Selur's Little Message Board

Full Version: Vaporsynth QTGMC bug in ver. 2017.10.05.1
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2

andrewschen

(29.12.2017, 17:50)Selur Wrote: [ -> ]Did a fast test her, by reencoding a 7.1 24bit 96kHz stream to flac 24 bit 48kHz.
The encoding call that was used, was:
Code:
ffmpeg -y -threads 8 -loglevel fatal -i "F:\TestClips&Co\files\audio\DTS Master Audio, 7.1 channels, 24 bits, 96kHz.dtsma" -ac 8 -ar 96000 -sample_fmt s32 -f sox - | sox --multi-threaded --ignore-length --temp "H:\Temp\16_46_44_251001" --buffer 524288 -S -t sox - --endian little -b 24 -t wav - rate 48k | flac -f -F --ignore-chunk-sizes -5 -o "H:\Temp\iId_1_aid_0_16_46_44_2510_02.flac" -
and encoding worked fine here.
Also tried the same with flac 24 bit 96kHz.
Code:
ffmpeg -y -threads 8 -loglevel fatal -i "F:\TestClips&Co\files\audio\DTS Master Audio, 7.1 channels, 24 bits, 96kHz.dtsma" -ac 8 -ar 96000 -sample_fmt s32 -f sox - | sox --multi-threaded --ignore-length --temp "H:\Temp\16_49_21_801001" --buffer 524288 -S -t sox - --endian little -b 24 -t wav - | flac -f -F --ignore-chunk-sizes -5 -o "H:\Temp\iId_2_aid_0_16_49_21_8010_02.flac" -
also no crash here.
-> so I really need the debug output to know why it isn't working for you.

Cu Selur

I checked the 2017.10.5.1 and 2017.5.6.1 version after I checked the current version.

On my system, 2017.5.6.1 can successfully encode flac24.

For 2017.10.5.1 and 2017.12.27.1, the program begins with the flac24 encode then soon reports finishing encoding, but doesn't really generate the flac24 audio for muxing to destination file, so the muxing process reported that it can't find the flac24 file and the job is crashed.

Here's the debug output of the 3 version of Hybrid I checked.
Only looked at the output of 2017.12.27.1 (don't care about old versions).

According to the debug output:
Code:
"C:\PROGRA~1\Hybrid\mkvmerge.exe" --ui-language en -o "T:\flac32stereo.mkv" --global-tags "H:\TMP\USERTMP\flac32stereo_03_38_37_6310__03.xml" --default-track 0:yes --forced-track 0:no -a 0 --compression -1:none --no-video --no-subtitles --no-chapters "H:\TMP\USERTMP\iId_5_aid_0_03_38_37_6310_01.flac"
is called, finished after 00:00:11.856 and created T:\flac32stereo.mkv (1112.91 MB)
-> looks fine to me if you want a raw flac file as output, use '.raw' as extension in the Main-tab,...

Cu Selur

andrewschen

(29.12.2017, 23:05)Selur Wrote: [ -> ]Only looked at the output of 2017.12.27.1 (don't care about old versions).

According to the debug output:
Code:
"C:\PROGRA~1\Hybrid\mkvmerge.exe" --ui-language en -o "T:\flac32stereo.mkv" --global-tags "H:\TMP\USERTMP\flac32stereo_03_38_37_6310__03.xml" --default-track 0:yes --forced-track 0:no -a 0 --compression -1:none --no-video --no-subtitles --no-chapters "H:\TMP\USERTMP\iId_5_aid_0_03_38_37_6310_01.flac"
is called, finished after 00:00:11.856 and created T:\flac32stereo.mkv (1112.91 MB)
-> looks fine to me if you want a raw flac file as output, use '.raw' as extension in the Main-tab,...

Cu Selur

Allow me to explain...

I tested each hybrid version with more than 1 job, flac16/24/32 all tested. the error is only happen on flac24 encode, and the output results is located in the middle of HybridDebugOutput 2017.12.27.1.txt, flac24stereo is the name of crashed file/job, not flac32stereo. 

Sorry for confusing. Blush
Ah, problem is probably related to you using flacCL (which I never use Smile).
Might be that it doesn't support 24bit encoding.
What happens if you call:
Code:
"C:\PROGRA~1\Hybrid\ffmpeg.exe" -y -threads 4 -i "H:\sample.thd" -ac 8 -ar 48000 -f sox - | "C:\PROGRA~1\Hybrid\sox.exe" --multi-threaded --ignore-length --temp "H:\TMP\USERTMP\03_38_05_901001" --buffer 524288 -S -t sox - --endian little -b 24 -t wav - remix -m 1v0.243804,7v0.062798,3v0.172392,6v0.044404,8v0.044404,5v0.432198 2v0.243804,7v0.062798,3v0.172392,5v0.044404,8v0.044404,6v0.432198 | "C:\PROGRA~1\Hybrid\CUETools.FLACCL.cmd.exe" -f -F --ignore-chunk-sizes -8 -o "H:\TMP\USERTMP\iId_4_aid_0_03_38_05_9010_02.flac" -
inside a 'Windows Command Prompt' ?

Hybrid atm. doesn't check whether the output of FLacCL was created, since I totally forgot about it. Big Grin

Cu Selur
Looking at the FLACCL command line options:
Code:
CUETools FLACCL 2.1.6, Copyright (C) 2010-2013 Grigory Chudov.
This is free software under the GNU GPLv3+ license; There is NO WARRANTY, to
the extent permitted by law. <http://www.gnu.org/licenses/> for details.
Usage    : CUETools.FLACCL.exe [options] <input.wav>

Options:

-0 .. -11             Compression level, default 8; 9..11 require --lax
-o <file>             Output filename, or "-" for stdout, or nul
-p #                  Padding bytes
-q --quiet            Quiet mode
--lax                 Allow non-subset modes
--verify              Verify during encoding
--no-md5              Don't compute MD5 hash
--no-seektable        Don't generate a seektable
--ignore-chunk-sizes  Ignore WAV length (for pipe input)
--cpu-threads         Use additional CPU threads

OpenCL Options:

--opencl-type <X>     CPU or GPU, default GPU
--opencl-platform <X> "NVIDIA CUDA", "Intel(R) OpenCL"
--group-size #        Set GPU workgroup size (64,128,256)
--task-size #         Set number of frames per multiprocessor, default 8
--slow-gpu            Some encoding stages are done on CPU
--fast-gpu            Experimental mode, not recommended
--define <X> <Y>      OpenCL preprocessor definition

Advanced Options:

-b #                  Block size
-s <method>           Stereo decorrelation (independent,search)
-r #[,#]              Rice partition order {max} or {min},{max} (0..8)

LPC options:

-w <func>[,<func>]    Window functions (bartlett,welch,hann,flattop,tukey)
-l #[,#]              Prediction order {max} or {min},{max} (1..32)
    --max-precision    Coefficients precision search (0..1)
It doesn't seem like there isn't an option to use 24bit with FlacCL. Sad

Cu Selur
Okay, I'm a bit confused how you even managed to get Hybrid to use flacCL with 24 bit encoding, because for me as soon as FLACCL is used, there 16bit is enforced.
Okay, I figured it out.
I through flaccl out as the default flac audio encoder, but you set it manually in the tools section, by default Hybrid uses flac.exe there.
And now comes the bug, normally Hybrid should remove '24bit' support when flaccl is used and this fails. Wink
->FLACCL only supports 16 and 32bit not 24bit and Hybrid should hide the 24 from the user, but here it fails. Wink

Cu Selur
After some sleepless hours,..
a. flac through flaccl only supports 16bit
b. flac through ffmpeg only supports 16 and 32bit
c. flac through flac supports 16/24/32bit
-> fixed that + a bug during the ffmpeg call optimization in the dev build I send you a link to in a few minutes.

Cu Selur

andrewschen

(30.12.2017, 03:28)Selur Wrote: [ -> ]After some sleepless hours,..
a. flac through flaccl only supports 16bit
b. flac through ffmpeg only supports 16 and 32bit
c. flac through flac supports 16/24/32bit
-> fixed that + a bug during the ffmpeg call optimization in the dev build I send you a link to in a few minutes.

Cu Selur

I didn't enable flaccl actually.....

But I tested the new fix you send me, and it does solve the problem!

Thanks you! Smile
Pages: 1 2