Thanks Daniel, that works! Plus using soundarray should be faster than first converting to raw.
So far I don't need to cache, but rollback netcode demands lots of CPU so I suspect I probably will end up optimizing this to improve robustness vs lag. A pity mixer won't let you start Channels partway through a Sound, as then there'd be no CPU cost. On Wed, Jul 14, 2021 at 9:15 AM Daniel Foerster <pydsig...@gmail.com> wrote: > Instead of using the raw sound binary data and winging the slice, I'd > recommend throwing things into a sndarray. > > > def clip_sound(sound, start_time): > freq = pygame.mixer.get_init()[0] > offset = round(freq * start_time) > arr = pygame.sndarray.array(sound) > clipped = arr[offset:] > return pygame.sndarray.make_sound(clipped) > > > It would be worthwhile probably to cache the frequency and the converted > array and only do the slice + flip back to Sound but that's premature > optimization right now. > > On Wed, Jul 14, 2021, 03:15 Jasper Phillips <jasperisja...@gmail.com> > wrote: > >> Alas, that is precisely what I did: >> >> def clipEffect( effect, startTime ): >> raw = effect.get_raw() >> rawStart = int( len(raw) * startTime / effect.get_length() ) >> return mixer.Sound( raw[rawStart:] ) >> >> Depending on what sound effect I play the code either: >> - Works fine! >> - Horribly distorts the audio, though the original Sound plays fine... >> >> Hi Jasper. Try the fourth answer here >>> <https://stackoverflow.com/questions/57960451/how-can-i-play-a-portion-of-an-audio-file-in-a-loop-in-pygame>, >>> by Ted Klein Bergman. >>> On 7/13/2021 4:24 PM, Jasper Phillips wrote: >>> >>> I'm working on a networked action game with rollback netcode, and so I >>> need to play sound effects starting an arbitrary amount of time from the >>> beginning of the file so they sync up with game events. >>> >>> pygame.mixer.music does this via set_pos()... but apparently there's no >>> corresponding method when playing effects directly via a Channel. >>> Unfortunate, as I need to use Channels! >>> >>> The best I've come up with is to use Sound.get_raw(), trim the starting >>> bytes off then, create a new Sound. But this fails several different ways, >>> I'm guessing because I'm arbitrarily slicing the bytes and there's some >>> format details I don't know. >>> >>> >>> Does anyone have any tips for hacking raw Sound data, or perhaps some >>> better approach? >>> >>> -Jasper >>> >>>
