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 > (HD) DVD, Blu-ray & (S)VCD > DVD & BD Rebuilder

Reply
 
Thread Tools Search this Thread Display Modes
Old 20th September 2014, 21:00   #1  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
BD-Rebuilder - HEVC

I have decided to dedicate this thread to HEVC and BD-RB discussion, instead of using bug reports thread.

I recently finished testing 4K animation source and posting result here.

Common between x264 and x265

--crf 20 --input-res 4096x2160 --fps 24000/1001 --input-csp i420

X264
--output "F:\ Sintel 2010 4K.264"
Bitrate average 20694.40 kb/s
PSNR Mean Y 48.611 | U 54.778| V 54.238
PSNR Average 49.564
Final File Size 2.13 GB

X265
--output "F:\ Sintel 2010 4K.hevc"
Bitrate average 16249.31 kb/s
PSNR Mean Y 47.644 | U 54.269 | V 53.687
PSNR Average 49.288
Final File Size 1.67 GB

*Source is freely available for download to public under creative commons license, do google search if you need it.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.

Last edited by HWK; 20th September 2014 at 21:07.
HWK is offline   Reply With Quote
Old 21st September 2014, 00:03   #2  |  Link
gonca
Registered User
 
Join Date: Jul 2012
Posts: 1,026
here are some comparisons
Attached Images
File Type: pdf comparison.pdf (95.3 KB, 255 views)
gonca is offline   Reply With Quote
Old 21st September 2014, 01:05   #3  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
I am gone be doing tear of steel next in 4K and post result, my aim is to see how psnr change with different source.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 21st September 2014, 08:54   #4  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,555
What does this have to do with BD-RB? This looks like a simple x265 vs x264 match. A large number of GUIs will or already do have x264 and x265 support. It doesn't make sense to create new threads for every GUI as the underlying encoders and results will be the same for all them. This rather belongs into the HEVC sub-forum.

Aside from that: CRF scales of x264 and x265 are not identical, so be careful. Since you provided both bitrate and PSNR this might have at least some value, though. Then again different psy defaults might skew the PSNR results.

Last edited by sneaker_ger; 21st September 2014 at 08:59.
sneaker_ger is offline   Reply With Quote
Old 21st September 2014, 16:14   #5  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
I created this thread in hope of taking off hevc discussion posts found in bd-rb bugs report thread.

Only reason you see comparison is bd-rb uses x264 to compress and some of us want to do test and see how efficient x265 is over x264.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 21st September 2014, 16:18   #6  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,679
BD-RB will support x265 backups in the next release, so this can be useful so I can adjust default settings. What? Are we now limiting what people can say specific to their own method of backup? An extra thread isn't going to hurt anything.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Old 21st September 2014, 16:52   #7  |  Link
gonca
Registered User
 
Join Date: Jul 2012
Posts: 1,026
HWK, jdobbs
any testing you need done let me know, I have some idle cpu time.
gonca is offline   Reply With Quote
Old 21st September 2014, 17:08   #8  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
Quote:
Originally Posted by gonca View Post
HWK, jdobbs
any testing you need done let me know, I have some idle cpu time.
Actually I was thinking so far I am the only one who has done 4K sources, beside me if someone else do it might help establish baseline.I recently got tear of steel and I think it would be good candidate since it is mix of animation and footage short on camera.

Here is my to do list

1. Do grainy video such as saving private ryan
2. Perform test on black and white video
3. Conduct two pass test on both encoders with challenge footage

Jdobbs, is there something we can do to help you or need help with?
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.

Last edited by HWK; 21st September 2014 at 17:53.
HWK is offline   Reply With Quote
Old 21st September 2014, 20:31   #9  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,555
Quote:
Originally Posted by jdobbs View Post
An extra thread isn't going to hurt anything.
It is. A lot of people are looking for reviews of the x265 development progress but only few people will look into the BD-Rebuilder sub-forum. Outside people will miss out on the infos collected here and you guys will get less feedback and suggestions concerning your test setups. It's lose-lose.
sneaker_ger is offline   Reply With Quote
Old 21st September 2014, 20:33   #10  |  Link
gonca
Registered User
 
Join Date: Jul 2012
Posts: 1,026
Here's an idea, since you started the thread you can name test samples and parameters, etc to use. this way a more valid set of results can be achieved, instead of being all over the place
gonca is offline   Reply With Quote
Old 21st September 2014, 20:56   #11  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
Quote:
Originally Posted by gonca View Post
Here's an idea, since you started the thread you can name test samples and parameters, etc to use. this way a more valid set of results can be achieved, instead of being all over the place
Good call, I will post setting and sample later.

For testing purpose all of my sources are 4K and yuv format so far.

x264.exe --psnr --tune psnr --crf 20 --input-res 4096x2160 --fps 24000/1001 --input-csp i420 --output "F:\movie name.264" "movie name.yuv"

pause

x265.exe --psnr --tune psnr --crf 20 --input-res 4096x2160 --fps 24000/1001 --input-csp i420 --output "F:\movie name.hevc" "movie name.yuv"

pause

I will try to upload source later
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.

Last edited by HWK; 21st September 2014 at 22:24. Reason: clear up confusion
HWK is offline   Reply With Quote
Old 21st September 2014, 21:20   #12  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
Quote:
Originally Posted by sneaker_ger View Post
It is. A lot of people are looking for reviews of the x265 development progress but only few people will look into the BD-Rebuilder sub-forum. Outside people will miss out on the infos collected here and you guys will get less feedback and suggestions concerning your test setups. It's lose-lose.
Ok, let me try to explain the purpose of this thread in case there is any doubt about it. I created this only for one purpose to remove post which is related to hevc and are posted in BD-RB bug report thread.

I never thought from feedback perspective, however I have included tags to help search engine.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 21st September 2014, 22:04   #13  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
PSNR (Tears of Steel 2012 UHD 4K)

Common between x264 and x265

--crf 20 --input-res 4096x2160 --fps 24000/1001 --input-csp i420

X264
--output "F:\ Tears of Steel 2012 UHD 4K.264
Bitrate average 36121.39 kb/s
PSNR Mean Y 44.808 | U 49.709| V 49.770
PSNR Average 45.791
Final File Size 3.08 GB

X265
--output "F:\ Tears of Steel 2012 UHD 4K.hevc"
Bitrate average 17710.16 kb/s
PSNR Mean Y 44.012 | U 49.213 | V 49.277
PSNR Average 45.321
Final File Size 1.51 GB

One thing I noticed bitrate was almost double for x264 to come close. Normally it is up to 30 % of x265.
I am gone upload the source for this one to try out.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 21st September 2014, 22:16   #14  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,555
Because of psy both x264 and x265 developers suggest using "--tune psnr" if you want to compare psnr. The psy optimizations increase subjective quality but reduce quality of metrics such as psnr and ssim.

But again: same crf does not result in same quality between 264 and x265.
sneaker_ger is offline   Reply With Quote
Old 21st September 2014, 22:21   #15  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
I did use --tune psnr and --psnr, look at post 11 by me.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 21st September 2014, 22:25   #16  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,555
Ah, I see. I only looked at #13. To put into easier to grasp numbers this suggests x264 offered +9.4% quality but at +104% the bitrate (if measured via objective PSNR metric).
sneaker_ger is offline   Reply With Quote
Old 21st September 2014, 22:29   #17  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,679
Quote:
Originally Posted by sneaker_ger View Post
It is. A lot of people are looking for reviews of the x265 development progress but only few people will look into the BD-Rebuilder sub-forum. Outside people will miss out on the infos collected here and you guys will get less feedback and suggestions concerning your test setups. It's lose-lose.
I disagree. So in your mind, once you made a post about any subject -- nobody can post anything similar anywhere else? When did DOOM9 become that dictatorial?
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Old 21st September 2014, 22:34   #18  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
Quote:
Originally Posted by sneaker_ger View Post
Ah, I see. I only looked at #13. To put into easier to grasp numbers this suggests x264 offered +9.4% quality but at +104% the bitrate (if measured via objective PSNR metric).
Much better that is why I am uploading the source for people to try out and report there findings.
This is my findings maybe someone will come up with something different.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.

Last edited by HWK; 21st September 2014 at 22:42.
HWK is offline   Reply With Quote
Old 21st September 2014, 22:53   #19  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,555
Quote:
Originally Posted by jdobbs View Post
I disagree. So in your mind, once you made a post about any subject -- nobody can post anything similar anywhere else?
No, I didn't say you can't make a new post. I'm saying you should do it in the correct sub-forum. There's a reason doom9 is divided into a bunch of them. Since this thread offers nothing specific to BD-RB it is in the wrong sub-forum.

But, well, I think I made my point. Last post from me on the sub-forum discussion.
sneaker_ger is offline   Reply With Quote
Old 21st September 2014, 23:12   #20  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 3,729
Black borders -- if any -- should be cropped for the PSNR tests because they bias (mask) the result.
Sharc is offline   Reply With Quote
Reply

Tags
bd-rb, h.265, hevc

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 00:53.


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