The fascinating quality of x265!1080p video with fast movements in 1300kb/

[deadrats]

2 things:

some of the test sources being used in this comparison are not indicative of the sources that most people will be using with the various encoders.

AND, posting screenshots is not an accurate way of representing quality as we have no way of knowing if the screenshot in question is of a I, P, or B frame, since each codec will use a different gop structure and it's already been established that there may be some circumstances where h264 looks better in still pictures but significantly worse when viewed as a movie.

[Atak_Snajpera]

That's why I always post links to encoded video files
Everyone can check how it looks in motion.

[Atak_Snajpera]

Ok The last test from me. This time troll's sample
x265 --crf 28 --preset medium --aq-mode 1 --aq-strength 1 --fps 25
https://mega.co.nz/#!pBsEzAxQ!RMsqjjAyyyJQ_TH74kFNgEw2Tj953IWYVu__Xo9LDMg
AVERAGE BITRATE: 1948 kbps

x264 --crf 34.25 --preset veryslow
https://mega.co.nz/#!RN9xlaZB!aiRSSGtZIxFERVgmrtbc9RnJWn8Y1879BfCkTeBFc7s
AVERAGE BITRATE: 1972 kbps

Screenshots
x265
http://i.cubeupload.com/bb6tcp.png
http://i.cubeupload.com/8behZ6.png

x264
http://i.cubeupload.com/dfh31Y.png
http://i.cubeupload.com/I4XZl2.png



This is really funny. Encoder decided to remove shadow and that tiny window looks now like some square hole.

[Mephesto]

When the windows Vista and Win7 appeared, many young people demanded the XP op system for their brand new PC and Laptop. I first used MS DOS than Windows 3.1 than Win95 Wing98 etc...
So XP was just an newer Op. system for me. However the XP means THE OP.SYSTEM for the young people, because when they started to use computer, the XP was THE FIRST op.system which they knew.

*s* My first OS was Windows95 when I was in kindergarten, the first OS I have memory of anyway, I switched to every new OS right when it came out except Vista because it blows **** and this is a sentiment shared by the majority of the PC community, go figure.

The same childish emotional prepossessions are true for video codecs too.
These forum members are teens or so-called junior aged (under 25years) x264 "fan-boys". When they started to edit / compress videos, the h264 (the x264) was the king of compression. When I started to edit / compress videos , the mpeg 1 mpeg II (DVD age) and h.263 (divx than xvid) became the king, the h264 based codecs were JUST A NEWER codec. For these young guys, the x264 IS THE CODEC.
That's why they are so emotionally and irrationally biased.

Wrong again, I've messed around with other codecs in the past but I hated them all and never used them for anything serious until I tried H264 and saw the awesome quality that didn't require a bitrate that would take hours and hours to transfer.
I've spent a lot of fruitless energy advertising and pressing all those xvid fanboys to move on to x264 back in 2007 where I was met with nothing but resistance and relentless bitching. I've also been following and awaiting H265 since 2008 so your accusation is void. I look forward to the future but I'm not gonna start hopping on the prototype jet in 1911 when I saw with my own eyes that it crashes and falls apart half the time. I'd stick to the automobile until that gets fixed.

Atak Snajpera posted numerous tests which you ignored and kept posting defensively like right now and projected your irrational, biased tantrum onto everybody else.

[Atak_Snajpera]

The main problem with HEVC is that encoding time will always be about 3-5 times slower than AVC (for example highly optimized x265 vs x264). And you can not expect more than 40% of size reduction. Most likely it will be something between 25-33%. (x265 vs x264).

Well something tells me that HEVC will end up like Jpeg2000 or Jpeg-XR. Both offer better compression than old jpeg but camera manufactures, users still use older less efficient format. Reasons: higher compression is not needed. We have big memory cards / hdd. Internet is alot faster than in mid 90s (56k modems

The same with audio formats. Most people still prefer mp3@256 kbps than AAC/Vorbis/Opus@128 kbps

[Mephesto]

The main problem with HEVC is that encoding time will always be about 3-5 times slower than AVC (for example highly optimized x265 vs x264). And you can not expect more than 40% of size reduction. Most likely it will be something between 25-33%. (x265 vs x264).

Well something tells me that HEVC will end up like Jpeg2000 or Jpeg-XR. Both offer better compression than old jpeg but camera manufactures, users still use older less efficient format. Reasons: higher compression is not needed. We have big memory cards / hdd. Internet is alot faster than in mid 90s (56k modems

The same with audio formats. Most people still prefer mp3@256 kbps than AAC/Vorbis/Opus@128 kbps

You're making way too many assumptions. x264 ended up being faster than xvid, almost twice as better quality and widely adopted. In the beginning it was intended to be twice as better than MPEG-2. Some implementations were WORSE than MPEG-2 like the QuickTime H264 codec. We don't know what will happen with HEVC.

Higher compression not needed? Are you sure? 4K UHD won't download in a matter of minutes or even hours, my friend.

We have big memory cards / hdd.

Which are not free and HDDs are slow as shit.

Internet is alot faster than in mid 90s (56k modems

Actually the net is a lot SLOWER today when you take into account that size of content has increased a lot more than speed.

A mouse driver was 50KB in the 28.8 K modem days, it took less than 20 seconds to download.
Now, a mouse driver is 60MB in size because the ****-smokers running the manufacturer sites decide to include every version of .NET Framework in the package (which I already had installed) and uncompressed BMPs.

For this to download as fast as the same driver 15 years ago on a 28.8 K modem would require a 24 Mbit connection which is not even available in my area.

Same trend with video. The bitrates are growing a lot faster than network capacities are.

[Brad]

x264 ended up being faster than xvid

Is that the case when both are run single-threaded?

[addu]

This discussion is useless and ugly we are not going to use x265 for hd but definitely for ultra hd because x264 quality and speed is enough for the current hd.

[Mephesto]

x264 ended up being faster than xvid

Is that the case when both are run single-threaded?

When you tweak the settings yes, x264 encodes faster and still delivers better quality.

[Tommy Carrot]

Here is my attempt at encoding park joy with x265:

http://www.load.to/j7J8OPY9zY/x265_q37.h265
average bitrate: 4979 kbps

screenshot: http://i.cubeupload.com/ot6MAF.png

The quality is still not very good, but imo significantly better than Atak_Snajpera's encoding. Why? Because i encoded with build 0.3+293, back from august. The quality/efficiency of x265 has seriously regressed since the earlier builds, and while it's good they are trying to improve encoding speed, i'd like them to try improve quality/efficiency as well, at least back to the level of the earlier builds.

[Selur]

@TommyCarrot: I was under the impression that only the presets changed a lot, but in general slow and quality encoding link in the 0.3 builds should still be possible.
-> this someone try to encode park_joy with x265 without using the presets?

[Tommy Carrot]

@TommyCarrot: I was under the impression that only the presets changed a lot, but in general slow and quality encoding link in the 0.3 builds should still be possible.

0.3 builds still used a lot of stuff from the HM reference encoder. Since then, they replaced most of them with their own algorythms and some modified from x264. I don't know if the quality regression is due to missing features or some bugs, but not even placebo preset can approach the quality of the old builds.

[deadrats]

0.3 builds still used a lot of stuff from the HM reference encoder. Since then, they replaced most of them with their own algorythms and some modified from x264. I don't know if the quality regression is due to missing features or some bugs, but not even placebo preset can approach the quality of the old builds.

my guess is that any quality regression is due to them trying to shoehorn algorithms designed for h264 encoding into an h265 encoder coupled with them taking math shortcuts in their calculations in an attempt to speed up their encoder.

[Stears555]

Ok The last test from me. This time troll's sample
x265 --crf 28 --preset medium --aq-mode 1 --aq-strength 1 --fps 25
https://mega.co.nz/#!pBsEzAxQ!RMsqjjAyyyJQ_TH74kFNgEw2Tj953IWYVu__Xo9LDMg
AVERAGE BITRATE: 1948 kbps

x264 --crf 34.25 --preset veryslow
https://mega.co.nz/#!RN9xlaZB!aiRSSGtZIxFERVgmrtbc9RnJWn8Y1879BfCkTeBFc7s
AVERAGE BITRATE: 1972 kbps

Screenshots
x265
http://i.cubeupload.com/bb6tcp.png
http://i.cubeupload.com/8behZ6.png

x264
http://i.cubeupload.com/dfh31Y.png
http://i.cubeupload.com/I4XZl2.png



This is really funny. Encoder decided to remove shadow and that tiny window looks now like some square hole.

[Attachment 21865 - Click to enlarge]
[Attachment 21866 - Click to enlarge]

Your pattern is very simple.

you chose lower quality options for x265, and a better quality option for x264, then you post the misleading photos and videos.

[Atak_Snajpera]

x264 --preset veryslow is still 4 x faster than x265 --preset medium you idiot ! What was I suppose to use for x265? --preset placebo!? And wait whole day to encode F-U-C-K-I-N-G 1 min sample!? Why don't you encode your sample by yourself then you genius?!

[deadrats]

x264 --preset veryslow is still 4 x faster than x265 --preset medium you idiot ! What was i suppose to use --preset placebo!? And wait whole day to encode F-U-C-K-I-N-G 1 min sample!?

this thread isn't about the relative speeds of the encoders in question but rather the quality of the encodes it produces, thus if you're going to make a fair quality comparison then yes, you should choose options of each that are as close as possible regardless of encode time.

now if you want to say that x264 is a more practical choice because of it's superior speed, then that's a different story but this thread is about the quality that can be achieved by using either one.

in fact, that douche DS once did an entire rant on that half-assed blog of his about people that cheat in encoder comparisons by using different settings between encoders and then declaring the encoder that used lower quality settings the loser. he was of course complaining that this was being done to his little baby but it seems that the x264 faithful are not above using the same tactics.

and for the record, in case you haven't seen most of my posts, i am in no way an x265 fan boy, in fact i have been highly critical of the project from the get go but fair is fair and your comparison encodes are not fair.

[Atak_Snajpera]

and for the record, in case you haven't seen most of my posts, i am in no way an x265 fan boy, in fact i have been highly critical of the project from the get go but fair is fair and your comparison encodes are not fair.

I know that ! So far only Stears555 is a blind x265 fan boy. Now let's wait for his sample with --preset placebo because I'm not going to waste ~6 kWh for that idiot.

in fact, that douche DS once did an entire rant on that half-assed blog of his about people that cheat in encoder comparisons by using different settings between encoders and then declaring the encoder that used lower quality settings the loser. he was of course complaining that this was being done to his little baby but it seems that the x264 faithful are not above using the same tactics.

The fact is that douche DS as you called him has put alot of work to improve quality in x264. AQ / Psy-RDO / MB-tree and so on. Not to mention about those handmade asm optimalizations. So far x264 offers the best quality and encoding speed.