@MazHoot So I fired this up in BlastEm and was quite pleased with the voice edit view, even though it uses pure register values instead of MIDI like I'm used to. If I could get that with genMDM it would be next level in a mixed workflow like mine.

Just dreaming is all, don't want to distract from a perfectly good project

created discord server: https://discord.gg/CE22HNNk5y

@category.zd do you mean, use the genMDM interface in reverse to make midi out notes with MDtracker instead of midi in? Or do you mean, keep normal using of genMDM and do midi out from emulator?

The first should be so interesting, I don't know if it's possible or not but I already thinked about that... if you mean the second, I think having save sram working on everdrive or other cartridge is the way to go (with research, coding, I hope it will be possible in the future)

Because for the second, if we go this way then I would prefer just use little piggy tracker to control genMDM which already works great (works great on raspberry pi and probably on computer too)

In my opinion now the first thing to do before all is resolve everdrive (or other cartridge) sram save bug.

Now if we can just combine this with GenMDM...

I also successfully used it (it was a previous version) on picodrive on RG350.

Just tested on android retroarch+picodrive core. Works (even better when overclocked) and saves fine :3

It is in C. Assembly etc goes from SGDK.
Compilation is as in any other SGDK projects. https://github.com/Stephane-D/SGDK/wiki … stallation

Woaw!!! Thank you!!

It's all in c language or there is parts in assembly?

Is there instructions for compilation?

1. - MD tracker is not based on Deflemask. Some values are flipped to make more sense going from 0 to max like the most of others.
AMS/FMS are flipped WHOOOPS wrong docs to blame again...

2. - Everdrive X7 has 256Kb (S)RAM, but thats all I know. IDK how it's really handled by flashcart, maybe only supported to read/write to SD is 64Kb.

3. - If there is less than 256Kb of real (S)RAM, nothing will work.

now opensourced https://github.com/corthax/mdtracker

Hum ok so I think my megadrive 1 is TMSS because I can see a splash screen. But for me save states don't work too. You have to look carefully because if it does like me you can think it works at first but when you go in note screen, the notes or instrument are not here

But i'm not sure if I have TMSS or not I will check it

Correct -
I have a Model 1 - High Definition Graphics model - NON TMSS - which means it doesn't have the lockout chip Sega put it later revisions on the model. This one has a genuine YM2612 inside and probably the best filters.

My other Model 1 - does not say High Definition Graphics and has the TMSS lock out - so a "PRODUCED OR UNDER LICENSE FROM SEGA" splash screen appears first before game starts. Sound is still decent on this one - but I can't confirm it's AS good my top one.
This guide should help:
https://www.sega-16.com/forum/showthrea … m-bad-ones

@diagamblic : model 1 is sega genesis 1 right? But what is TMSS and NON-TMSS?

1)
I've created a spreadsheet that can convert Deflemask instrument values to MD tracker!
I'd like to share it with you all in case you find it helpful while testing:
https://docs.google.com/spreadsheets/d/ … sp=sharing

The only instrument so far that I've been able to MOSTLY faithfully reproduce is a kick drum - by moving around the release value on OP3. I've included the kick drum instruments (among others instruments).

Since Deflemask is pretty accurate on actual hardware playback - I'm thinking there are still some tweaks that need to be done on the instruments in MD Tracker.

2) Everdrive v7 at the moment corrupts save state data on a Non TMSS Model 1. And SRAM doesn't work.
HOWEVER - Saves States seem to work on my Model 1 TMSS model.

3) I haven't been able to get any effects working (vibrato, porta, etc). I'm not sure it's me or the program however.

Corthax - would you consider setting up a Discord for the program? There has to be an easier way to catalogue errors and share wips.

Well. Seems that 68K cannot run current logic fast enough, when trying to write instrument changes at all channels at the same time. More stable timer not helping much here.
Overclocking can fix that, but thats not realistic.