EPTHere's some code I whipped up
Run vdub and edit the variables to reflect your file locations, but if you have a file full of AVIs that are 4GiB that Fraps outputs this will run 4 instances of conversion at the same time.
Essentially it means you'll transcode in 1/4 of the time, each ~2 minute capture takes me about 440 seconds to transcode to a nice avi, I have some filters to reduce the resolution (need to experiment with more later) but if you have Perl and the modules installed and run this you'll find nothing faster at making great fast threaded transcodes of your Fraps castings.
I warn you, I have a quad core (Phenom II 940BE) and it's killed my computer more than anything I've done before. DO NOT go above the amount of cores your computer has, you'll probably die.
I actually wanted real time transcoding as I made the captures but I can't work the -r and -w options (check if file is read / writable) with Windows, it just seems to fail and opens the file and dies.
Anyway I thought any of you casters / programmers (not progamers, horribly close those two words) might find this script handy.
![[image loading]](http://i.imgur.com/BfUDm.png)
As you can see, with my settings, each thread will use almost 800MiB of RAM and 100% of one processor. If you're lucky enough to have one of the 6-core procs out there and 6GiB of RAM then you'll get beautiful speeds.
You do have to disable GPU optimizations in vdub unfortunately as your GPU simply can't do it. It would be possible to do one thread with GPU processing, I may do that just for the slight improvement in one thread.
Hope someone else finds a use for this! It does mean 400 seconds for 8 minutes of files, and this is HDD to the same HDD and it's only a Samsung F1, nothing special!
I'll keep working on the real-time transcoding, and I intend to add features to use blip.tv's FTP features to upload on completion as well as joining of the AVI files, that can most likely be done with more tuning though.
I wrote the script as legibly as I possibly could in Perl, it shouldn't be hard to understand. You will need the Paralell::ForkManager module installed.
