Hi all, and Happy New Year!
I might be about to display my ignorance here with my first post but here goes...
When playing back 23.976fps media on my system the actual refresh rate reported by ReClock is 23.949fps. This means that the audio needs to be slowed down slightly for smooth playback.
The accepted method would seem to be to drop the rate of the PCM stream and resample. Fair enough. In this case though would it not make more sense to use audio timestretching instead? This would avoid any resampling artifacts, pitch change (not that I'd notice it in my example) and CPU usage required for resampling (I'm getting close to the limit for 1080p playback on my system).
I'm making a few assumptions here:
1. Timestretching works in a similar (or maybe identical) way to the SoX tempo effect, which uses WSOLA, and works very well in my own testing.
2. Timestretching and resampling are mutually exclusive options. If I enable timestretching it disables resampling.
3. Timestretching is more CPU friendly than resampling. It should be if my first assumption is correct.
I think my assumptions might be wrong though. With audio timestretching enabled CPU usage still seems to vary depending on the resampling quality setting, suggesting that resampling is still going on. If this is the case could anyone explain to me what the audio timestretching setting is actually doing please? The readme just says that it allows the media speed to be changed without altering the pitch, but doesn't say how this is achieved.
Thanks very much (and thank you for continuing with ReClock - it's invaluable).
I might be about to display my ignorance here with my first post but here goes...
When playing back 23.976fps media on my system the actual refresh rate reported by ReClock is 23.949fps. This means that the audio needs to be slowed down slightly for smooth playback.
The accepted method would seem to be to drop the rate of the PCM stream and resample. Fair enough. In this case though would it not make more sense to use audio timestretching instead? This would avoid any resampling artifacts, pitch change (not that I'd notice it in my example) and CPU usage required for resampling (I'm getting close to the limit for 1080p playback on my system).
I'm making a few assumptions here:
1. Timestretching works in a similar (or maybe identical) way to the SoX tempo effect, which uses WSOLA, and works very well in my own testing.
2. Timestretching and resampling are mutually exclusive options. If I enable timestretching it disables resampling.
3. Timestretching is more CPU friendly than resampling. It should be if my first assumption is correct.
I think my assumptions might be wrong though. With audio timestretching enabled CPU usage still seems to vary depending on the resampling quality setting, suggesting that resampling is still going on. If this is the case could anyone explain to me what the audio timestretching setting is actually doing please? The readme just says that it allows the media speed to be changed without altering the pitch, but doesn't say how this is achieved.
Thanks very much (and thank you for continuing with ReClock - it's invaluable).