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 > One click suites for DVD backup and DVD creation

Reply
 
Thread Tools Search this Thread Display Modes
Old 30th January 2016, 00:33   #2901  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
Hi MrC and SD,

did a few more tests, and it sure has nothing to do with permissions - I am using Win XP.

The user from VideoHelp is right IMO, the issue only appears if the destination for MuxMan is on a different drive than the drive where the program resides.

In my case my system drive is E:, and this is where my program files folder is located. The only way to avoid the error is to specify a folder on E: as my output folder in AVStoDVD. Using any other drive as the destination causes MuxMan to crash.


Cheers
manolito
manolito is offline   Reply With Quote
Old 30th January 2016, 23:00   #2902  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Ok, let's give a try to this one (BatchMux v1.4.1, hotfix, beta):

http://www93.zippyshare.com/v/8ZFyL4Lq/file.html

Once you download it, it is intended as a direct replacement for the version 1.4 of BatchMux.exe, located in the AVS2DVD programs folder.

In case issues are still present with this version, please let me know...
Sir Didymus is offline   Reply With Quote
Old 31st January 2016, 12:14   #2903  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
Thanks SD, this hotfix works here. Beautiful...

Cheers
manolito
manolito is offline   Reply With Quote
Old 31st January 2016, 17:00   #2904  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
OK! Great... Thanks, manolito, for the very rapid testing & report. Let's wait for a while for possible feedbacks from other users, incurred in other troubles related to the same bug, before considering the problem solved. OK, another thing to consider for the future is that obviously the already released BM versions 1.5 and 1.6 are from now on deprecated. Clearly, there is no urgency to provide new versions of BM 1.5 and 1.6. It would be not even wise, since maybe it is better to wait for other possible bugs & issues in the current BM 1.4.1; the relevant features of BM 1.5 and 1.6 are already known (from the other thread dedicated to BM), so whenever MrC requires, their updates will be distributed...

Quote:
... Beautiful ...
Mhhh, yes!
How to disagree?

Cheers,
SD
Sir Didymus is offline   Reply With Quote
Old 1st February 2016, 09:01   #2905  |  Link
Deach
Registered User
 
Join Date: Jan 2016
Posts: 3
Can confirm the issue is fixed on my end too.
Deach is offline   Reply With Quote
Old 1st February 2016, 09:59   #2906  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Oh, thank you Deach, for your report, and welcome in the doom9 community.
:-)
Sir Didymus is offline   Reply With Quote
Old 1st February 2016, 11:13   #2907  |  Link
Deach
Registered User
 
Join Date: Jan 2016
Posts: 3
Quote:
Originally Posted by Sir Didymus View Post
Oh, thank you Deach, for your report, and welcome in the doom9 community.
:-)
Cheers. I was the user MrC was talking about earlier, had to wait 5 days though obviously to be able to post!

Relevant I suppose at the minute, is there much performance benefit to having temp files assets and whatever else across difference drives?
Deach is offline   Reply With Quote
Old 1st February 2016, 13:20   #2908  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Ah, I see! Good to know...

Yes, at least partially. The authoring engine (MuxMan) and the video encoder (HcEnc, for instance) are the most time-consuming components in AVS2DVD, I think... The encoder especially...

Both are heavily using I/O operations to read large chunks of input assets and produce equally big output files. So, reading from one drive and writing to a different (physical) one is surely beneficial. I do not have recent figures on the matter, to tell you frankly; furthermore nowadays SSD are quite diffused, so this benefit of distributing the I/O jobs on different drives is not as huge as it was some years ago...

There is also to consider that both encoding and authoring, in AVS2DVD, are single thread modules, using in a very intense way just one of the available CPU cores, so the gain you can obtain will be in any case bounded by the CPU, at a given point...

Last edited by Sir Didymus; 1st February 2016 at 13:23.
Sir Didymus is offline   Reply With Quote
Old 1st February 2016, 13:39   #2909  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
Quote:
Originally Posted by Sir Didymus View Post
There is also to consider that both encoding and authoring, in AVS2DVD, are single thread modules, using in a very intense way just one of the available CPU cores, so the gain you can obtain will be in any case bounded by the CPU, at a given point...
Hi SD,

this is not true for HCenc and FFMpeg...
AVStoDVD configures both encoders so they will use all available CPU cores (at least if you have configured it this way under Preferences->System).

But for my encodes the CPU load is still only at about 30% because the real bottleneck is AviSynth and the filters I use. These are all single threaded, and switching to AviSynth MT or AviSynth+ and using only MT plugins is something I do not dare right now.


Cheers
manolito

Last edited by manolito; 1st February 2016 at 13:43.
manolito is offline   Reply With Quote
Old 1st February 2016, 16:32   #2910  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Ah, I was not aware about that... Do you know if this is achieved by the controlling environment (e.g. AVS2DVD is splitting the overall work into multiple individual encoding jobs) or these two encoders are real multithread applications (allowing to achieve one single encoding job to be completed individually by using multiple threads)?
Sir Didymus is offline   Reply With Quote
Old 1st February 2016, 18:51   #2911  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
It is entirely the encoders...

From the A2D manual:
Quote:
If you have a multi-processor / multi-core CPU and you use this option, video encoding speed will be improved a lot.
HCenc multithread encoding is activated by "*SMP" command in HCenc.ini file
FFmpeg multithread encoding is activated by "-threads n" command line flag, where n is the number of the cores used.
Cheers
manolito
manolito is offline   Reply With Quote
Old 1st February 2016, 21:13   #2912  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Very relevant features indeed (for both encoders...). Regarding HCenc I already clarifyed to me any doubts, via the related FAQ, explaining its multiprocessing feature. That's great!

So, forgive me the wrong statement above...

Well, back to the question about the usefulness of splitting encoding and authoring activities by keeping on different physical volumes the related input and output assets, I am still positive that there are tangible benefits by selecting different volumes for the reading and for the writing operations.

However, me too curious about some recent evaluations about this aspect, with modern HW configurations, including the usage of SSD devices...

Cheers,
SD
Sir Didymus is offline   Reply With Quote
Old 16th February 2016, 18:16   #2913  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
It's quiet here... The software seems to be working perfectly

Well, I don't want to spoil the party, but I did find a small issue with chapter creation. It only happens for muxed MPEG2 source files. I have set an interval for automatic chapter creation and also ticked the checkbox to prefer the source chapters if present.

In this case the fixed chapters routine should take over (MPG files cannot have chapters). But it doesn't, no chapters will be created. I believe that the reason is that MediaInfo reports one empty chapter for these files. This is the source info from AVStoDVD:
Quote:
GENERAL INFO
PathName: I:\test.mpg
Format: MPEG-PS
Time: 2:54:25 hours
Audio Tracks: 1
Subs Tracks: 0

VIDEO TRACK INFO
PathName: I:\test.mpg
Track ID: 224
Stream Order: 0
Time: 2:54:25 hours
Compression: MPEG-2 Video
BitRate Mode: VBR
Avg BitRate: 3256 kbps
Min BitRate:
Max BitRate:
Resolution: 720x576
Display AspectRatio: 1,778
Pixel AspectRatio: 1,422
Frame AspectRatio: 1,25
FrameRate Mode:
FrameRate: 25 fps
Frame Count: 261636
Scan Type: Interlaced
Interlaced: Yes
Field Order: TFF
Color Space: YUV
Colour Coefficients:
DVD Compliance: 1

AUDIO TRACK INFO (Track 1)
PathName: I:\test.mpg
Track ID: 189-128
Stream Order: 1
Time: 2:54:25 hours
Compression: AC3
BitRate Mode: CBR
Avg BitRate: 224 kbps
Channels: 2
SampleRate: 48000 Hz
Resolution:
Language:
Delay: 0 ms

CHAPTERS INFO
Chapter 1: Count

This MediaInfo output is not correct IMO, but it has been there forever (I went back to version 0.7.58). So I think that A2D should just assume that there never is any source chapter info for MPEG2 files.

BTW if I first demux the MPG file and feed the elementary streams to A2D, the chapters will be created correctly.


Otherwise the latest alpha version with the fixed BatchMux version is in frequent use here, and for me it is absolutely stable.


Cheers
manolito
manolito is offline   Reply With Quote
Old 26th February 2016, 00:28   #2914  |  Link
Deach
Registered User
 
Join Date: Jan 2016
Posts: 3
Quote:
Originally Posted by manolito View Post
Hi SD,

this is not true for HCenc and FFMpeg...
AVStoDVD configures both encoders so they will use all available CPU cores (at least if you have configured it this way under Preferences->System).

But for my encodes the CPU load is still only at about 30% because the real bottleneck is AviSynth and the filters I use. These are all single threaded, and switching to AviSynth MT or AviSynth+ and using only MT plugins is something I do not dare right now.


Cheers
manolito
Would there be much of a benefit from other AviSynth forks? I've tampered with both + and the MT fork briefly but I couldn't say for sure if there was any noticeable improvement based on the odd few encodes I'd run with them installed? I suppose it depends on the project settings as well.

Recent BatchMux working fine here still too, no issues to report!
Deach is offline   Reply With Quote
Old 27th February 2016, 09:53   #2915  |  Link
Sir Didymus
Registered User
 
Join Date: Mar 2004
Location: Italy
Posts: 948
Ok. Regarding BM, thanks manolito and Deach for your confirmation. Your support is great! With low priority, in the next days/weeks I will update the thread dedicated to the BM development with some related information, just for future tracing purposes [...hem, especially to remind myself about the "status quo", since three versions of BatchMux are affected: 1.4.1, 1.5.1 and 1.6.1... :-( ...].

manolito, concerning the other small glitch (chaptering based on MPG container) I totally agree with your considerations and proposed solution. It's a relatively minor issue, indeed, since the backup of unticking the checkbox to prefer the source chapters allows to live with the problem...

About speeding up the AviSynth operations, IMO things are definitely not simple to deal with. The (radically different) applicative approach "A la DVD Rebuilder", of dividing the source assets in chuncks (titles, chapters or segments), with one CPU dedicated to the operations for each individual chunk, would be probably beneficial. But it is just a matter of speculations, since in the context of AVS2DVS is would be probably almost impossible to implement (again IMHO, these conjectures here are just "free wheel thinking") and the B/C ratio would be for sure extremely low...

Maybe exploring furthermore the usage of AviSynth MT or AviSynth+, strictly with MT plugins, would give higher benefits, with a relatively minor effort, if the throughput of the job is the priority...

Last edited by Sir Didymus; 27th February 2016 at 10:20.
Sir Didymus is offline   Reply With Quote
Old 10th March 2016, 18:23   #2916  |  Link
MrC
AVStoDVD Dev
 
MrC's Avatar
 
Join Date: Apr 2006
Location: Italy
Posts: 1,280
Guys, I'm not dead... maybe just zombified by my job... within few days I should publish a new alpha with some minor improvements (BM 1.4.1, tnx SD, more subs tools) and bugfixes (mpeg2 auto chapters issue, tnx manolito, AVS join titles chapters flags with fps conversion, tnx davexnet from VH forum).



Bye
__________________
MrC

AVStoDVD Homepage
AVStoDVD @ Doom9 Forum
MrC is offline   Reply With Quote
Old 10th March 2016, 21:07   #2917  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
Quote:
Originally Posted by MrC View Post
Guys, I'm not dead... maybe just zombified by my job...
This is very reassuring - I was almost ready to email you inquiring about your health...


Take care
manolito
manolito is offline   Reply With Quote
Old 13th March 2016, 11:53   #2918  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
Problem with default FFMpeg HQ parameters

By default AVStoDVD uses the following HQ parameters when the FFMpeg HQ mode is activated:
-bf 2 -mbd rd -trellis 2 -cmp 2 -subcmp 2

These parameters are officially recommended, see here:
https://ffmpeg.org/faq.html#Which-ar...PEG_002d2_003f

But at least for current versions of FFMpeg the "trellis" parameter causes heavy artifacts with a lot of sources. It looks like the background is "pulsing", the reason is that all I-frames are bitrate-starved, i.e. their quantizer is way too high.

Here is an example source file:
http://www67.zippyshare.com/v/PpKnauBU/file.html

Even with the highest bitrate these artifacts are clearly visible. Removing the "-trellis 2" from the HQ parameters fixes the problem.


Thanks to SilverSolver from the DVDStyler forum for discovering the issue and being a great help in finding the culprit.

I recommend to everyone to remove the "trellis" parameter from the HQ FFMpeg parameter list...


Cheers
manolito
manolito is offline   Reply With Quote
Old 15th March 2016, 13:37   #2919  |  Link
MrC
AVStoDVD Dev
 
MrC's Avatar
 
Join Date: Apr 2006
Location: Italy
Posts: 1,280
Thanks manolito for the hint



Bye
__________________
MrC

AVStoDVD Homepage
AVStoDVD @ Doom9 Forum
MrC is offline   Reply With Quote
Old 19th March 2016, 22:35   #2920  |  Link
manolito
Registered User
 
manolito's Avatar
 
Join Date: Sep 2003
Posts: 2,262
New versions of my plugins for AVStoDVD

I just uploaded new versions of my AVStoDVD plugins, here's the link:
http://www72.zippyshare.com/v/vQjJSXBw/file.html


Changes:

Wavi_Mod now uses the latest version 0.4.9 of BS1770Gain

ffmpeg_vbr features some usability improvements.

1. The bitrate thresholds for switching between different encoding modes and the thresholds for switching between Custom Quant Matrices can now be specified independently.
2. The 1-pass VBR and 2-pass VBR modes can now be forced regardless of the video bitrates.
3. In forced 2-pass VBR mode the desired HQ mode can now be specified (Normal Quality, Regular HQ, Extreme HQ).
4. In the Extreme HQ mode the MPEG Adapted matrix and a DC precision of 10 are forced.


The plugins were tested under WinXP and Win7 (32-bit and 64-bit) and AviSynth 2.60. I used the latest FFMpeg point release 3.0.



Cheers
manolito

Last edited by manolito; 26th March 2016 at 17:10. Reason: Various small quality improvements
manolito 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 Off
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 21:34.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2018, vBulletin Solutions Inc.