Softmpu uses the code of dosbox, so that should in theory provide the same level of emulation.
There can be of course various reasons why certain sound effects do not play, not limited to your usb-midi cable.

Settings I am currently using for DOSBox:

mpu401=intelligent
mididevice=default
midiconfig=1

Here are two recordings I made, no voices. Just pure Midi comparison using Day of The Tentacle between the two.

DOSBox with MT-32: http://reeseriverson.com/rriverson/stuff/DoTT … X-RealMT-32.mp3

Retro PC with MT-32, SoftMPU running: http://reeseriverson.com/rriverson/stuff/DoTT … C-RealMT-32.mp3

Also I just checked the ROM Version, which is Version 1.07.

I made some comparisons a while ago: http://www.youtube.com/watch?v=ZB8zqSRdofk&li … 1vekBzFG68SVzCh

Bottom line is you will struggle to hear a difference.

With sound effects you likely should be emulating the CM-32L and not the MS-32 (Old), but this depends on the game. Lucasarts was mostly using the LAPC-I / CM-32L, whereas Sierra used the MS-32 (Old).

Mau1wurf1977 wrote:

I made some comparisons a while ago: http://www.youtube.com/watch?v=ZB8zqSRdofk&li … 1vekBzFG68SVzCh

Bottom line is you will struggle to hear a difference.

With sound effects you likely should be emulating the CM-32L and not the MS-32 (Old), but this depends on the game. Lucasarts was mostly using the LAPC-I / CM-32L, whereas Sierra used the MS-32 (Old).

But I'm not emulating the MT-32, I am using a real MT-32 in both configurations.

Even right now some of my Sierra games sound different between DOSBox and my Retro machine with the same MT-32 device.

Oh I See 😀

At least that justifies building old computers for lots of money 🤣

DOTT doesn't require intelligent mode so we can remove SoftMPU from the equation. You may also want to try mpu401=uart in your DOSBox config.

It sounds like the custom patches aren't being uploaded to the MT-32 correctly under DOSBox. The usual cause for this is speed-dependent code in the game coupled with a fast machine. I would reducing the emulated CPU speed in DOSBox (reduce cycle count?)

bjt wrote:

It sounds like the custom patches aren't being uploaded to the MT-32 correctly under DOSBox.

This actually is the perfect description of what seems to be going on. The case being for every game I have tried between the two so far.

I didn't think of trying to slow down DOSBox (I heard that it is usually needed if you run into overflow error issues), but that is a great suggestion to try! I shall do that when I get back home tonight. Plus UART mode too for games that will accept it. 😀

Try a cycle setting of 10000. This works well for me for older adventure games.

You may also want to try using

1midiconfig=1 delaysysex

The status window of DOSBox should tell you that it is activated.

The problem may lie on the USB-to-MIDI interface used in the Dosbox machine.

For me, there was no difference either in Dosbox with the USB MIDI cable or the retro system with the Roland MPU-IPC-T.

rcblanke wrote:

You may also want to try using […]
Show full quote

You may also want to try using

Copy code to clipboard

1midiconfig=1 delaysysex

The status window of DOSBox should tell you that it is activated.

Hmm, adding 'delaysysex' doesn't seem to change anything nor show up in the status window for me.

Mau1wurf1977 wrote:

Try a cycle setting of 10000. This works well for me for older adventure games.

My cycles are at 3000 by default, but I tried 10,000 anyway, but it made no difference. 🙁

Malik wrote:

The problem may lie on the USB-to-MIDI interface used in the Dosbox machine.

For me, there was no difference either in Dosbox with the USB MIDI cable or the retro system with the Roland MPU-IPC-T.

This is the USB MIDI cable I am using: http://www.ebay.com/itm/321274065658?ssPageNa … 984.m1439.l2649

Heh. delaysysex option assumes you're using an SVN build and not an official 0.74 DOSBox.

Malik is right: USB-MIDI can appear really surprising. Fortunately, it's quite easy to find out how well your one performs. Make a MIDI capture using DOSBox and play it back on the retro PC.
In fact, a USB-MIDI thing should be carefully tested (especially if you are offered the money back 😉 ). It'd be best to measure how reliable it can transmit sysex messages of length up to 267 (the longest sysex I've ever experienced in the wild). I did such testings using two USB-MIDI cables - one for transmiting test data and one for receiving.