VOGONS


First post, by KingGuppy

User metadata
Rank Member
Rank
Member

There are a lot of things I'd like to do with the Munt project that I lack the time, expertise and/or equipment to accomplish in any reasonable timeframe. Here's a quick list of ways in which you could help:

Improve the web site
http://munt.sourceforge.net/ is currently... uninspiring.

Build a digital sound ripper
Munt is getting to the stage where it's very difficult to improve its accuracy using analogue samples. It should be possible for someone with some electronics experience, equipment, time and a bit of cash to build a board for digitally ripping the MT-32's audio.

Reverse-engineer the ROM
There are some subtle issues with partial management, MIDI command handling and so on that could - I think - be resolved by studying the control ROM. If you're proficient in assembler (or willing to learn), and have or can write a disassembler for Intel 8096 code, this might be a job for you.

Port
I'd especially love to see a driver for Mac OS X, but support for any platform/system would be very welcome.

Test
Though I may not always respond, I do always take notice of messages posted to this forum about problems with the emulator. Attaching sound samples where useful is always appreciated. If you want to make extra sure that your report doesn't get forgotten, please use the bug tracker for Munt on SourceForge.

Analyse
I've written a few tools to gather data on the sounds produced by the MT-32, but it turns out to be quite tricky to interpret the output in some cases. If you're mathematically minded or have a knack for finding patterns, I have a lot of graphs and tables to throw at you.

Code
You don't necessarily need any knowledge of sound synthesis to usefully work on Munt's code.

If you're interested in doing any of these things - or something else that I haven't thought of - please let me know (preferably before you get started, to prevent duplication of effort).

If you're not interested or able, and know someone who might be, please spread the word.

Reply 1 of 9, by Reckless

User metadata
Rank Oldbie
Rank
Oldbie

A more 'inspiring' web site would be a cool plan 😀 I could lend a hand/do it if you're short of someone. I don't have uber web skills mind you!

Reply 2 of 9, by taiken7

User metadata
Rank Member
Rank
Member

The 8096 family is now called the 'mcs96' series:

http://www.intel.com/design/mcs96/docs_mcs96.htm

There is also a debugger and compiler suite

http://www.geocities.com/rjkuhn_2000/software.htm?20053#kpc

And a more professional demo (Sim/Emu/Debugger/Assember...etc)

http://microcontrollershop.com/phyton_project-196_spec.php

Reply 3 of 9, by Lord Nightmare

User metadata
Rank Newbie
Rank
Newbie

Kingguppy:
I am in the process of adapting a GB-z80 disassembler to disassemble i809x code (it isn't easy). The i809x/i839x series is like a cross between a motorola cpu registerwise (it has a 232-byte register file as opposed to discrete registers) and an intel 8080 (CISC instructions).
The most important document, assembler/coding-wise from the intel site is:
http://www.intel.com/design/mcs96/manuals/272258.htm
I'll keep you updated as I finish the disassembler.

I also have a copy of the original 809x/839x datasheet from 1984, contact me if you need a copy (its missing a LOT of info which only the newer datasheet has).

Lord Nightmare
lord_nightmare @t users d0t sf d0t net

Reply 4 of 9, by adventur

User metadata
Rank Newbie
Rank
Newbie

You've got my support! cheers, I'm sorry I can't do much code though, but Munt is awesome & the best thing that should be done is something about speed issues because it really slows down Dosbox to almost a half!

Keep up the good work 😅

Reply 5 of 9, by Snejs

User metadata
Rank Newbie
Rank
Newbie

I'm willing to atleast give the coding part a try.

The analyze part could be interesting as well. Could you give some more examples of what you want here so I know if it's totally out of my leage?

BRs
Snejs

Reply 6 of 9, by adventur

User metadata
Rank Newbie
Rank
Newbie

Forget what I just said! I just realised Dosbox 0.65 is twice as fast as before!!! this said I haven't tried it without Munt so I cannot say if it is EVEN faster, but it shows amazing performance enough already! 😲

This said, more options of the the original Mt-32 Module could be a great addition, I don't know how much the hardware can do since I'm no expert but it would be nice to emulate everything it can do since mine is stored pretty far away at the moment

Reply 7 of 9, by Spikey

User metadata
Rank Oldbie
Rank
Oldbie

I can certainly capably test any new releases and compare with previous ones.

- Spike

Reply 8 of 9, by Forza

User metadata
Rank Newbie
Rank
Newbie

Hello!

I'd like to help with analysis or testing, a port to Mac OS X would be very interesting as well but first of all let me tell you that my knowledge of programming is very basic, however I love computers and i'm always willing to learn new things, besides that I'm a skilled musician / sound engineer, and I'm familiar with most kinds of popular synthesis and several more obscure ones.

if you think I could be of help to the project, by all means please let me know!

regards

Peace, Love, Light and Harmony

Reply 9 of 9, by vladr

User metadata
Rank Oldbie
Rank
Oldbie

Attached is the HDL design for a buffered digital acquisition setup I designed for the MT32 (but did not actually have time to physically implement, PCB and all.)

-----------------------+          +-------------+      |              |
MT32 | | Ext. RAM | | COMPUTER |
+-----+ | |(capt.buffer)| +--------------+
digital | DAC | | analog +-------------+ ^
audio --+-/->| |-> audio ||||| | USB
in | |PCM54| | out +-------------+ +--------------+
| +-----+ | | XILINX -or- | | DLP or eq. |
| | | Cypress PLD | | USB to FIFO |
+------/--------------->| probectl /w |--/-->| parallel i/f |
probe | | fifoctl | | module |
-----------------------+ +-------------+ +--------------+

The HDL (fifoctl+probectl) was designed to fit on a cheap XILINX or Cypress PLD, and use a configurable amount of external RAM for buffering. It would interface with the computer via USB using a $25 "USB-to-FIFO parallel I/F module" (like the DLP Design USB245M which was available back in those days.)

The acquisition would be performed at the entrance of the DAC, and the captured data would be retrieved on a computer e.g. via DLP Design's driver and sample app.