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 > Video Encoding > MPEG-4 AVC / H.264

Reply
 
Thread Tools Search this Thread Display Modes
Old 3rd June 2008, 00:28   #1  |  Link
KornX
primus inter pares
 
KornX's Avatar
 
Join Date: Feb 2004
Posts: 70
h264 LOSSLESS experiences

if you do a 1080p24 encode with the lossless mode, what bitrates are we talking about (yeah i know, content...)
But what are your experiences?
And how many of you use it (at least sometimes)

KornX
__________________
Field Application Engineer
Advanced Micro Devices
KornX is offline   Reply With Quote
Old 6th June 2008, 19:29   #2  |  Link
bond
Registered User
 
Join Date: Nov 2001
Posts: 9,770
depends on what video you try to encode losslessly of course

so try it yourself on the for you relevant content
__________________
Between the weak and the strong one it is the freedom which oppresses and the law that liberates (Jean Jacques Rousseau)
I know, that I know nothing (Socrates)

MPEG-4 ASP FAQ | AVC/H.264 FAQ | AAC FAQ | MP4 FAQ | MP4Menu stores DVD Menus in MP4 (guide)
Ogg Theora | Ogg Vorbis
use WM9 today and get Micro$oft controlling the A/V market tomorrow for free
bond is offline   Reply With Quote
Old 6th June 2008, 22:16   #3  |  Link
Dust Signs
Registered User
 
Join Date: Jun 2004
Location: Salzburg, Austria
Posts: 219
Hi,

are you talking about perceptually lossless or really lossless (QP = 0)? In the latter case it's possible that you get data rates exceeding those of raw uncompressed YUV files, in the former case - of course depending on the content - you can expect some 20 to 50 MBits/s

Dust Signs
__________________
The number you dialed is imaginary. Please turn your phone by 90° and try again
Dust Signs is offline   Reply With Quote
Old 6th June 2008, 22:18   #4  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by Dust Signs View Post
In the latter case it's possible that you get data rates exceeding those of raw uncompressed YUV files
This is extremely doubtful. I've only ever seen a single macroblock ever exceed its uncompressed size with lossless x264 encoding, and technically the standard requires that any such block be coded as I_PCM anyways.
Dark Shikari is offline   Reply With Quote
Old 6th June 2008, 22:27   #5  |  Link
Dust Signs
Registered User
 
Join Date: Jun 2004
Location: Salzburg, Austria
Posts: 219
I wrote that it is possible and I experienced one such case myself when IPCM mode was accidently disabled. But anyways, using lossless mode makes no real sense to me because the bit rate required is extremely high and the difference to a file encoded with p.e. CRF 16 is perceptually zero.

Dust Signs
__________________
The number you dialed is imaginary. Please turn your phone by 90° and try again
Dust Signs is offline   Reply With Quote
Old 6th June 2008, 22:30   #6  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by Dust Signs View Post
I wrote that it is possible and I experienced one such case myself when IPCM mode was accidently disabled.
x264 doesn't actually support I_PCM mode; its in the code right now but it isn't enabled.
Quote:
Originally Posted by Dust Signs View Post
But anyways, using lossless mode makes no real sense to me because the bit rate required is extremely high and the difference to a file encoded with p.e. CRF 16 is perceptually zero.

Dust Signs
Its far better than most other lossless options in the case that inter prediction is very useful, such as for CGI or animated sources which were originally lossless.
Dark Shikari is offline   Reply With Quote
Old 6th June 2008, 22:33   #7  |  Link
Dust Signs
Registered User
 
Join Date: Jun 2004
Location: Salzburg, Austria
Posts: 219
Quote:
Originally Posted by Dark Shikari View Post
x264 doesn't actually support I_PCM mode; its in the code right now but it isn't enabled.
I wasn't talking about x264 concerning that matter - this time I was using a different encoder. Concerning CGI I agree with you, but I think many people are using "normal" video material (movies etc.) much more often.

Dust Signs
__________________
The number you dialed is imaginary. Please turn your phone by 90° and try again

Last edited by Dust Signs; 6th June 2008 at 22:36.
Dust Signs is offline   Reply With Quote
Old 11th June 2008, 23:32   #8  |  Link
elguaxo
Registered User
 
elguaxo's Avatar
 
Join Date: Jun 2006
Posts: 260
I'm doing my first lossles try right now. The source is a 01:31:20 old anime movie on DVD. It needs a lot of denoising, which I'm not doing right now. The lossles pass is only for AnimeIVTC. After 75% the projected filesize is of 12.8GB.

I'm using MeGUI's CQ-Lossless profile:
Code:
 --qp 0 --nf --subme 1 --partitions p8x8,b8x8,i4x4 --merange 12 --threads auto --thread-input --progress --no-psnr --no-ssim --output
elguaxo is offline   Reply With Quote
Old 11th June 2008, 23:45   #9  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by elguaxo View Post
I'm doing my first lossles try right now. The source is a 01:31:20 old anime movie on DVD. It needs a lot of denoising, which I'm not doing right now. The lossles pass is only for AnimeIVTC. After 75% the projected filesize is of 12.8GB.

I'm using MeGUI's CQ-Lossless profile:
Code:
 --qp 0 --nf --subme 1 --partitions p8x8,b8x8,i4x4 --merange 12 --threads auto --thread-input --progress --no-psnr --no-ssim --output
Subme 1? That's a waste; most time is spent in bitstream encoding, so a higher subme hardly takes any more time (until you get to 6 or above, at least). --nf does nothing there, and --no-psnr and --no-ssim do nothing too. The partitions is what is used as default, and the --merange is silly with --me hex.
Dark Shikari is offline   Reply With Quote
Old 12th June 2008, 00:04   #10  |  Link
elguaxo
Registered User
 
elguaxo's Avatar
 
Join Date: Jun 2006
Posts: 260
This encode is almost done, but I will use subme 5 or 6 next time. What would be a good merange value if I keep --me hex? Thanks
elguaxo is offline   Reply With Quote
Old 12th June 2008, 00:07   #11  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by elguaxo View Post
This encode is almost done, but I will use subme 5 or 6 next time. What would be a good merange value if I keep --me hex? Thanks
For hex, there's no reason to change it to anything but the default (16).
Dark Shikari is offline   Reply With Quote
Old 12th June 2008, 00:07   #12  |  Link
LoRd_MuldeR
Software Developer
 
LoRd_MuldeR's Avatar
 
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
Quote:
Originally Posted by elguaxo View Post
This encode is almost done, but I will use subme 5 or 6 next time. What would be a good merange value if I keep --me hex? Thanks
I think --merange does not apply to "hex" mode, only "umh" and higher...
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊
LoRd_MuldeR is offline   Reply With Quote
Old 12th June 2008, 00:13   #13  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by LoRd_MuldeR View Post
I think --merange does not apply to "hex" mode, only "umh" and higher...
It does, its just that its clipped to 4<=X<=16.
Dark Shikari is offline   Reply With Quote
Old 12th June 2008, 00:15   #14  |  Link
elguaxo
Registered User
 
elguaxo's Avatar
 
Join Date: Jun 2006
Posts: 260
elguaxo is offline   Reply With Quote
Old 12th June 2008, 00:21   #15  |  Link
LoRd_MuldeR
Software Developer
 
LoRd_MuldeR's Avatar
 
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
Quote:
Originally Posted by Dark Shikari View Post
It does, its just that its clipped to 4<=X<=16.
Sorry, I was mistaken
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊
LoRd_MuldeR is offline   Reply With Quote
Old 12th June 2008, 23:12   #16  |  Link
foxyshadis
ангел смерти
 
foxyshadis's Avatar
 
Join Date: Nov 2004
Location: Lost
Posts: 9,558
If you want a lot more speed out of lossless, just turn cabac off. It'll be slightly larger but not as much as you'd expect. You lose a lot less efficiency than you do at low bitrates, since so many of the randomer high quantizers are present.
foxyshadis is offline   Reply With Quote
Old 12th June 2008, 23:16   #17  |  Link
Dark Shikari
x264 developer
 
Dark Shikari's Avatar
 
Join Date: Sep 2005
Posts: 8,666
Quote:
Originally Posted by foxyshadis View Post
If you want a lot more speed out of lossless, just turn cabac off. It'll be slightly larger but not as much as you'd expect. You lose a lot less efficiency than you do at low bitrates, since so many of the randomer high quantizers are present.
The benefit of CABAC in lossless is actually higher than at low quantizers:


(thanks to akupenguin for this chart)

But yes, overall, its definitely worth turning off given the very large speed boost one gets; (de|en)coding lossless is nearly pure entropy (de|en)coding.

Last edited by Dark Shikari; 13th June 2008 at 01:58.
Dark Shikari is offline   Reply With Quote
Old 13th June 2008, 01:58   #18  |  Link
foxyshadis
ангел смерти
 
foxyshadis's Avatar
 
Join Date: Nov 2004
Location: Lost
Posts: 9,558
Oh wow, that's crazy, thanks for letting me know. I normally use might-as-well-be-lossless qp 1 or 2, so I assumed behavior would translate over, good to know the truth.
foxyshadis is offline   Reply With Quote
Reply

Tags
1080p24, avc, h264, lossless, x264

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 15:22.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.