VOGONS


Munt Reloaded - Development

Topic actions

Reply 361 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

Yep, it's close to LauraBow_CM64 made by Kaminari, but your CM-32L doesn't require any modification of the MIDI file to play it correctly. 😀

Thus, we must find the exact cause of Thunder MS partials abortion. Whereas channel 3 really contains PC message to change Thunder MS into the organ, channels 4 and 5 don't. Nevertheless, both Lightning (chan 4) and Thunder MS (chan 5) shut down approx. simultaneously.

So, if you try to delete each of these SysEx, Pitch Bend, Modulation and Pedal messages sequentially as well as try each of channels 3, 4, 5 solo, maybe it would at last be clearer what your CM-32L does but Munt doesn't. 😀

Reply 364 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie
Kaminari Redux wrote:

I have another bug to report with the complete soundtrack of Colonel Bequest.

At around 08:15, there's a small organ theme which starts with a thunder patch. In Munt, the patch is heavily distorted, saturating the rest of the music for a few seconds. In the console, Munt complaints about insufficient free partials for that patch.
...

This one seems to be fixed now...
Thanks to Jerome's recent deep researching of the Sysex processing, we now have much more adequate channel assign modeling.

Reply 365 of 964, by Kaminari

User metadata
Rank Oldbie
Rank
Oldbie

Sorry for not popping up sooner. Great job all around! The partials bug at 8:15 seems fixed indeed.

However, there is still a similar distorsion at 1:15, albeit smaller yet noticeable. I'm not sure what patch is involved, since the console of the latest Munt version is much less verbose than in previous versions (now it only displays "Play SysEx message xxx bytes"). It sounds like a small thunder followed by a low rumble. Do you need a recording?

Reply 366 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

I'll appreciate any recordings due to I still don't have any LA module for some reason. 😀
Yep, I noticed that distortion caused by Thunder MS, of course. However, IIRC, KG said that Munt plays Thunder MS patch identically to CM-32L. So, maybe we have here another place of partial abortion that Munt ignores. Let's listen to the recordings...

Reply 368 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

From what I can hear, Munt's partial allocation / abortion algorithm rather very different. This pattern generates a plenty of "... (Thunder MS): Insufficient free partials ..." messages.

It seems that Munt never aborts PLAYING polys, though it aborts RELEASING and HELD ones. In contrast, this Thunder / Lightning pattern contains no polys in HELD / RELEASING state, therefore, Munt never tries to abort them.

Reply 370 of 964, by Kaminari

User metadata
Rank Oldbie
Rank
Oldbie

Good job, works a charm 😀

... but I have yet another one to report! Interesting thing is, to my knowledge it didn't happen in previous versions of Munt. Still the same kind of distorsion produced by a sort of rumble/thunder sound. This one is around 16:10.

Here is a recording from Munt.

Reply 372 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

Please check DAC input emulation mode. I encounter similar crackling if turned GENERATION1/2 mode only.

I encountered subtle distortions with other tunes when the mode isn't NICE. This is really strange to me since KG and I crunch almost all out of amp levels emulation. Errors were about +-1 both for synth and PCM partials. Seems like we need to retest with GENERATION 1/2 modes...

Reply 373 of 964, by Kaminari

User metadata
Rank Oldbie
Rank
Oldbie

I will record the segment ASAP.

In the meantime, you are right, I am using Generation 2 since it is the default setting for the win32 app (I'm not using the win driver) and I suppose it should sound closer to my CM-64 and my CM-500 when I'm making comparison tests.

I tried the Nice setting, and indeed, it sounds perfect! Even some small volume distorsions I could hear before in some tunes disappear when using Nice setting.

Reply 375 of 964, by endre1952

User metadata
Rank Newbie
Rank
Newbie

How is compatibility with a first gen. MT-32? Somebody has posted on QuestStudios that a sound effect in Heart of China is only heard properly on a first generation MT-32. It's the sound effect when you're flying to a location and the postcard gets stamped. I can make a recording if you wish.

Others here could point out an issue in Dune 2 and Willy Beamish. I'm not sure about those.

Reply 376 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

@Kaminari Redux:
Too bad, the capture you made shows significant differences. We need a detailed analysis of this Thunder patch at the partial level (which is impossible w/o KG's wonder-box).

@endre1952:
Good question... Without a detailed spec, there is almost nothing to do. To succeed, we need a bunch of digital captures of 1st gen. device but the setup we currently have unfortunately involves CM-32L. So, in the nearest time, I doubt we can advance much in the matter 🙁

Reply 378 of 964, by KingGuppy

User metadata
Rank Member
Rank
Member

Still on board but, realistically, probably unable to do any Munt work for another 4 weeks or so. In the mean time, sergm is doing a fantastic job on the new GUI application, so good things are coming.

You can read a bit about the DAC capture method here: Munt development ?

It's not what I'd recommend, though! The whole setup is very, very fragile - physically, electronically and in terms of driver/firmware/USB host interactions (it's almost completely unbuffered). That's why, although I have an MT-32 and could switch back to capturing from it, it's something I want to avoid doing until we're sure we've squeezed absolutely all the information we can get from the CM-32L while the capture's working.

I'm always on the lookout for a better capturing solution at a reasonable price.

Reply 379 of 964, by robertmo

User metadata
Rank l33t++
Rank
l33t++

I think it will be better to have another "wonder-box" for mt-32 as there will always be someone showing up with a bug from time to time, as no one is able to test all midi games one by one.