Is there 2GB file size limit for WAV files processed by AviSynth?

[Silicon]

@FranceBB: OK thanks for letting us know. So I'm gonna use WAVsource in my AviSynth scripts from now

[FranceBB]

I'm gonna use WAVsource in my AviSynth scripts from now

Nice.

[Silicon]

@FranceBB
Just out of curiosity, I did a test with AviSynth script using WavSource and it seems to me it has the same problem with > 2GB WAV file as ffms2.
The AviScript is as follows:

m_clip=WavSource("%s_original_full%")

A1=GetChannel(m_clip, 1)
A2=GetChannel(m_clip, 2)
A8=GetChannel(m_clip, 8)
Odd=MixAudio(A1, A8, 1.0, 1.4)
Even=MixAudio(A2, A8, 1.0, 1.4)
m_clip=MergeChannels(Odd, Even)

# The following MUST be the last line of your script
Return m_clip

And the workflow is simple

8-channel WAV to stereo MP3 WF.PNG (6.75 KiB) Viewed 6074 times

What I'm missing here? If you say there is A/V decoder missing in front of AviSynth, I would replay "it didn't work at all".

P.S. I'm using Custom FFmpeg node to achieve desired outcome, but i just wanted to test the AviSynth method as well.

[emcodem]

Hehe the proposed workaround and avs script was coming from me:D
Anyway, i am not available the next weeks so @FranceBB to the help!

I just retried it and had no problem, the worfklow seems to be ok this way too.

[FranceBB]

WAVSource doesn't have any problem on my side.
About A/V Decoder, that's not needed as it's only needed if you want to use indexers and open the file inside Avisynth and use a Custom Avisynth Script.
I'll tell you what: first you should get the script working in a normal IDE like AVSPmod and test it with either VirtualDub x64 or with FFMpeg opening the AVS Script.
There's one thing, though: you cannot index an audio file without a video, otherwise Avisynth is gonna return an error at compile time, so you're gonna need BlankClip().

Can you tell me exactly what your source file is and what you're trying to achieve?

'cause if we can avoid going to PCM and index the source directly, that's even better.

With a mockup example I can come up with an AVS Script and a BAT and help you properly.
This sounds more like an Avisynth question than an FFAStrans one eheheheh

[Silicon]

Hi FranceBB
The source file is 8-channel WAV exported from our news production system (these 8 audio channels are extracted from XDCAM HD422/MXF file).
The goal is to get audio channel 8 (containing Audio Description), boost it by +3dB and then add it to audio channels 1 and 2 (and encode to mp3).

[FranceBB]

Well but then it's easier than expected.
This should definitely work:

Code: Select all

video=LWLibavVideoSource("\\10.64.6.11\avisynth\Scambio\test.mxf")
ch1=LWLibavAudioSource("\\10.64.6.11\avisynth\Scambio\test.mxf", stream_index=1)
ch2=LWLibavAudioSource("\\10.64.6.11\avisynth\Scambio\test.mxf", stream_index=2)
ch8=LWLibavAudioSource("\\10.64.6.11\avisynth\Scambio\test.mxf", stream_index=8).AmplifydB(3)

stereo1=MergeChannels(ch1, ch2)
stereo2=MonoToStereo(ch8, ch8)
audio=MixAudio(stereo1, stereo2)
AudioDub(video, audio)

At least, it does for me:




Then you just need to feed this "AVS Script.avs" to FFMpeg and you're good to go.


p.s If I may, AAC is nowadays as compatible as MP3 but achieves a much better quality due to the fact that it's not using the old poly-phase filter-bank which was implemented back when MP3 was around.

[FranceBB]

Fixed in ffms3. Release pending as I have to test other things not related to this as well, but if everything works I'll submit it to grandmaster for FFAStrans 1.2.2.
Don't hold your breath, though.
So far the recommended way is to use LSMASH with LWLibavVideoSource() and LWLibavAudioSource()