https://github.com/tEFFx/SnoozeTracker

SN7 with custom waveforms and PCM. ;)

All we need is binary export, or binary music rip exports (SGC/KSS). ;)

THANK YOU for your efforts tEFFx! Let's put it to good use!

What's that feeling called when you get something you haven't even realized you wanted?

>thinking only of gamedev

I don't think this was made in mind with software running in background


Anyways, since I see some people bashing on it already:
Yes the .vgm files it outputs are very very big, but you can reduce it with ValleyBell's VGM compressor (no it does make it a .vgz), usually reduces it to half the original size and the second thing:

as of writing this -- It's only a second release of something that is literally titled something that is far from done; alpha


I for one welcome the tracker, even if it's a bit unpractical in mind and such, but directly bashing on it seems a stupid thing to do already, famitracker wasn't perfect in the beginning as well

Asparah: Dunno if I'm correct in reading your posts as addressed to me or not, but I am not related to tEFFx, despite a rather similar name, and was merely expressing intrigue for the tracker.

SN with custom waveforms, and written on Unity's framework - who wouldda thought.

To be fair, though, memory space is a legitimate concern for the PCM samples regardless of whether or not you're doing anything other than playing music (well, unless you don't actually intend on packing it in a ROM). I remember a thread on SMSpower that was detailing single samples that would consume a large amount of ROM on cart, not to mention basically all the channels just to play the PCM back.

Don't see too much of an issue with PCM or wavetables... You've got 1MB of ROM to work with. You guys are worried way too much over the sequencer pattern compression in relation to playback routine.

SnoozeTracker doesn't even have a Z80 playback routine at the moment. It's up to tEFFx or the community to make one based on VGM output or gleaning from the TFM file. The bytecode from the sequencer data could also be dupe/loop checked and loop characters could be added to decrease ROM space.

Regarding PCM, you're already taking 16-bit PCM, mixing stereo, and then dividing the 16-bit sample by 4 and crunching the value into a nybble. The sample is already 4x smaller than it previously was. ;) Just have to determine a good rate to play the samples back at; given the overhead of underlying processes within the sound engine. (This is likely already determined in that C# code somewhere.)

Apsarah, you're gonna make the Game Gear playback routine, right? ;) (Given that SnoozeTracker is stereo, we can assume that SnoozeTracker is a Game Gear tracker. :D)

>we can assume that SnoozeTracker is a Game Gear tracker

I think it's advertised the same way as Sneventracker in that regard

This forum should allow SnoozeTracker TFM files to be uploaded!

Hey, wouldn't the 'not doing things in the background' side of things be pretty relevant for a Z80 sound driver on the Genesis/Mega Drive?

Or at least only doing other sound related things in the background. (though there's the 68k sound driver route with Z80 only acting as a sample playback slave)

You've got the shift-register bank switching overhead and 32kB bank size constraint to deal with, especially with the larger file sizes involved here. (also issues with bus contention with DMA is used to update VRAM)

OTOH tunes done with SnoozeTracker with RAM/ROM is no object, but do-able on a 3.58 MHz Z80 would still be a pretty solid representative demo for what more system-specific, lean coded drivers could be used for. (getting good sound out of the logarithmic 4-bit amplitude seems to be the bigger challenge that's been overcome here)

Plus you could argue hardware mappers on-cart that help the Z80 on the MD or even on the SMS. (Tiido even prototyped a mapper+DMA board for the SMS around 6-7 years back I think, but I didn't see it go anywhere, but it looked to have DMA both for writing to VDP RAM ports and to SN76489 PSG registers, I think with 8 pulse-wave generator sound channels mixed into the SN channels too, I assume sort of like how Pitfall II did it)

As far as all-software techniques go, I think Afterburnet II on the MD was the only game to bother with PSG sample playback. Seems to do 2 channels/voices with 1 dedicated to speech/SFX and the other to percussion. (might also all be crammed into a single 32k block of ROM in this case, sounds like <4 kHz and heavily filtered in preprocessing, but not too bad for an early MD game, especially compared to early Z80 PCM drivers or some of the crappier drivers used years later on)