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 13th September 2014, 08:37   #21081  |  Link
daberti
Registered User
 
Join Date: Apr 2004
Posts: 172
Quote:
Originally Posted by jdobbs View Post
Ok. I'll test with LAVF set and output to ALTERNATE. Which ALTERNATE preset did you use? Better yet, post your log on the one that failed.
The two logs (the failing one comes first):

[09/12/14] BD Rebuilder v0.48.05 (beta)
[18:21:48] Source: L_UOMO_CHE_FISSA_LE_CAPRE_00000
- Input BD size: 20,18 GB
- Approximate total content: [01:34:03.971]
- Windows Version: 6.1 [7601]
- MOVIE-ONLY/ALTERNATE OUTPUT mode enabled
- Mode: MKV Container, 1920x1080, Intact Audio
- Decoding/Frame serving: X264/LAVF
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[18:21:51] PHASE ONE, Encoding
- [18:21:51] Processing: VID_00000 (1 of 1)
- [18:21:51] Extracting A/V streams [VID_00000]
- [18:33:32] Reencoding video [VID_00000]
- Source Video: VC-1, 1920x1080
- Rate/Length: 23,976fps, 135.320 frames
- [18:33:32] Reencoding: VID_00000, Pass 1 of 1
- [19:17:24] Video Encode complete
- [19:17:24] Processing audio tracks
- Track 4352 (ita): Keeping original audio
[19:17:24]PHASE ONE complete
[19:17:24]PHASE TWO - Rebuild Started
- [19:17:24] Building ALTERNATE OUTPUT Structure
[19:18:14] - Encode and Rebuild complete
[19:18:14] JOB: L_UOMO_CHE_FISSA_LE_CAPRE finished.
----------------------
[09/12/14] BD Rebuilder v0.48.05 (beta)
[19:54:59] Source: L_UOMO_CHE_FISSA_LE_CAPRE_00000
- Input BD size: 20,18 GB
- Approximate total content: [01:34:03.971]
- Windows Version: 6.1 [7601]
- MOVIE-ONLY/ALTERNATE OUTPUT mode enabled
- Mode: MKV Container, 1920x1080, Intact Audio
- Decoding/Frame serving: DirectShow
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[19:55:02] PHASE ONE, Encoding
- [19:55:02] Processing: VID_00000 (1 of 1)
- [19:55:02] Extracting A/V streams [VID_00000]
- [20:07:00] Reencoding video [VID_00000]
- Source Video: VC-1, 1920x1080
- Rate/Length: 23,976fps, 135.320 frames
- [20:07:00] Reencoding: VID_00000, Pass 1 of 1
- [21:18:29] Video Encode complete
- [21:18:29] Processing audio tracks
- Track 4352 (ita): Keeping original audio
[21:18:29]PHASE ONE complete
[21:18:29]PHASE TWO - Rebuild Started
- [21:18:29] Building ALTERNATE OUTPUT Structure
[21:21:23] - Encode and Rebuild complete
[21:21:23] JOB: L_UOMO_CHE_FISSA_LE_CAPRE finished.

*****
Both done at Fastest encoding option, CRF 20, NO Multiprocess.
Feel free to ask further information-
daberti is offline   Reply With Quote
Old 13th September 2014, 13:44   #21082  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by jdobbs View Post
Thanks. I'm doing testing now using SSIM to try and determine the relationship between CRF values for each encoder (X264 & X265). There'll also have to be a conversion to H.264 if importing HEVC to make it compatible with BD (at least until they finalize H.265 support)... but I do that now with a DIVX or MPEG-4 sources, so it shouldn't be a big deal. I just have to do it as lossless as possible while keeping a reasonable interim size.
Just an update. It looks like X264 and X265's CRF values are very closely aligned. The PSNR is very close for each setting I tested. Close enough that it makes you think it may be the metric used for CRF. That's a good thing -- as I won't have to do any conversions to give consistent results.

Based on my tests, it looks like you will see a reduction in file size of between 31-33% for an identical level of quality when you create a backup MKV using HEVC rather than AVC. That doesn't reach the 50% reduction you see advertised often -- but that's probably because I'm testing at higher quality levels and bitrates (typical of a backup). HEVC would likely get higher reduction levels while maintaing the same quality when it is forced to fit into highly constrained bitrates (like you might see in satellite transmission). Also, I'm using a single source video across all my testing... but I think that's ok for what I'm trying to accomplish.

I'm now doing SSIM tests to see how closely it aligns with PSNR. But SSIM testing is pretty slow...
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 13th September 2014 at 14:38.
jdobbs is offline   Reply With Quote
Old 13th September 2014, 13:54   #21083  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 4,041
You have probably already seen the x264/x265 comparative discussion here.
Sharc is offline   Reply With Quote
Old 13th September 2014, 14:36   #21084  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by Sharc View Post
You have probably already seen the x264/x265 comparative discussion here.
Yeah. But I like to do my own testing. Sometimes people like to report what they want to see as opposed to objective facts. I set pretty strict rules for myself when I do testing of this type. I also tailor my testing for BD-RB and BD backups as opposed to possible other uses.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 13th September 2014 at 14:39.
jdobbs is offline   Reply With Quote
Old 13th September 2014, 18:03   #21085  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
Quote:
Originally Posted by jdobbs View Post
That doesn't reach the 50% reduction you see advertised often -- but that's probably because I'm testing at higher quality levels and bitrates (typical of a backup). HEVC would likely get higher reduction levels while maintaing the same quality when it is forced to fit into highly constrained bitrates (like you might see in satellite transmission). Also, I'm using a single source video across all my testing... but I think that's ok for what I'm trying to accomplish.
Jdobbs, I think 50 % reduction is theoretical maximum, also hevc compression would truly shine on 4K sources.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.
HWK is offline   Reply With Quote
Old 13th September 2014, 19:12   #21086  |  Link
daberti
Registered User
 
Join Date: Apr 2004
Posts: 172
Quote:
Originally Posted by Lathe View Post
Thanks! I do have an on board monitor that keeps track of all the temperatures and such. They seem fine; I even stuck a big fan on the side, blowing in on the MOBO and CPU, thinking that might help also. I sure don't know; very odd... I've got the guy who built it though and he is pretty sharp, so I just need to drag it over there. I will do what JD said though and give it a good cleaning.

As far as the speed of encoding goes, it is SO dang weird now. I have absolutely NO idea what changed, but now encodes (using LAVF like I always do) take like twice as long (12 hours instead of 6) I'm pretty sure everything is the same and it LOOKS like I am topping out my CPU while encoding too. Most odd... I even used version 4707 just as a momentary test comparison, just to start it and see how fast it was going, but it appeared to be about the same as the newer version. I can't for the life of me figure out WHAT has changed. Oh well, at least it is getting done, and I guess if I INSIST on using these x264 tweaks (the same ones I've always used on demanding encodes) I'm just gonna hafta live with it taking longer. I can tell immediately that the fps are a lot slower than they were before. I'll play around with it and see if I am indeed missing something. But, I AM glad that the program is working and doing an excellent job though!
Question: are you sure that Thermal compound is still there between Cpu and its cooler?

Other than this building to ALTERNATE, CRF 20, Better speed, and FFDshow frame serving I scored 1.7x speed than with LAVF. No MULTIPROCESS FYI.

@JD
When your H265 code will be ready it will be time for another donation, as you're doing a heck of a work.
daberti is offline   Reply With Quote
Old 13th September 2014, 19:26   #21087  |  Link
Lathe
Registered User
 
Lathe's Avatar
 
Join Date: Aug 2005
Posts: 1,268
Quote:
Originally Posted by jdobbs View Post
You weren't, by chance, encoding a DVD import and resizing it, were you? I noticed something odd early today in TSMUXER that could (very rarely, I'd guess) cause it to detect a 480i/29.97 source as 480p/29.97 (it might also happen on 576i/25, but I haven't seen it). That caused an encode failure on a source I tried. Interestingly TSMUXER lists it at 480i -- but sets the MPLS as 480p/29.97 (which is illegal for a BD primary source). I've added a workaround that will be included in the next release of BD-RB.
No, I pretty much never resize anything (unless in rare cases where I'm starting with an out of spec AR or something, but not in this case, it was a full, normal Blu-ray) But, thanks for the information though; good to know.
Lathe is offline   Reply With Quote
Old 13th September 2014, 19:57   #21088  |  Link
Lathe
Registered User
 
Lathe's Avatar
 
Join Date: Aug 2005
Posts: 1,268
Quote:
Originally Posted by daberti View Post
Question: are you sure that Thermal compound is still there between Cpu and its cooler?

Other than this building to ALTERNATE, CRF 20, Better speed, and FFDshow frame serving I scored 1.7x speed than with LAVF. No MULTIPROCESS FYI.

@JD
When your H265 code will be ready it will be time for another donation, as you're doing a heck of a work.
Yes, you could very well be right; this has occurred to me before. The way it is acting, despite the additional fan, would likely suggest something heat related.

Thanks for the information about encoding speed; it is indeed quite curious as to the difference in speed now. I'm gonna keep working on it to see if I can figure it out.
Lathe is offline   Reply With Quote
Old 13th September 2014, 20:38   #21089  |  Link
Lathe
Registered User
 
Lathe's Avatar
 
Join Date: Aug 2005
Posts: 1,268
This was interesting. A reply on the basic BDRB guide thread at ClubMice came up and the fellow who updates the guide there said this:

"The latest version as I am writing this post is 0.48.05. I've noticed some different behavior in the program regarding frame-serving. Jdobbs has added a line that can make DirectShowSource the default frame server, and has removed an option to use FFVideoSource that was only in the program for a few editions before being abandoned.

The reason I mention this is that CRF encoding seems to be disabled when using the internal LAV filters for frame serving. You'll need to select DirectShowSource if you intend to make a one pass CRF encoding."


I didn't really quite understand what he was saying about the LAV filters (which I use) Would this have anything at all to do with my slower encoding speeds now as opposed to before...? I DO normally use High Quality, 2 Pass though; it doesn't seem like that would be affected by what he is talking about. Do you suppose that some of my 'Tweaks' are being processed or treated differently than before? These are the standard tweaks that I use when encoding either a rather large and/or a rather demanding Blu-ray:

TWEAK_PASS_TWO=--deblock -1:-1 --psy-rd 1.00:0.20 --me umh --subme 8 --trellis 2 --direct auto --qcomp 0.50 --b-adapt 2

Again, in the past when incorporating these same tweaks and primary settings with just about any Blu-ray, the 2nd pass would normally take around 6-7 hours. Now it takes 11-12 (BTW, your HQ setting I believe already sets the me & subme at these values if I'm not mistaken - at 'Highest' it sets the subme at 9 which is indeed overkill and did take forever, but these specific tweaks never really added that much time before)
Lathe is offline   Reply With Quote
Old 13th September 2014, 23:40   #21090  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 4,041
Quote:
Originally Posted by Lathe View Post
TWEAK_PASS_TWO=--deblock -1:-1 --psy-rd 1.00:0.20 --me umh --subme 8 --trellis 2 --direct auto --qcomp 0.50 --b-adapt 2
- Post your lastcmd.txt
- Why tweak like this? You risk to spoil pass 1 by forcing conflicting settings in pass 2. Leave --trellis 2 and --direct auto and --qcomp 0.50 and --b-adapt 2 away, or better just select Auto-quality and see what you gain in speed, and tell us the difference in quality.
Sharc is offline   Reply With Quote
Old 14th September 2014, 01:05   #21091  |  Link
Lathe
Registered User
 
Lathe's Avatar
 
Join Date: Aug 2005
Posts: 1,268
Quote:
Originally Posted by Sharc View Post
- Post your lastcmd.txt
- Why tweak like this? You risk to spoil pass 1 by forcing conflicting settings in pass 2. Leave --trellis 2 and --direct auto and --qcomp 0.50 and --b-adapt 2 away, or better just select Auto-quality and see what you gain in speed, and tell us the difference in quality.
Yeah, I was thinking that I probably should just leave it with maybe tune--film. I'm pretty sure that with Auto-quality it would indeed be a lot faster. But, even with that said, still... it IS kind of weird that before this WITH these tweaks, the encode times were ALWAYS considerably shorter (about half) So, I figured at THAT time that it didn't really take that much longer, so I'd throw them in. Now, it doesn't seem to be the case and it takes twice as long. But, you could very well be right and I probably should leave it on Auto or just tune--film (I do want the Deblock though to sharpen the picture as per 'Film' settings and I DO kind of like the filmlike Psychovisual bump too - neither of these are ever activated in BDRB, and neither is Trellis) Also, from what I've READ, the qcomp 0.50 contributes to a slightly sharper picture too, but I'm just going by what Selur, the guy who created HYBRID said about his settings. I'm certainly NOT an expert, by any means!

When you say 'Conflicting settings', isn't adding Tweaks merely choosing stronger settings for the final 2nd pass than would normally be determined by a first pass based upon an Auto or lesser setting? In other words, aren't you just 'Telling' x264 to use those higher settings rather than relying on what it's first pass 'Determines' based upon the Auto or a lighter setting? Is there really a 'Conflict'? I'm still VERY green at this, so I'm probably just missing something important here, but just going by what I've learned so far, that is how I understand it.

FWIW, the encodes that I've been doing have turned out breathtakingly beautiful! The delineation and the depth of field especially in long shots or dark shots is pretty amazing. But, perhaps I would get that anyway with Auto

Oh, I'll check my last command txt too next time I do an encode.

BTW, it appears that perhaps my major PC issue may indeed be heat related; I'm working on that now. Thanks for the suggestions!

Last edited by Lathe; 14th September 2014 at 01:08.
Lathe is offline   Reply With Quote
Old 14th September 2014, 01:11   #21092  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by jdobbs View Post
Just an update. It looks like X264 and X265's CRF values are very closely aligned. The PSNR is very close for each setting I tested. Close enough that it makes you think it may be the metric used for CRF. That's a good thing -- as I won't have to do any conversions to give consistent results.

Based on my tests, it looks like you will see a reduction in file size of between 31-33% for an identical level of quality when you create a backup MKV using HEVC rather than AVC. That doesn't reach the 50% reduction you see advertised often -- but that's probably because I'm testing at higher quality levels and bitrates (typical of a backup). HEVC would likely get higher reduction levels while maintaing the same quality when it is forced to fit into highly constrained bitrates (like you might see in satellite transmission). Also, I'm using a single source video across all my testing... but I think that's ok for what I'm trying to accomplish.

I'm now doing SSIM tests to see how closely it aligns with PSNR. But SSIM testing is pretty slow...
Finished the SSIM testing... it pretty much just mirrored what I saw in PSNR. I'm now making updates to the ALTERNATE routines to support HEVC and implementing the encoding. After that I'll have to make some changes to the Video Import area to support importing the MKVs created.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 14th September 2014 at 01:14.
jdobbs is offline   Reply With Quote
Old 14th September 2014, 02:03   #21093  |  Link
HWK
Registered User
 
HWK's Avatar
 
Join Date: Feb 2009
Location: Toronto, Ontario, Canada
Posts: 1,059
I am assuming jdobbs at this stage it gone be only MKV.

On a side note I am thinking when time comes you will be able to offer 4K bluray support since tsmuxer already works with 4K and hevc.

May I ask which hevc encoder you are using and build if possible.
__________________
If you fail to plan; you plan to fail, would you not agree? Think about it.

Last edited by HWK; 14th September 2014 at 02:23.
HWK is offline   Reply With Quote
Old 14th September 2014, 07:35   #21094  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 4,041
Quote:
Originally Posted by Lathe View Post
Is there really a 'Conflict'? I'm still VERY green at this, so I'm probably just missing something important here, but just going by what I've learned so far, that is how I understand it.
You can't wildly play with special settings for pass 2. Some are ok, others must be identical to pass 1. Let BD-RB do it's job without any "smart tweaks" which will disqualify for a BD-RB bug report -- see jdobb's statement in this respect.
Sharc is offline   Reply With Quote
Old 14th September 2014, 14:25   #21095  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by HWK View Post
I am assuming jdobbs at this stage it gone be only MKV.

On a side note I am thinking when time comes you will be able to offer 4K bluray support since tsmuxer already works with 4K and hevc.

May I ask which hevc encoder you are using and build if possible.
X265. There are daily builds, so I just grab one every now and then.

Yeah, I'm hoping to get ahead of the power-curve on 4K bluray (and HEVC at HD res). Right now I'm just implementing MKV support.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 14th September 2014 at 14:36.
jdobbs is offline   Reply With Quote
Old 14th September 2014, 20:30   #21096  |  Link
daberti
Registered User
 
Join Date: Apr 2004
Posts: 172
Quote:
Originally Posted by jdobbs View Post
X265. There are daily builds, so I just grab one every now and then.

Yeah, I'm hoping to get ahead of the power-curve on 4K bluray (and HEVC at HD res). Right now I'm just implementing MKV support.
You're already doing a very good job...take your time JD
daberti is offline   Reply With Quote
Old 14th September 2014, 22:13   #21097  |  Link
worknstiff
Registered User
 
Join Date: Jun 2009
Posts: 314
RE: Yeah, I'm hoping to get ahead of the power-curve on 4K bluray (and HEVC at HD res).

I think they are going to have to make some bigger bluray's for 3D @ 4K + dolby atmos, lol.
worknstiff is offline   Reply With Quote
Old 15th September 2014, 00:35   #21098  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by worknstiff View Post
RE: Yeah, I'm hoping to get ahead of the power-curve on 4K bluray (and HEVC at HD res).

I think they are going to have to make some bigger bluray's for 3D @ 4K + dolby atmos, lol.
I'm sure at some point you'll see BD-XL with 100GB become more common. I'm using a BDXL-RE now and it's pretty handy having that much storage on a single disc.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Old 15th September 2014, 01:05   #21099  |  Link
Lathe
Registered User
 
Lathe's Avatar
 
Join Date: Aug 2005
Posts: 1,268
Quote:
Originally Posted by Sharc View Post
You can't wildly play with special settings for pass 2. Some are ok, others must be identical to pass 1. Let BD-RB do it's job without any "smart tweaks" which will disqualify for a BD-RB bug report -- see jdobb's statement in this respect.
Okidoke. No more stray 'Tweak'n' reports here then
Lathe is offline   Reply With Quote
Old 15th September 2014, 01:47   #21100  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 21,108
Quote:
Originally Posted by jdobbs View Post
Hmm... I'll have to try the Directshow filter I'm using, but I don't think has the ability to demux from M2TS. Of course that wouldn't prevent another type of playback... but for now I'll probably limit it to MKV.
Quote:
Originally Posted by HWK
Yes, it can. I recently ran test just for you.
I tried it. I muxed a H265 M2TS with TSMUXER and then tried to view it in Media Player Classic using the directshow filter. It crashed MPC. What player are you using?
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Reply

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 02:37.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2025, vBulletin Solutions Inc.