I've been trying to figure out how to compose music that accurately models the music that can come from a SEGA Genesis and Super Nintendo. FM synthesis is fairly easy for me to understand, however the SNES has had me pretty darn confused because it seems to be a little more complicated than FM.
I gathered a bunch of information from around the internet, including here (Neil Baldwin has been pretty helpful from reading other older threads). Here is what I have written down so far.
"The audio chip of the Super NES/Super Famicom (Sony SPC700 via 65816 RAM(?)) has 8 adaptive differential pulse-code modulation (ADPCM) channels for bitrate reduction- (BRR) compressed samples. All the samples had to be compressed with ADPCM/BRR formats at a ratio of around 3.5:1 (32:9(?)) to less than 64 Kilobytes to fit on the discrete audio chip’s RAM. This means that for every 32 bytes of 16-bit PCM there are 9 bytes of BRR. The size of all the samples should in total not exceed more than 150 kilobytes. The original, raw samples of the instruments are tweaked for downsampling, then downsampled to 8 to 22 kHz, and cut into small loops that are then truncated to 8 bits. Equalizing to retain brightness can occur before or after downsampling. Samples are often no larger than 2 kilobytes, having 16 samples to be looped.
The DSP also allowed for echo effects. [How can this be recreated accurately?]
The SNES is capable of Dolby Surround Sound, albeit it was pseudo-surround sound using stereo Q-Sound.
Many games used Nintendo’s ready-made samples for its SNES games.
Because there are currently no programs that have simplified the process of making accurately modeled SNES music like FamiTracker does with the NES, the alterations to the original, raw samples must be done manually. I can use BRRTools to compress the raw sample."
Did I get anything wrong here? I want to get these things down pat before I begin shouting from the mountaintops that I can compose music that can (theoretically) be played on an actual SNES.