See Pirkle DSSPC++ p.159. They can say which note they think they are, for example.
I suppose the reason the industry migrated toward metadata in filenames is that you can grab everything via a directory listing, rather than opening every file. Plus you probably have to put metadata in the filename anyway just to disambiguate two samples that are identical except for pitch. I wonder whether this is still a good tradeoff.
I've done enough of this by reading Acidized RIFF chunks (c6f5ceed05213da96e2156f4c6c287c03911247f), and there's a separate issue #99 for handling the smpl chunk.
See Pirkle DSSPC++ p.159. They can say which note they think they are, for example.
I suppose the reason the industry migrated toward metadata in filenames is that you can grab everything via a directory listing, rather than opening every file. Plus you probably have to put metadata in the filename anyway just to disambiguate two samples that are identical except for pitch. I wonder whether this is still a good tradeoff.