Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > High Efficiency Video Coding (HEVC)

Reply
 
Thread Tools Search this Thread Display Modes
Old 10th October 2017, 00:03   #1  |  Link
x265_Project
Registered User
 
Join Date: Jul 2013
Posts: 596
HEVC wins an Emmy Award!

https://www.microsoft.com/en-us/rese...definition-tv/
x265_Project is offline   Reply With Quote
Old 10th October 2017, 14:20   #2  |  Link
IgorC
Registered User
 
Join Date: Apr 2004
Posts: 1,318
Interesting.
It's a last thing that Microsoft can do for HEVC because now they support AV1 which comes soon.
IgorC is offline   Reply With Quote
Old 12th October 2017, 02:12   #3  |  Link
x265_Project
Registered User
 
Join Date: Jul 2013
Posts: 596
Quote:
Originally Posted by IgorC View Post
Interesting.
It's a last thing that Microsoft can do for HEVC because now they support AV1 which comes soon.
Igor... Dude... the latest updates from the AOM report that it's 200x more complex than VP9, and while the AOM thinks it's 25% more efficient than HEVC (against x265 v1.9 measured with psnr without --tune-psnr), independent tests (experts at some of the world's biggest video services and solution providers) show that it's barely able to show any measurable improvement over HEVC. I'm not saying it's never going to succeed, but clearly AV1 has a long way to go before it's a competitive threat to HEVC which was standardized almost 5 years ago, and is now built into every new TV, set-top box, premium smartphone, tablet, PC, and GoPro video camera (not to mention security cameras, photo compression, etc.). Building and winning market share for a new video coding standard is a REALLY HARD THING, that takes many years and billions of dollars of investment by hundreds of companies. It's hard to fathom the amount of investment in HEVC implementations worldwide by thousands of companies. If you're a fan of AV1, you'll need to be very patient. In the meantime, HEVC is proliferating much faster than we've seen in past years. Apple was the big catalyst. GoPro is another. More dominoes will fall.

Last edited by x265_Project; 12th October 2017 at 22:01.
x265_Project is offline   Reply With Quote
Old 12th October 2017, 08:29   #4  |  Link
iwod
Registered User
 
Join Date: Apr 2002
Posts: 757
Quote:
Originally Posted by x265_Project View Post
Igor... Dude... the latest updates from the AOM report that it's 200x more complex than VP9, and while the AOM thinks it's 25% more efficient than HEVC, independent tests (experts at some of the world's biggest video services and solution providers) show that it's barely able to show any measurable improvement over HEVC. I'm not saying it's never going to succeed, but clearly AV1 has a long way to go before it's a competitive threat to HEVC which was standardized almost 5 years ago, and is now built into every new TV, set-top box, premium smartphone, tablet, PC, and GoPro video camera (not to mention security cameras, photo compression, etc.). Building and winning market share for a new video coding standard is a REALLY HARD THING, that takes many years and billions of dollars of investment by hundreds of companies. It's hard to fathom the amount of investment in HEVC implementations worldwide by thousands of companies. If you're a fan of AV1, you'll need to be very patient. In the meantime, HEVC is proliferating much faster than we've seen in past years. Apple was the big catalyst. GoPro is another. More dominoes will fall.
This. To me it isn't really AOM, it happens all the way back at On2. And as many who said Google is just a bunch of small company doing things their own way, very likely On2 gets to carry on within Google.

After all we are on Doom9, which started to get famous around the Web for Video Encoding because of Ripping DVD for Backup and its subjective comparison test between different codec like Divx, Xvid and many others. We knew early on, ( which is 15+ years ago) PSNR and SSIM wont good enough. Hence why Doom9 did all the long, hard, resource intensive Subjective Comparison, similar to Hydrogenaudio at the time.

And yet after all this year it is still the same

VP6 / 7 is similar or better then H.263
VP8 is similar or better then H.264
VP9 is similar or better then H.265
VP10 / AV1 will exceed H.265

Truth to be told I really wanted to like AV1. But it is late, I am hoping AV2 will finally shake things up in this ridiculous patents fees cartel.
iwod is offline   Reply With Quote
Old 15th October 2017, 11:19   #5  |  Link
MoSal
Registered User
 
Join Date: Jun 2013
Posts: 89
@iwod

You are both affirming the consequent:
* On2 was developed by VP8. It wasn't as good as claimed (initially).
* AV1 is developed by On2. Therefore, it's not as good as claimed.

And making a false equivalency. As AV1 is not solely developed and marketed by Google/On2.

And I find it quite ironic that you're agreeing with the mind projections and cherry-picked
selective truths of HEVC PR, while promoting skepticism and emphasizing the importance of subjective evaluations, appealing to the traditions of the mighty Doom9
community.
__________________
saldl: a command-line downloader optimized for speed and early preview.
MoSal is offline   Reply With Quote
Old 18th October 2017, 18:30   #6  |  Link
Sagittaire
Testeur de codecs
 
Sagittaire's Avatar
 
Join Date: May 2003
Location: France
Posts: 2,488
Quote:
VP8 is similar or better then H.264
No for objective and subjective comparison ... and by far.

Quote:
VP9 is similar or better then H.265
No for objective and subjective comparison ... and by far.
__________________
Le Sagittaire ... ;-)

1- Ateme AVC or x264
2- VP7 or RV10 only for anime
3- XviD, DivX or WMV9
Sagittaire is offline   Reply With Quote
Old 20th October 2017, 19:03   #7  |  Link
benwaggoner
Moderator
 
Join Date: Jan 2006
Location: Portland, OR
Posts: 3,390
Quote:
Originally Posted by iwod View Post
VP6 / 7 is similar or better then H.263
VP8 is similar or better then H.264
VP9 is similar or better then H.265
VP10 / AV1 will exceed H.265
Yeah, so I've been hearing for years. And I have tried, oh so many times, to try to get VP8 to match x264, or VP9 to match x265. VPx never was as good even. And the encoders were much less flexible, and much slower on multicore systems. Parallelization from horizontal slices? Please. H.264/HEVC encoders all nailed single-slice parallelism within a few years after standardization.

And the decoders were so single-thread bound. And no B-frames, so no ability to skip non-reference frames for fast random access. No Profile @ Level definition, so it was impossible to know how far settings could be pushed to work on generic players.

Of course, we also only had one real encoder implementation for those codecs. the H.26x codecs benefited enormously from furious competition between multiple vendors. We don't know how good a VP8/9 encoder could have been, since it was really tuned by and for one major customer.

Quote:
Truth to be told I really wanted to like AV1. But it is late, I am hoping AV2 will finally shake things up in this ridiculous patents fees cartel.
The good thing about AV1 is that it had a whole lot of other experts involved, it'll be easier to implement for HW, and maps cleanly into MPEG file formats with useful SEI and header metadata. Lots of eyeballs and experts help shake out all those small little things that in aggregate matter hugely. Tons of things that were needless hard blockers for VPx in commercial applications have been fixed.

And, if the final version and all its tools look reasonably competitive (which is far from clear yet), it will hopefully get multiple implementations competing against each other.

The critical thing to remember is that it isn't a bitstream competition, but an encoder/decoder competition. Real-world AV1 encoders need to have a material advantage over real-world HEVC encoders at the time of comparison. And there need to be deployed decoders, or predicted deployment of decoders, to make it a viable target.

And this is where HEVC's five year head start can really matter. There are a bunch of existing implementations, and we're seen huge advances year-on-year in perceptual quality and performance. AV1's future commercial-grade encoders will compete against HEVC's future commercial-grade encoders. Which will be a lot better than they are now, which is already a lot better than x265 1.9 default settings. And VP6+ were always super-tuned around PSNR; On2 decided that PSNR was what drove decisions, so they worked hard to win that metric. x26? never cared about PSNR, and so has a significantly better perceptual quality at a given PSNR than any VPx encoder to date. The "25% better" comparison is entirely meaningless for market decisions.

I think AV1 could be successful if either
  1. It has commercial encoders that deliver at least 25% lower bitrate than the best HEVC encoders at the same perceptual quality
  2. There is big addressable market of devices with AV1 but not HEVC, and AV1 offers at least a 30% bitrate reduction versus H.264

The first is theoretically theoretically possible (it is theoretically possible that the bitstream syntax would offer a theoretical encoder that advantage), but we won't know for 1+ years I think. Getting a well-tuned encoder to market is a big deal.

The second is hard to predict. Certainly I can't point to many devices I know will have AV1 and not HEVC. All the future SoC designs that could support AV1 already have HEVC.

On net, the reason that the MPEG codecs keep winning despite the patent fees is that they are economically worth it. The advantages in compatibility, integratability, available encoders, and compression efficiency make using them and paying the patent fees a better deal than using other available codecs.

AV1 promises to address some of those issues, particularly integration, better than past royalty-free alternatives. It should be capable of at least roughly comparable compression efficiency in a mature encoder. But is it enough better to justify an aggregate billions of dollars in infrastructure replacement? It is quite possible any advantage to AV1 will only become clear after H.266 is on the horizon.

What would be cool is if H.266 and AV2 wind up being the same thing, just like H.265 and HEVC are the same thing...
__________________
Ben Waggoner
Principal Video Specialist, Amazon Prime Video

My Compression Book
benwaggoner is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 19:41.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.