Postby Competern » 20 Mar 2012 11:54
Since VLC media player will play back at twice speed (audio and video) you should be able to use Snapzxpro on a mac to record video from the screen during play-back (at twice or whatever speed) and then export that. SnapzXPro is not free, but it I've found it good. You say that you can use Mac - good, because I understand that SnapzXPro is only for Mac - if not, someone please correct me. Incidentally Apple Technical Support told me to use VLC Media player when my then new MacBook Pro wouldn't play some DVDs using Apple supplied software.
As I understand it, speeding up or slowing down the video is fairly easy - to slow down, frames are repeated; so speed up frames are omitted. But however it is achieved, VLC Media player does this for you.
VLC Media player also does the audio processing for you, but it helps to know what the processing does so that you understand the limitations of speeding-up or slowing down audio. If the speed change is done by simply changing the clock/sampling rate that changes the pitch. To prevent pitch-change, the audio must be played back at normal speed (with normal clock/sampling rate) and short sections of audio must be snipped out to speed up: or repeated to slow down. The hard part, the art of it, is to snip out, or add in, short sections of audio such that the audio-joins are effectively silent, while the overall audio remains in sync with the video. To do this, the software has to match the cuts so that the magnitude and slope of the audio waveform matches on each side when the audio sections are spliced together. When this is done properly, the join is inaudible. If it is done poorly, it will sound noisy.
In practice audio is often very complex with different unrelated sounds and frequencies all mixed together. Hence, even though all individual splices in the audio may have no click or noise, and so the dominant theme or speech sounds OK, the background themes, background speech or background noise tend to sound weird as they wax and wane in an unpredictable manner.
The net result is that audio processing to keep the same pitch works reasonably well for clear speech, especially when there is only one person speaking and no background noise. But when the audio is complex with speech or singing plus music or other speech in the background, as in your intended use, the end result may sound strange and the speed-shifted audio quality may be noticeably worse. This is not a fault in the audio processing - it is simply a reflection of the fact that there is no perfect solution for a complex mixed-audio original.
So try out the sped-up playback first: check that the audio quality is acceptable to the intended audience: and then decide if you want to invest in Snapz Pro X to capture and then export the sped-up playback.
If the speed-change technique was done professionally, the way forward would be to record multi-track so that each track consists of a dominate theme with no background noise, so each track would be spliced in different places and then the different tracks mixed after the speed change. That's a mammoth task - but it helps you understand the problems.