AviSynth Audio Overlay

[legocrazy1]

Well, I have yet another small question about avisynth, except this one is much simpler in my opinion.

Basically I am wanting to add a soundclip on top of any audio that might already exist at a specified time interval.


Is there some kind of function or plugin that would allow me to do this easily? If there is not one I'm guessing I will have to add some silence before the audio clip and then mix the two audio files together?

Thank You,
Adam

[AlanHK]

http://avisynth.org/mediawiki/MixAudio

But yes, only mixes the entire track, no offset.

Use Tone() to generate some silence:
http://avisynth.org/mediawiki/Tone


But you need a video track to use most functions, so maybe BlankClip would be simpler

#Overlay Comments.mp3 (length 1000 frames) starting at 2000 frames offset.
AviSource ("Original.avi")
voiceover=BlankClip(last,2000) ++BlankClip(last, 1000).AudioDub(DirectShowSource("comments.mp3")))
MixAudio(last,voiceover)

Need to calculate the durations in frames; there are ways to do that in Avisynth if you need to do it often.

[legocrazy1]

Yes, but that only allows you to specify volumes, not actual time positions. That was what I was going to use if such a function doesn't exist.

[AlanHK]

I expanded my answer, see revised post.

[legocrazy1]

Okay, that was quite informative. Thank you for your assistance.

[Gavino]

But you need a video track to use most functions, so maybe BlankClip would be simpler

#Overlay Comments.mp3 (length 1000 frames) starting at 2000 frames offset.
AviSource ("Original.avi")
voiceover=BlankClip(last,2000) ++BlankClip(last, 1000).AudioDub(DirectShowSource("comments.mp3")))
MixAudio(last,voiceover)

The voiceover clip doesn't need a video track, so you can just use:

#Overlay Comments.mp3 (length 1000 frames) starting at 2000 frames offset.
AviSource ("Original.avi")
voiceover=DirectShowSource("comments.mp3")
Trim(0, 1999)++Trim(2000, 2999).MixAudio(voiceover)++Trim(3000, 0)

A more serious problem with your version is that the sound track outside the voiceover section would be reduced in volume by 50% (since mixed with silence).

[AlanHK]

A more serious problem with your version is that the sound track outside the voiceover section would be reduced in volume by 50% (since mixed with silence).

Right. Your method is better.*
Sorry you have to keep correcting me.

Actually, the only times I really use Avisynth mixing is to downmix to stereo or mono.

For overlaying commentary tracks I use Audacity; so I was thinking like that, as mixing silence + track in Audacity causes no change in volume.

Perhaps this method, which I actually use, might be helpful:

For some TV shows I found a full-length podcast commentary online; if I cue it right it applies to what you see on screen. So I export the audio from the video, and then in Audacity use the "Compressor" filter to knock the volume down to -29 dB (works for me); then add the commentary track; mix them to make an alternate audio track. The original audio is still audible in the background but the commentary is clear.

* Though one possible problem with dubbing a bare mp3 is that if it's VBR, it quickly gets out of sync -- Avisynth doesn't know how long it really is. If CBR then it's okay. I use CBR MP3 for patching audio (replacing advertising VO with music on TV shows, for instance).

[legocrazy1]

Thank You both. I will probably just use audacity for a little bit, but I may tweak around with the suggested avisynth code, and see where it brings me.