x264 vs x265??

[xenotox]

Hello...

which are the best encodes... x264 with the maximum settings and 2 pass or the x265 with only 1 pass.

thanks in advance

[enim]

i wud rather say, it would be rather better if you tell us.

[El Heggunte]

It depends.

[Vidd]

I don't think x265 is mature enough to do a proper comparison. In theory, HEVC should offer better compression than AVC, but all the features of x265 have not been implemented yet.

[enim]

I don't think x265 is mature enough to do a proper comparison. In theory, HEVC should offer better compression than AVC, but all the features of x265 have not been implemented yet.

"In theory, there is no difference between theory and practice. But, in practice, there is."
-Yogi Berra

Just for an example,
In theory I do know what encoding is, In practice What is encoding?
I absolutely forgot it.

[deadrats]

Hello...

which are the best encodes... x264 with the maximum settings and 2 pass or the x265 with only 1 pass.

thanks in advance

what do you mean by "x264 with the maximum settings"?

if you mean x264 + placebo, that would mean with psy-rd, psy-trellis and mb-tree enabled, 3 things i absolutely despise and feel destroys the image.

x265 on the other hand i can't make up my mind, on some tests it looks like a winner and on some tests it looks like i need to kick all the developers in the balls.

i keep hoping someone creates a multithreaded vp9 encoder that will allow us to say F U to both x264 and x265.

[jagabo]

The real issue is that few devices outside of a computer will play h.265 encoded video. Whereas almost everything plays h.264 these days.

[enim]

The real issue is that few devices outside of a computer will play h.265 encoded video. Whereas almost everything plays h.264 these days.

ATM comparison is meaningless as one is fully matured while other is new born.

[racer-x]

ATM....x265 wins in stupid low bitrates, x264 wins at everything else. That's based on my countless tests so far. That could all change as x265 matures, assuming it becomes a useful standard.

[LigH.de]

It depends.

+1. There is no "best" until you tell us your personal formula to calculate "how good" it is. Quality is subjective.

[Atak_Snajpera]

ATM....x265 wins in stupid low bitrates, x264 wins at everything else. That's based on my countless tests so far. That could all change as x265 matures, assuming it becomes a useful standard.

I see the same results. Check this simple comparison
http://forum.doom9.org/showthread.php?p=1682910#post1682910

[Bjs]

Would't even compare ... the equation is biased

If your doing 2 pass for one then you must 2 pass for both using the same source for a fair comparison

[hello_hello]

Hello...

which are the best encodes... x264 with the maximum settings and 2 pass or the x265 with only 1 pass.

thanks in advance

I assume you'd be using the same bitrate each time?

The question is flawed because if you encode a video using a single pass quality based encoding method (CRF encoding) using x264, then make a note of the resulting bitrate and use it for a 2 pass encode with the same x264 settings, the video will be (for all practical purposes) encoded exactly the same way each time.

So you can't really compare x264 and x265 based one 2 passes vs 1 pass, as such.

[hello_hello]

if you mean x264 + placebo, that would mean with psy-rd, psy-trellis and mb-tree enabled, 3 things i absolutely despise and feel destroys the image.

Any chance you might post some "with and without" sample encodes at the same bitrate to demonstrate the image destroying effect to which you refer? Given you seem to be the only one complaining......

[Atak_Snajpera]

i keep hoping someone creates a multithreaded vp9 encoder that will allow us to say F U to both x264 and x265.

Why are you so aggressive towards those great codecs? vp9 specification is even weaker than AVC specs. If I recall correctly vp9 does not even support b-frames. Not to mention that encoding is practically impossible due to slooooooooooooooooooow encoding speed. Multithreading won't change much.

if you mean x264 + placebo, that would mean with psy-rd, psy-trellis and mb-tree enabled, 3 things i absolutely despise and feel destroys the image.

I would also like to see some examples. My brain prefers videos with psy-rd enabled. Without it I immediately see blurring.

[jman98]

i keep hoping someone creates a multithreaded vp9 encoder that will allow us to say F U to both x264 and x265.

Why are you so aggressive towards those great codecs?

He has personal issues with the developers of X.264 that have caused him to lose all objectivity. I saw this thread and his post and I immediately thought "Not again!" Please note that to him H.264 is very different from X.264 so you can't use the terms interchangeably because he doesn't use them that way. When he complains about X.264 or X.265 he means those projects specifically and not commercial implementations like H.264 and H.265.

[deadrats]

i keep hoping someone creates a multithreaded vp9 encoder that will allow us to say F U to both x264 and x265.

Why are you so aggressive towards those great codecs? vp9 specification is even weaker than AVC specs. If I recall correctly vp9 does not even support b-frames. Not to mention that encoding is practically impossible due to slooooooooooooooooooow encoding speed. Multithreading won't change much.

if you mean x264 + placebo, that would mean with psy-rd, psy-trellis and mb-tree enabled, 3 things i absolutely despise and feel destroys the image.

I would also like to see some examples. My brain prefers videos with psy-rd enabled. Without it I immediately see blurring.

first of all i can't stand DS, he spent more time spreading FUD about all competing products than he did developing his encoder and his supporters turned the use of x264 into a religion.

with regards to the psy optimizations, my feelings come down to this:

1) it's not the job of an encoder to "enhance" the image, the job of the encoder is to reproduce the source as closely as possible.

2) with some previously compressed sources, at stupid low bit rates, yes the 3 above mentioned optimizations do seem to be of some benefit BUT with high quality uncompressed sources, like the elephant's dream i used for my other tests, i find that they kill image quality.

re: vp9 - while it may not support b-frames that's by design, b-frames are a patented technology and any codec that supports them needs a license from the mpeg-la (i think that's the clowns that administer the rights of the mpeg family of codecs). it does use this other type of bi-directional frame which i can't remember what it's called and in my tests it produces some great video quality.

the so-called "slow" encoding speed is due to the total lack of threading and total lack of any simd optimizations, there no reason why it can't be optimized like any other codec. the best thing about vp9 is that it's patent free and youtube will be switching over to it come 2015, which means in a few years forums like this will be full of people looking for ways to edit vp9 and create vp9 videos for uploading to youtube.

one last thing, avc in general and x264 in particular is way overrated, it appeals to people that get off on seeing how little bit rate they can use, something i just can't wrap my head around. i can pick up a 3tb hard dive for about $150, by 2016 we will have 8tb and 16tb SSD's, a standard blu-ray can hold 25gb, a dual can hold twice that, i don't why people seem to enjoy taking a blu-ray movie and wasting hours to get it down to 2gb.

from everything i gathered, x264 really was a hit with the anime crowd, that fruity portion of the population that is into that sick Japanese demon rape cartoon porn crowd, that wanted to be able to share their sick little cartoons via p2p over their parents slow dsl connections.

those of us that were not into that garbage and could afford faster cable and fios internet connections and had an IQ higher than a kumquat never really found x264 all that appealing.

[deadrats]

He has personal issues with the developers of X.264 that have caused him to lose all objectivity. I saw this thread and his post and I immediately thought "Not again!" Please note that to him H.264 is very different from X.264 so you can't use the terms interchangeably because he doesn't use them that way. When he complains about X.264 or X.265 he means those projects specifically and not commercial implementations like H.264 and H.265.

x264 and x265 are very different than h264 and h265, the former are simply open source implementations of the latter.

[Gravitator]

Is not impressive in HEVC containing demos from Elecard ? > TOMSK ( 4K/3mbps)
And the problem with encoding "park_joy_1080p50.y4m" (p-frames / pyramid) speaks of a large reserve capacity
More HEVC a little better encodes image 64x64 proportions

[Atak_Snajpera]

which means in a few years forums like this will be full of people looking for ways to edit vp9 and create vp9 videos for uploading to youtube.

Youtube will always re-encode your uploaded footage. So it does not matter if it is encoded as vp9 or avc.

1) it's not the job of an encoder to "enhance" the image, the job of the encoder is to reproduce the source as closely as possible.

At some point (low bitrate) you have to "cheat" (fake details and so on) in order to get acceptable quality. Just think about HE-AAC@64kbps where half of frequency is cleverly reconstructed. It obviously does not sound like original but still sounds good especially if you do not have access to source

2) with some previously compressed sources, at stupid low bit rates, yes the 3 above mentioned optimizations do seem to be of some benefit BUT with high quality uncompressed sources, like the elephant's dream i used for my other tests, i find that they kill image quality.

You have freedom of choice. You can always disable those options if they "destroy" quality of source footage. No need to hate those optimizations.

first of all i can't stand DS, he spent more time spreading FUD about all competing products than he did developing his encoder and his supporters turned the use of x264 into a religion.

For me this is just an excellent tool. If you know about faster and with better compression efficiency encoder let me know.

[deadrats]

just to give you a sample of no-psy-rd vs psy-rd, i tested with the uncompressed 720p24 elephant's dream source, i used the latest Hybrid build as my gui front end and ran the tests with the following settings:

test 1: no-asm, no-mbtree, no-psy-rd, no-psy-trellis, no trellis, all psy disabled, 4mb/s, 3 b-frames, 3 reference frames, no deblock, scenecut 100, weighted b and p, optimal, automatic b, normal b pyramid, mixed refs, no fast skip, sub me 7, all partition types, dia

test 2: same as above, except psy-rd = 1

to my eyes test 1 looks closer to the source and cleaner overall, if one of you wants to download the source and the samples and compare yourselves, you're free to do it.

i also did an x265 test, here's the command line as copied from Hybrid:

x265 --preset fast --threads 8 --frame-threads 8 --input - --input-res 1280x720 --fps 24 --frames 15691 --me 0 --subme 0 --merange 16 --max-merge 1 --no-strong-intra-smoothing --keyint 100 --bframes 3 --weightb --b-adapt 0 --scenecut 100 --bitrate 4000 --rd 0 --no-signhide --aq-mode 2 --no-lft --no-psnr

i couldn't find an option to disable asm, so...

[hello_hello]

first of all i can't stand DS, he spent more time spreading FUD about all competing products than he did developing his encoder and his supporters turned the use of x264 into a religion.

I'm still not clear on the "which encoder is better than x264" thing...... If there is a better quality encoder, I suspect people would use it. Which AVC encoders are better than x264?

1) it's not the job of an encoder to "enhance" the image, the job of the encoder is to reproduce the source as closely as possible.

In a perfect world..... but the reality of it is, lossy encoding invariability involves compromises in order to compress efficiently, so if psy-rd helps distribute the bits in a manner which improves the perception of "quality", or encodes certain parts with less blurring, then it's not exactly doing something odd or unusual. Isn't that what the encoder tries to do anyway? Encode in the most efficient way? If you don't like psy-rd, you can always disable it.

2) with some previously compressed sources, at stupid low bit rates, yes the 3 above mentioned optimizations do seem to be of some benefit BUT with high quality uncompressed sources, like the elephant's dream i used for my other tests, i find that they kill image quality.

I'm keen to have a look, but they're taking forever to download.

re: vp9 - while it may not support b-frames that's by design, b-frames are a patented technology and any codec that supports them needs a license from the mpeg-la (i think that's the clowns that administer the rights of the mpeg family of codecs). it does use this other type of bi-directional frame which i can't remember what it's called and in my tests it produces some great video quality.

B-frames are patented technology? Are you sure you're not making that up?
vp9 uses B-frames which aren't really B-frames? If you remember what the non-b-frame b-frames are called I'd be keen to know.

one last thing, avc in general and x264 in particular is way overrated, it appeals to people that get off on seeing how little bit rate they can use, something i just can't wrap my head around. i can pick up a 3tb hard dive for about $150, by 2016 we will have 8tb and 16tb SSD's, a standard blu-ray can hold 25gb, a dual can hold twice that, i don't why people seem to enjoy taking a blu-ray movie and wasting hours to get it down to 2gb.

What a load of nonsense. There's lots of reasons for re-encoding, besides the "enjoyment" of reducing the file size. Try transferring a 25GB file to a thumb drive, or a phone, or a tablet, and if you do that often enough you'll probably want to reduce the file size too. Not to mention the fact I invariably apply some sort of filtering when I re-encode, even if it's just a bit of noise removal. That process does require the video to be re-encoded. And I prefer to crop so the video isn't stuck at 16:9 or 4:3. Unfortunately it can't be done without re-encoding because nothing supports the MKV crop flag thingy properly.

from everything i gathered, x264 really was a hit with the anime crowd, that fruity portion of the population that is into that sick Japanese demon rape cartoon porn crowd, that wanted to be able to share their sick little cartoons via p2p over their parents slow dsl connections.
those of us that were not into that garbage and could afford faster cable and fios internet connections and had an IQ higher than a kumquat never really found x264 all that appealing.

Given x264 is way better than Xvid, and (in my opinion) the best AVC encoder, I find it quite appealing. I don't watch/encode anime, and while I'm no genius, I have a higher than average IQ. Would you mind not speaking for the rest of us while you're exercising your x264 non-objectivity?

[hello_hello]

just to give you a sample of no-psy-rd vs psy-rd, i tested with the uncompressed 720p24 elephant's dream source, i used the latest Hybrid build as my gui front end and ran the tests with the following settings:

test 1: no-asm, no-mbtree, no-psy-rd, no-psy-trellis, no trellis, all psy disabled, 4mb/s, 3 b-frames, 3 reference frames, no deblock, scenecut 100, weighted b and p, optimal, automatic b, normal b pyramid, mixed refs, no fast skip, sub me 7, all partition types, dia

test 2: same as above, except psy-rd = 1

to my eyes test 1 looks closer to the source and cleaner overall, if one of you wants to download the source and the samples and compare yourselves, you're free to do it.

i also did an x265 test, here's the command line as copied from Hybrid:

x265 --preset fast --threads 8 --frame-threads 8 --input - --input-res 1280x720 --fps 24 --frames 15691 --me 0 --subme 0 --merange 16 --max-merge 1 --no-strong-intra-smoothing --keyint 100 --bframes 3 --weightb --b-adapt 0 --scenecut 100 --bitrate 4000 --rd 0 --no-signhide --aq-mode 2 --no-lft --no-psnr

i couldn't find an option to disable asm, so...

That was disappointing...... and after your long rant on shrinking movies down tp 2GB, you run sample encodes using constant bitrate encoding......
Hands up anyone who normally encodes using constant bitrate encoding? Anyone? I don't, so the effect, or lack thereof, of psy-rd when running constant bitrate encoding is fairly irrelevant to me. Now if CRF encoding had been used...... chances are one encode would even have a slightly higher bitrate than the other, so you could probably tell how much effect psy-rd was having on sharpness.

I haven't looked at the samples closely yet, I'm out of time today and I'll need to wait until my disappointment has decreased a little..... probably tomorrow.

[deadrats]

B-frames are patented technology? Are you sure you're not making that up?
vp9 uses B-frames which aren't really B-frames? If you remember what the non-b-frame b-frames are called I'd be keen to know.

i first found out about it from his darkness himself:

http://x264dev.multimedia.cx/archives/377

The lack of B-frames in VP8 is a killer. B-frames can give 10-20% (or more) compression benefit for minimal speed cost; their omission in VP8 probably costs more compression than all other problems noted in this post combined. This was not unexpected, however; On2 has never used B-frames in any of their video formats. They also likely present serious patent problems, which probably explains their omission. Lack of weighted prediction is also going to hurt a bit, especially in fades.

this also says that it's patented technology:

http://en.wikipedia.org/wiki/H.264/MPEG-4_AVC

as for what the non-b b-frames are called they are referred to as "compound prediction frames":

https://forum.doom9.org/showthread.php?p=1679627

However, VP9 does support “compound prediction”, which really is just another word for bi-prediction where there are two motion vectors for each block and the two resulting prediction samples are averaged together. In order to avoid patents on bi-prediction, compound prediction is only enabled in frames that are marked as not-displayable. A frame like this is never output for display, but may be used for reference later. In fact, a later frame may consist of nothing but 64x64 blocks with no residuals and 0,0 motion vectors that point to this non-displayed frame, effectively causing it to be output later using very little data.

vp9 also has what's known as a golden frame and alt ref frames.

from everything i have read vp9 is designed to be free in every sense of the word, free of patents, free to use, open source, etc.

of course this is America, where anyone can file a lawsuit for practically anything, so i wouldn't be surprised if the mpeg-la sues google at some point inthe future over a supposed patent infringement.

[Selur]

no clue why this thread now is about vp9, but http://en.swpat.org/wiki/WebM,_VP8_and_VP9 might be interesting in what happened so far regarding patents inside VP8&VP9

[deadrats]

Now if CRF encoding had been used...... chances are one encode would even have a slightly higher bitrate than the other, so you could probably tell how much effect psy-rd was having on sharpness.

fine, i redid the tests with crf 22, test 1 is as before and test 2 is with psy-rd=1.

[hello_hello]

I can't say I've compared every single frame of each encode, but they're obviously encoded slightly differently in places. Number 2 appears to have a little more fine detail at times, and the slightly higher bitrate seems to indicate that's the case. Which one was encoded more accurately compared to the original.... well I don't have the original to compare them to. If there's "image destroying" differences between them, I think you'll need to point them out to me. I don't appear to be clever enough to find them on my lonesome.

Mind you I think you moved the goalposts a bit. You originally declared your hatred for psy-rd, psy-trellis and mb-tree, but not your hatred for psy-rd, psy-trellis and mb-tree with qcomp reduced a little from the default setting and AQ in auto variance mode, and you left mb-tree disabled for both encodes.

Anyway..... I assume next you'll be uploading another encode of the same video at the same bitrate using a different h264 encoder? One of the encoders which do a better job of encoding the video "accurately" and therefore justify your dislike of x264 and it's psy enhancements?

[Atak_Snajpera]

For more common film footage psy really does great job

Uncompressed frame
https://forum.videohelp.com/images/imgfiles/MU2RNcd.png


x264 --crf 22 --preset veryslow (average bitrate 886 kbps)
https://forum.videohelp.com/images/imgfiles/gTqmkjJ.png

x264 --crf 20 --preset veryslow --no-psy (average bitrate 903 kbps)
https://forum.videohelp.com/images/imgfiles/6DTT23H.png



x265 --crf 20 --preset medium (average bitrate 992 kbps)
https://forum.videohelp.com/images/imgfiles/TT9h3Ok.png

x265 --crf 20 --preset medium --rd 6 --psy-rd 1.0 (average bitrate 1060 kbps)
https://forum.videohelp.com/images/imgfiles/42gksNo.png

[deadrats]

@atak:

WHAT?!? "great job"? in what way? all the screenshots are too dark to see any real details, the bit rate of the encodes is absurdly low and with one of the test encodes psy-rd used more bit rate than without.

this is the type of silly BS that the x264 faithful use to justify their idiotic beliefs, seriously who the F actually encodes their video with 900 kb/s? i can barely stand audio with 640 kb/s and you want to encode video with barely more?

[LigH.de]

The more the developers ask not to test psy-rd but other options until it is known to be fixed, the more people test psy-rd ...


@ deadrats: You are getting personal and insulting.