x264 vs x265??

[deadrats]

seriously, if as has been reported by 2016 we have 16tb ssd's and they cost in the $500 range

That is a fantasy.

assume it's not fantasy or assume that the 8tb versions are in the $500 range, would you seriously consider upgrading a your cpu/mb/ram or would you just buy a new ssd and call it a day?

[hello_hello]

Is "philosophical" a nice way of saying "not completely rational"?

what's "irrational" is starting out with a high quality 20-30 mb/s source, cropping and resizing the crap out of it, dropping the bit rate to stupid low levels, i've seen encodes of Ghost Rider at 720p@1mb/s which was unwatchable, and thinking that it's ok because you enabled psy-rd, mb-tree, aq or whatever other "optimization" in the most over-rated half-assed encoder ever conceived.

You're just ranting irrationally again. You've made claim after negative claim regarding x264 in numerous threads and not once managed to provide a single piece of evidence to back them up. Not a single comparison encode. You haven't even been able to provide an example where psy "destroyed the image" as you claimed it does.

Any encoder will output crap if the bitrate is too low. Any encoder will output quality if the bitrate is high enough. Your argument seems to be solely based on the fact that because x264 does better than other AVC encoders at low bitrates and because people use it at lower bitrates, that somehow makes it a bad encoder. It's nonsense. I've got plenty of encodes here, filtered, resized or otherwise, which look better than the source at lower bitrates, so your argument is just silly generalisation and nothing to do with the quality of the encoder.

You posted a couple of CRF22 encodes of Elephants Dream, but you don't seem to be able to offer an encode of the same video at the same bitrate using another encoder which matches either of the x264 encodes for quality, but somehow that makes x264 over-rated. Pfffftttt.....

And there's lots of reasons for re-encoding which don't revolve solely around disc storage space. Try regularly transferring 30GB files to your ipad.

[deadrats]

@gonca

thanks, it seems all the encoders undershot the target bit rate a bit, i personally have trouble seeing any real differences between any of the encodes, in some sequences it seems to me x264 produced the better image, in some it seems DivX265 and in other x265.

i would like to see what a pro caliber hardware encoder could do.

btw, you did use placebo for x264 and x265 and aq 5 for divx, right?

what kind of encoding speeds did you see with x264 and x265 with placebo?

[deadrats]

It's nonsense. I've got plenty of encodes here, filtered, resized or otherwise, which look better than the source at lower bitrates, so your argument is just silly generalization and nothing to do with the quality of the encoder.

And there's lots of reasons for re-encoding which don't revolve solely around disc storage space. Try regularly transferring 30GB files to your ipad.

first let me laugh at you for owing an ipad and then for actually watching movies on it, i'm sure that makes you the coolest kid in your parent's basement.

as for the first statement, i have to congratulate you, in one sentence you managed to give proof to support what i've been saying all along, that x264 users have a reality distortion field of enormous strength turned on at all times.

you have "plenty" of encodes that have been filtered, resized or "otherwise" (i'm not familiar with this video editing process) and they look better than the source at lower bit rates and of course that's thanks to x264 and not the filters you used, right?

any chance of providing a before and after sample, a clip from the source and then after you got through molesting it, i mean improving it while using lower bit rate.

i think we would all like to see a sample of your twerk.

[gonca]

@gonca

thanks, it seems all the encoders undershot the target bit rate a bit, i personally have trouble seeing any real differences between any of the encodes, in some sequences it seems to me x264 produced the better image, in some it seems DivX265 and in other x265.

i would like to see what a pro caliber hardware encoder could do.

btw, you did use placebo for x264 and x265 and aq 5 for divx, right?

what kind of encoding speeds did you see with x264 and x265 with placebo?

All settings as before except for the bit rate. Placebo and aq5 were used.

Re speeds: with apologies to the other members who use this saying --- OY VEY!!!
Translated that means about 8 to 9 fps for x264 and 0.3 fps for x265.
X265 was an all nighter. Divx265 was a little faster than x265 but it still took 3 to 4 hours for a 10 minute sample

[deadrats]

Translated that means about 8 to 9 fps for x264 and 0.3 fps for x265.
X265 was an all nighter. Divx265 was a little faster than x265 but it still took 3 to 4 hours for a 10 minute sample

damn, and that's with an overclocked intel hexa core? LOL.

just to extrapolate, in order to get real time performance with x264+placebo you would need 3 overclocked i7 3930k's and to get real time with x265+placebo you would need 80 overclocked i7 3930k's.

let's see, at $500 per processor, for real time 24fps x265 encoding with a lossless source to 720p@4mb/s at the slowest settings you would have to spend a measly $40,000 on processors alone, but since the i7's are single processor only you need to get xeons if you want 4 cpu's per motherboard, plus the cost of the motherboards (we would need 20) plus the cost of the ram, plus the cost of 20 windows licenses, hard drives, power supplies, so basically we would need a $100,000 beowulf cluster.

one can't help but wonder if someone decided to take the reference encoder and get it to run on a Tesla K40 which only costs 5 grand, if perhaps that would be a more cost effective solution.

oh, i'm sorry, i forgot "it can't be done"™ because blah, blah <insert BS here> blah blah <pretend that Elemental hasn't already done this> blah blah <more BS about "no demand from corporate sponsors"™>.

sorry, sorry i forgot.

[hello_hello]

It's nonsense. I've got plenty of encodes here, filtered, resized or otherwise, which look better than the source at lower bitrates, so your argument is just silly generalization and nothing to do with the quality of the encoder.

And there's lots of reasons for re-encoding which don't revolve solely around disc storage space. Try regularly transferring 30GB files to your ipad.

first let me laugh at you for owing an ipad and then for actually watching movies on it, i'm sure that makes you the coolest kid in your parent's basement.

I don't own an ipad. It was just an example. Although I wouldn't expect someone who can't determine the aspect ratio of a video correctly, or who thinks 45 minutes to encode 10 minutes of 720p video using x264's default settings is normal, to understand that.
Let me laugh at you for living in the dark ages and not realising most portable devices capable of playing video these days are equipped with an exciting new output called HDMI.

as for the first statement, i have to congratulate you, in one sentence you managed to give proof to support what i've been saying all along, that x264 users have a reality distortion field of enormous strength turned on at all times.

Even if that wasn't a complete load of twaddle, it still wouldn't make x264 a bad encoder, but it does make for another of your irrational generalisations.

you have "plenty" of encodes that have been filtered, resized or "otherwise" (i'm not familiar with this video editing process) and they look better than the source at lower bit rates and of course that's thanks to x264 and not the filters you used, right?

Well of course they look better because of the filtering. Have you ever encoded a video yourself? Why do you ask such silly questions?
The lower bitrates are often due to the encoded video being "cleaner", but also due to the x264 encoder compressing more efficiently.

any chance of providing a before and after sample, a clip from the source and then after you got through molesting it, i mean improving it while using lower bit rate.

Have you ever encoded a video before? Are you at least aware that if you remove noise from a video it's easier to compress and therefore less bitrate will be required even if the same encoder is used? Do you really require samples of the obvious?

[deadrats]

Have you ever encoded a video before? Are you at least aware that if you remove noise from a video it's easier to compress and therefore less bitrate will be required even if the same encoder is used? Do you really require samples of the obvious?

i have about a half hour free and can think of nothing more mind numbing than seeing samples of your "work". who knows maybe someone will happen to be visiting this site from a major movie studio, see the "incredible" twerk you do, "improving" video and offer you a kick in the nuts.

[DarrellS]

Elephant's Dream at 720p isn't that high of quality. I was using Elephant's Dream at 1080p and got hammered because my x265 encodes didn't have near the bitrate that the x264 encodes had. This doesn't make much since to me since the advantage of x265 is that it can create files of the same quality at much lower file size (lower bitrate). I could set the crf as low as I could or tell x265 to encode 50Mbps files and it would still create low bitrate files. If everyone wants encodes at the same bitrate and same file size then why are we even using x265 since x264 can create the same bitrate files at much faster encoding speeds.

My Q6600 at stock speed encodes 1080p hevc using x265 ultrafast at 15 fps. X264 is five times faster but the files are five times bigger. 3840x2160 drops to about 4.3 fps (figured it would be 7.5 fps but) and at medium preset is .something or .0something (or not worth the time to even test it).

Until x265 is able to encode at least half as fast as x264 at half the bitrate then it will not be a serious contender against x264. Kinda reminds me of VC1 against DivX or XviD. VC1 could compress better but it took forever to encode and you needed a decent graphics card to play it. x264 was better at compressing than DivX or XviD but again, x264 was not a serious contender until Intel introduced the Core 2 Quad and people started using better graphics cards.

Contrary to what at least one person on this site believes, x264 is king and it's going to take a whole lot more effort from it's challengers to kill the king.

PS. DivX265 looked promising and although it's faster than x265, it still needs to be faster to compete with x264. My WDTV plays high resolution x264.mkv (3840x2160) that I've been creating for years,

[hello_hello]

i have about a half hour free and can think of nothing more mind numbing than seeing samples of your "work".

And once again your contribution to the topic has degenerated into mindless prattle.

[deadrats]

Elephant's Dream at 720p isn't that high of quality.

the uncompressed Elephant's Dream, the download is 4gb in a zip archive and when extracted it's 21 gb, it's raw video, never before compressed, my guess is you downloaded one of the 720p already compressed versions.

the 1080p uncompressed file is 46 gb in a zip archive and hundreds of gigs when extracted, i doubt you tested with these sources.

[deadrats]

My Q6600 at stock speed encodes 1080p hevc using x265 ultrafast at 15 fps. X264 is five times faster but the files are five times bigger. 3840x2160 drops to about 4.3 fps (figured it would be 7.5 fps but) and at medium preset is .something or .0something (or not worth the time to even test it).

you are so full of it, it's not even funny.

you have a stock speed Q6600 that encodes 1080p hevc at 15fps and x264 at 75 fps at 1080p? at what bit rate, 500 kb/s?

even better, your Q6600 3840x2160 you encode x264 at 4.3 fps and "at medium preset is .something or .0something".

hell, maybe i should downgrade my pc because i really want to be able to encode at "something or .0something" speeds.

really, the thought that a Q6600 is going to encode 4k content at 4.3 fps, maybe with a 128 kb/s bit rate for source and target.

[DarrellS]

Elephant's Dream at 720p isn't that high of quality.

the 1080p uncompressed file is 46 gb in a zip archive and hundreds of gigs when extracted, i doubt you tested with these sources.

Yes, this is the file that I downloaded and it took all night to download (maybe a couple of night, I don't remember). Not sure I still have it since I cleaned most of my HEVC stuff out of my PC when the latest x265 came out. I'll look when I get home from my radiation treatment. Gotta go.

[hello_hello]

My Q6600 at stock speed encodes 1080p hevc using x265 ultrafast at 15 fps. X264 is five times faster but the files are five times bigger. 3840x2160 drops to about 4.3 fps (figured it would be 7.5 fps but) and at medium preset is .something or .0something (or not worth the time to even test it).

you are so full of it, it's not even funny.

you have a stock speed Q6600 that encodes 1080p hevc at 15fps and x264 at 75 fps at 1080p? at what bit rate, 500 kb/s?

deadrats, have you ever encoded a video yourself?
I've never used anything faster than the x264 medium speed preset, but out of curiosity I thought I'd try re-encoding a 720p video using the x264 ultra fast preset. CRF18. Only with an old E6750 dual core, slightly over-clocked (3.2gig). Five minutes into the encoding job and it's sitting on around 82fps. At a very rough guess (I didn't run the whole thing) I think the output would be around 6000kb/s.
If I upscale that same 720p video to 1080p as a quick test, this dual core manages around 38fps, ultra fast preset, CRF18.

There's nothing particularly unusual about DarrellS's encoding speed.

[hogger129]

I'd say with x264 just for compatibility's sake. Plus unless you have a fast computer, it takes forever to encode using x265.

[Mephesto]

I can't believe this. I leave for a few months and immediately come back to see deadrats post 100 times about the same tired bullshtick as before.
x264's superiority over other implementations has already been established every time on MSU's annual codec competition so this is rather pointless but does anyone think that if I produced a video at the same bit rate as deadrat's chosen MainConcept encode and he saw the better quality for himself that this will finally shut his stupid-ass up?
Are you game, deadrats? Using your Do Elephants Dream as a source is fine with me.

Oh wait, he'll encode his with a super high bitrate to the point that there will be no difference in even the worst H264 codec and he'll go like "see? its only better at cellphone bitrates which no one should use anyway"

[deadrats]

I leave for a few months

yeah, those parole violations can be very inconvenient. are you even supposed to be using a computer? maybe you should just stick with the community service for a while.

[hello_hello]

I can't believe this. I leave for a few months and immediately come back to see deadrats post 100 times about the same tired bullshtick as before.

It'd be funny if he wasn't so tedious.

x264's superiority over other implementations has already been established every time on MSU's annual codec competition so this is rather pointless but does anyone think that if I produced a video at the same bit rate as deadrat's chosen MainConcept encode and he saw the better quality for himself that this will finally shut his stupid-ass up?
Are you game, deadrats? Using your Do Elephants Dream as a source is fine with me.

He'll ignore it and ramble on and on and on and on....... see...... he made a fool of himself yet again.....

Oh wait, he'll encode his with a super high bitrate to the point that there will be no difference in even the worst H264 codec and he'll go like "see? its only better at cellphone bitrates which no one should use anyway"

At least it'll give us something to laugh at..... well if he wasn't so tedious.

[Selur]

Come on, can some moderator please show some mercy and move this thread into the Off-topic section, since this thread long lost all seriousness a while ago and hardly has any useful information in it at side from:

a. deadrats doesn't like x264
b. lots of people are annoyed when he starts to post his opinions in the form he prefers
c. there seems to be a troll in each one of us
d. every few weeks someone posts in this thread and revives it

[deadrats]

@gonca:

any chance of updating your hexacore benchmarks now that the new divx265 cli is about 3 times faster?

[gonca]

@gonca:

any chance of updating your hexacore benchmarks now that the new divx265 cli is about 3 times faster?

Sure, just give me a day or so

[gonca]

file used
ducks_take_off_1080p50.y4m
bitrate 4000 kbps

balanced 23.29 fps
fastest 41.69 fps
highest quality 0.69 fps

[Kaustubh.patankar]

gonca @ What are the PSNR achieved for these tests?

[gonca]

didn't check. I'll have to redo the encodes

[gonca]

Hybrid does not show PSNR for DivX265 and I will not use the DivX installer

[Selur]

Does DivX265 even output any PSNR measurements? (if it does I could adjust Hybrid to show that output)

[gonca]

Does DivX265 even output any PSNR measurements? (if it does I could adjust Hybrid to show that output)

Don't think so.
It was kaustubh.patankar that raised the issue of PSNR for DivX265

[jagabo]

Since you have the source you can use MSU's video Quality Measurement Tool to generate PSNR values.

http://compression.ru/video/quality_measure/video_measurement_tool_en.html

I don't consider PSNR to be a very good measure of quality. For example, suppose I made a codec that was completely lossless except that it shifted the image left by one pixel. A video with spacial detail would get really bad PSNR scores. But a human watching the video would probably never notice the problem.

[gonca]

Can't make it accept Hybrid's output, and like you said, your eyes are the best test.
If Selur creates an output format that the tool can handle I'll try again

[poisondeathray]

You can load avs scripts into MSU VQMT ; ffms2 can load HEVC