Take care when testing: by opening the AVS-script the whole file gets scanned in order to return the exact sample number when normalize=false (default) this is done with partial decoding (~20sec for one hour audio on my PC). When normalize=true it is much slower (~100sec).
Version 0.2 creates a small file .D2A after the first scan, so there is no multiple scanning necessary.
When backward jumps are requested, the seeking is always done in fast mode from frame 0 to the requested frame.
Along with the audio, a small 8x8 video is generated, so you always have to put MPASource as SECOND argument in AudioDub.
LoadPlugin("C:\myprojects\mpasource25\release\mpasource.dll")
V = BlankClip(height=100, width=100, length=20000, fps=25)
A = MPASource("d:\_mp3\pop\liquido - narcotic.mp3", normalize = false)
AudioDub(V, A)$Date: 2004/08/13 21:57:25 $