From writing part of the GUS code, I can assure you that there is at least one thing different between the emulated GUS and the real thing. The real thing would reduce the outputted sampling rate as more channels were played simultaneously. The emulated version does not suffer from the same processor limitations the hardware version had, and will therefore use the sampling rate set in the configuration regardless of the number of channels playing. This may automatically make for a slightly different sound.

Would that change the pitch of the sound produced though?

HunterZ wrote:

Would that change the pitch of the sound produced though?

I wouldn't expect that to happen. I programmed the GUS quite intensively back then, and one of the tricks I created (to be used in a moduleplayer) was to adjust the number of voices to the absolute minimum by 'relocating' voices from higher numbers to lower, unused, numbers by implementing simultaneous, opposing volumeslides, adjusting the samplerates on all channels and changing the number of voices used. Unfortunately, changing the samplerates was slightly audible, so I never implemented it. What's important for this discussion however, is that adjusting the samplerates + the number of active voices did not create an audible difference in most samples.

JAL

canadacow wrote:

From writing part of the GUS code, I can assure you that there is at least one thing different between the emulated GUS and the real thing.(...) This may automatically make for a slightly different sound.

Did you have a 'real' GUS playing that you compared to the code you wrote? Or did you just assume it to sound right?

JAL

Whoa, a ninja-post by canadacow. Don't expect a reply for two months or so 😉