those issues but causes issues else where (can't pay Peter without robbing Paul) the person at this time might not have the experience to judge them as critically.

Anthony, now that’s funny.

What I mean is that according to all the marketing wanabees on the internet who fancy themselves as prognosticators of the ultimate success, or failure, of 4K adoption by the consumer-at-large… the feeling is that in order for 4K to truly succeed amongst the masses, the increased clarity or sharpness, if you will, must be apparent to the most inexperienced observers…not only videophiles or experienced imaging professionals who can identify things like aliasing, for instance. That’s been one of the premier arguments, i.e…Joe6pack has to appreciate it.

Anyway, if you are skeptical of the laboratory findings of some of the best imaging minds in the business as regards to their methodology related to subjective response, then I can tell you that all ‘experienced’ viewers (searching for compression related artifacts), whom I’ve spoken with, were convinced of the ~ 50% efficiency claim comparing HEVC with H.264/AVC, also by Broadcom, once they moved from the floor….

Getting 4K with Broadcom's HEVC Encoding - CES 2013 - YouTube

to seeing the Broadcom demo in their private booth.

P.S.

Also, for some perspective, if you have a partly philosophic problem with subjectivity testing and methodology in general, then you must be whole-heartedly disenchanted with the .RED codec roll-out, because to the best of my knowledge, ALL their demos have been based solely on public subjectivity and they’ve never posted any objective PSNR comparison metrics, like for instance Qualcomm did nearly a year ago -

http://reviews.cnet.com/8301-13970_7...n-h.265-video/

[Taikero]

Originally Posted by lobosrul

4k has FOUR times (2x horizontal, and 2x vertical pixels) the resolution as 1080, not double. So even if h.265 only needs half the bit rate as h.264 (and I've heard its more like 30% than 50) the disc needs to be twice the size, given the same length of content.

Part of quote bolded by me for emphasis. The bolded statement is not true. Why is this?

Because of P and B frames.

I frames are a current snapshot of what a given scene looks like.

After any I frame, you will have one or more P frames.

Each P frame uses data from previous frames (either I, P, or B frames) to construct the current image.

Between P and I frames, you may also have B frames.

Each B frame uses data from previous and future frames to attain high levels of compression.


Because of P and B frames, 4x the resolution does not linearly equate to 4x the bitrate required, simply because when you build on frames that already exist, that's data you get "for free" (at the expense of decoding time/power requirements).

[lobosrul]

Originally Posted by Taikero

Part of quote bolded by me for emphasis. The bolded statement is not true. Why is this?

Because of P and B frames.

I frames are a current snapshot of what a given scene looks like.

After any I frame, you will have one or more P frames.

Each P frame uses data from previous frames (either I, P, or B frames) to construct the current image.

Between P and I frames, you may also have B frames.

Each B frame uses data from previous and future frames to attain high levels of compression.


Because of P and B frames, 4x the resolution does not linearly equate to 4x the bitrate required, simply because when you build on frames that already exist, that's data you get "for free" (at the expense of decoding time/power requirements).

You are correct that an exact linear increase is probably inaccurate. But, there are still differences between each frame (unless its a static image), and an image that has 4x as many pixels, will, on average I think, have 4x as much difference between frames.

[Anthony P]

Originally Posted by 5th Of November

If this only needs half the bit rate, then wouldn't it allow 4K enough space to fit on a 50 GB Blu-ray?

don't know what "half the bit rate means" but 4k is 4x the resolution.