Speed improved with just AviSource+Prefetch?!

[shae]

On AviSynth+, I'm seeing speed improvement with the script:


Import("MtModes.avsi")
AviSource(...)
Prefetch(...)


Which is fed directly to x264.

Is this because x264 by default doesn't spawn a new thread for AVS, or is there simultaneous decoding of multiple frames (which I thought would anyway be insignificant relative to x264)?

[poisondeathray]

You can test which settings are the fastest with avsmeter (look at cpu usage and fps). That should translate to the script version resulting in the fastest x264 encoding . For a single QTGMC call, the fastest for me processing SD was prefetch(6)

[shae]

The surprising thing here is that there's no processing, only AviSource().