I recently used a notation tool to generate just a couple of notes for one instrument - at the end of a piece. The file generation took the best part of a minute - to generate an MP3 file. It would have been the same for other audio formats I think.
Given that 99% of the file was silence surely it would have been more effective to merge two files - one containing absolute silence and the other with the one or two seconds fragment at the end.
I assume that with PCM audio this is easy, though I'm not quite so sure about compressed formats such as MP3. It seems possible to me that there could be timing errors with some compressed formats - though how significant they would be I'm not certain - probably "usually" unimportant.
Given that the software which generated that file allows many instrument tracks/staves, then it occurs to me that if this level of "performance" is accorded to all the instruments in a piece that generating audio would use very many completely wasted CPU cycles.
Given that 99% of the file was silence surely it would have been more effective to merge two files - one containing absolute silence and the other with the one or two seconds fragment at the end.
I assume that with PCM audio this is easy, though I'm not quite so sure about compressed formats such as MP3. It seems possible to me that there could be timing errors with some compressed formats - though how significant they would be I'm not certain - probably "usually" unimportant.
Given that the software which generated that file allows many instrument tracks/staves, then it occurs to me that if this level of "performance" is accorded to all the instruments in a piece that generating audio would use very many completely wasted CPU cycles.