Next-Gen Video Encoding: x265 Tackles HEVC/H.265

Status
Not open for further replies.

Jindrich Makovicka

Honorable
Jul 23, 2013
1
0
10,510
In addition to PSNR comparison, I'd be much more interested in the SSIM metric, which is better suited for codecs using psychovisual optimizations.

PSNR can be usable for when testing varying parameters for one codec, but not so much when comparing two completely different codecs.
 
Nice intro to the new codec!
And to think that this is unoptomized... Once this is finalized it will really blow 264 out of the water and open new doors for 4K content streaming, or 1080p streaming with much better detail and contrast. This is especially important with the jump to 4K video. The 16x16 grouping limit on x264 is great for 1080p, but with 4K and 8K coming down the pipe in the industry we need something better. The issue is that we really do not have many more objects on the screen as we did back in the days of 480i video, it is merely that each object is more detailed. Funny thing is that a given object will typically have more homogeneous data across its surface area, and when you jump form 1080p to 4K (or 8K as is being done for movies) then it takes a lot more 16x16 groupings which may all relay the same information if it is describing a large simple object. Moving up to 64x64 alone allows for 8K groupings that take up the same percentage of the screen as 16x16 groupings do in 1080p.
 

nibir2011

Distinguished
Aug 28, 2011
131
0
18,680
Considering the CPU Load i think it wont be a viable solution for almost any home user within next 2-3 years unless CPUs gets exceptionally fast.

Of course then we have the Quantum Computer. ;)
 

Shawna593767

Honorable
Feb 26, 2012
127
0
10,690
Quantum computers aren't fast enough for this, the get their speed by doing less calculations.
For instance a faster per clock x86 computer might have to do say 10 million calculations to find something, whereas the quantum computer is slower per clock but would only need 100,000 calculations.
 

nibir2011

Distinguished
Aug 28, 2011
131
0
18,680



well a practical quantum computer does not exist . lol

i think that is not the case with calculation.i think what you mean is accuracy. number of calculation wont be different; it will be how many times same calculations need to be done. in theory a quantum computer[whatever qubits] should be able to make perfect calculations as it can get all the possible results by parallelism of bits[long stuff]. a normal cpu cant do that it has to evaluate each results separately. SO a quantum computer is very very efficient than any traditional cpu. Speed is different it depends on both algorithm and architecture. quantum algorithms is at its infancy. last year maybe a quantum algorithm for finding out primes was theorized. I do not know if we will see a quantum computer capable of doing what the regular computers do next 30 years.

thanks

 

InvalidError

Titan
Moderator
Most of the 10bit HDR files I have seen seem to be smaller than their 8bits encodes for a given quality. I'm guessing this is due to lower quantization error - less bandwidth wasted on fixing color and other cumulative errors and noises over time.
 

ddpruitt

Honorable
Jun 4, 2012
1,109
0
11,360
I know it's a minor detail but it's important:

H.264 and H.265 are NOT encoding standards, they are DECODING standards. The standards don't care how the video is encoded just how it's decoded, I think it should be made clear because the article implies they are decoding standards and people incorrectly assume one implies the other. x264 and x265 are just open source encoders that encode to formats that can be decoded properly by H.264 and H.265.

x264 has noticeable issues with blacks, they tend to come out grey. I would like to see if x265 resolves the problem. I would also like to see benchmarks on the decoding end (CPU Load, power usage, etc) as I see this becoming an issue in the future with streaming video on mobile devices and laptops.
 

Nintendo Maniac 64

Distinguished
Jan 24, 2012
73
0
18,630
You guys should really include VP9 in here as well, since unlike VP8 it's actually competitive according to the most recent testing done on the Doom9 forums, though apparently the reference encoder's 2-pass mode is uber slow.
 

InvalidError

Titan
Moderator

h264/h265 are methods of storing compressed video.

While the exact method of encoding and decoding it are at the individual algorithm developers' sole discretion, the structures and core algorithms related to how information ultimately needs to be structured dictate a fair chunk of how BOTH the encode and decode algorithms need to work.

Most of the decode steps are almost exact inverse transforms of their corresponding encode steps. That's why you have reference encoders and decoders to prove that every encoding step can be reversed by its corresponding decoding step. Saying that encode is unrelated to decode is very naive; they are very closely related - at least in reference implementations.

Some programmers may find shortcuts through the reference designs or ways to combine multiple steps into one or find other ways to achieve the same result for a given step or group of steps but the overall encode and decode algorithms usually retain the same general flow.
 

ElMoIsEviL

Distinguished


Quantum Computers do less calculations by virtue of being able to calculate every possible answer simultaneously (as the amount of Qbits rises the amount of possible solutions entertained in a Quantum State rises exponentially). Seems to me that this would be perfect when it comes to the concept of having to do multiple passes in terms of video encoding. You'd only have to do a single pass and during that single pass any possible outcome in terms of frame IQ can be processed simultaneously ensuring that each frame kept is the perfect frame free of errors.

It would be quite super.
 

nevilence

Distinguished
May 8, 2012
339
0
18,810
 

jkflipflop98

Distinguished
Basic quantum theory is very difficult to grasp, and understanding the operation of a quantum computer is beyond most humans. There are only a handful of people on this planet that can knowingly speak on the subject, and I'll almost guarantee none of them visit this site.
 

nevilence

Distinguished
May 8, 2012
339
0
18,810


Its only difficult if I observe myself trying to grasp it, wait does that make me a wave or a partical
 

Achoo22

Distinguished
Aug 23, 2011
350
2
18,780
Basic quantum theory is very difficult to grasp, and understanding the operation of a quantum computer is beyond most humans. There are only a handful of people on this planet that can knowingly speak on the subject, and I'll almost guarantee none of them visit this site.

So? Just as there are plenty of drivers that aren't very knowledgeable about internal combustion engines, there are plenty of expert programmers that have little knowledge of current computer organization at its lowest levels.
 

Sagittaire

Honorable
Jul 24, 2013
3
0
10,510
Well I know very well x264 and I make actually test H264 vs H265. Your test have problem:
- If you make psnr test, you must use best possible setting for x264 and use --tune-psnr
- you use quantizer mode for x264 and x265. Anyway x264 have really most advanced rate control with crf mode.
- you don't use the best possible x264 version. x264 10 bits produce better result (at least 0.2 dB for same bitrate).

If you combine all these tweak, I can say that x264 with best setting produce certainely same PSNR result than x265. Moreover H265 is superior to H264, no doubt about that. But I can say, when I read your article, than x265 is not really superior to x264 at this time (for PSNR result at least).
 

ElMoIsEviL

Distinguished
 

nevilence

Distinguished
May 8, 2012
339
0
18,810
The figures you are talking about are just the number of classical bits represented by qbits. Yes 16 qbits may represent what 65k+ classical bits can, but that doesnt mean it examines all answers simultaneously. The act of measuring the spin up or down of the qbit, gives a very clearly defined state, only before observation are there the possiblity of "ulimited" states.

Well that is at least my understanding, mind you I aint no quantum master either. The youtube link is my source, its a very good watch.

http://www.youtube.com/watch?v=g_IaVepNDT4
 

Filiprino

Distinguished
Dec 30, 2008
160
0
18,680
x265 is still unoptimized. If the performance improves and with the same bitrate as x264 you can get rid of banding on obscure scenes enabling higher contrast, it's a welcomed codec. I still do not have a CPU that can do x264 1080p real time and you come with this codec eating as much as 10x more cycles. I hope performance improves a lot :(
 

InvalidError

Titan
Moderator

If your PC cannot decode 1080p in real-time, you either have something ancient or have software issues - even my old Core2Duo E8400 (3GHz) could play 3-4 1080p/x264 files at the same time while an even more ancient dual-core P4 would likely manage to play at least one file. (My single-core Northwood P4 with HT could just barely play 720p files, as did my laptop's single-core Athlon XP 3000... those are ~9 years old CPUs.)

While the encoding effort may have increased by over 10X, the decode effort likely increased by a much smaller amount: encoding needs to find movement and solve optimization problems which involve lots of work to solve multiple possible decision trees to find the optimal solution with all other work getting discarded while playback is entirely deterministic; simply playing back the best solution the encoder found and recorded.

I would expect just about anything i3 and up to be able to handle x265/1080p decoding fine once the code is mature and properly optimized.
 

nibir2011

Distinguished
Aug 28, 2011
131
0
18,680


i assume you know a bit programming
in c language a double has double the precision than a float. because it evaluates the value more times than a float. video encoding is similar in terms of calculation to exactly know what is the thing really we need to be closer to the exact thing that is why we pass it more times.we need to generate different answers then we can evaluate it under our algorithm.but in a quantum computer we should be able to get all the answers at the same time.it is more like you have wall full of possible answers and choose the correct one.when it is a binary computer we can get only one result but in a quantum computer you must not forget about the 0/1 state.this should give us a fair amount of knowledge what the next answer should be.to understand this really you must have a little knowledge about flip flops[dld] and how cpu works in core level. all these things are difficult to grasp because the quantum algorithms are not yet present. but one thing is for sure quantum computer will be very efficient.i am sorry if i could be a help.i do not study in the subject right now but want to in my MS course.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
For the small increase in quality or what seems to be 20-25% less file size (?) for the same quality this just isn't worth multiple times more electricity used to get the work done. A bluray can hold ~17 episodes that are around 1.35GB (42min avg tv shows). So all this extra work will let me fit maybe a few more on a disc? Waste of time. I'll take x264 with a cuda card behind it and call it a day. This may be useful for streamers like netflix etc as they are streaming millions of shows, but not for a guy like me. I would have to be able to encode at better quality in the same time, or equal quality in less time for this to matter to me. Otherwise it just raises my electric bill and ties up my PC longer for so little in return.
 
Status
Not open for further replies.