I do believe that Wing Commander III had a Roland option. Privateer is indeed interesting. It sends a reset at the start of the game, which stuffs up your MT-32 if you configured it with the GM map 🤣

Apparently it uses some code from the "MT-32 days" or something like that.

But I think most companies back in those days transitioned from MT-32 to General MIDI. Lucasarts was the same, Fate of Atlantis was MT-32, DOTT supported both.

I wonder why they even bothered so long with MT32 music. No one had those things!

Dominus wrote:

I wonder why they even bothered so long with MT32 music. No one had those things!

Roland endorsement? More MONEY!

I had a LAPCI from about 1991, and it would have been Privateer and Ultima VIII that would have forced me to do something about GM, I remember there was a problem with the Roland SCC--1 but I can't remember if it was cost or a configuration issue (another ISA card). My eventual solution was to replace my Sound blaster pro with an AWE 32.

James-F wrote:

Roland endorsement? More MONEY!

Could be. Result is that a totally unrepresentative amount of games has MT-32 support, which most people now use as a 'fact' to show all the other hipsters that MT-32 was THE music standard in the days of DOS gaming.
Reality is that there are a ton of games with MT-32 support which hardly anyone ever got to hear, because hardly anyone ever even came near a PC with MT-32 setup.
I didn't even see MT-32/LAPC-I in computer shops at the time. They probably didn't bother to stock them, because they were out of the price range of their average customer anyway.
I only saw them in music shops.

I think the demoscene history tells the real story of what hardware was popular, because there were not a lot of endorsements going on. Demosceners were 'regular' people in the age of say 15-25, and they just used the hardware that they could get their hands on, either using the 'family' computer, or buying their own hardware from whatever job they had.
DOS demos mainly supported Sound Blaster, SB Pro and GUS, because that is what 'normal' people could afford.

To my knowledge, there is only one group that ever bothered to support any Roland hardware, and that is the group The Phoney Coders: http://www.pouet.net/groups.php?which=1451
Again, probably because they had a member (possibly more), where that person or their family was into music production with computers, so they had a Roland MIDI setup available at home.
I guess that single group amongst the hundreds of groups being active in the DOS demoscene at the time is an accurate representation of how widespread the MT-32 was.
I believe only 3 of their demos had any Roland sound. They didn't specify exactly, but they say a 'Roland soundcard', so presumably they mean the LAPC-I. The music seems to be MT-32 music anyway.
Here's a video that includes all of them: https://youtu.be/Y9jJYVM-dq0
So there you have it, probably only 3 demos ever made for an MT-32 (with two of them sharing the same music, so only two MT-32 tracks) in all of demoscene's history.

I have a couple of alternative theories, as music became more important especially in graphics adventure games the musicians brought in to compose the music really wanted to work with the best tools and that was Roland. Secondly even if very few people had Roland, computer mags reviewing the games did and made a big deal of it in the reviews. When I saw a used lapci in a computer shop in 1991 I was well aware of the significance of the of card at the time even though I don't believe I'd ever heard one play and bought it in blind faith. I'm very glad I did!

lvader wrote:

I have a couple of alternative theories, as music became more important especially in graphics adventure games the musicians brought in to compose the music really wanted to work with the best tools and that was Roland.

That doesn't make sense. MIDI isn't compatible with any other common audio devices either on PC or on other platforms such as C64, NES, Atari, Amiga etc.
The ONLY thing you could do with MIDI was to make MT-32 music basically (okay, and IBM Music Feature Card, if you want to talk obscure... but you still had to write a separate MIDI track for that, since neither MT-32 nor the IMFC were general MIDI compatible, and required custom instrument maps etc).
For any other music chips, you'd need custom sound routines, and at best, you could try to translate the MIDI data to a native format. But since MIDI is aimed at devices with massive polyphony, MIDI data doesn't usually translate well to devices that have just 3-4 channels, as was common in those days.

I did a two-part blog on that recently, actually:
https://scalibq.wordpress.com/2017/03/29/trackers-vs-midi/
https://scalibq.wordpress.com/2017/05/29/trac … midi-part-deux/

lvader wrote:

Secondly even if very few people had Roland, computer mags reviewing the games did and made a big deal of it in the reviews.

That one may have factored in. I guess the only way to tell would be to read the reviews, and see if the reviewers commented on the MT-32 audio at all, and to what extent they let it weigh in on their overall rating of the game.
I mean, reviewers that are conscious of the game-buying public and their hardware, would have to factor in that not many people would have access to MT-32.

I always felt that game developers wanted to push the existing technology to its limits, that's why they often pursued higher-end music options for their games. Especially when graphically games were stuck at VGA resolutions for so long. Part of my nostalgia for DOS games is the fact that the music could bring games to life when the graphics couldn't.

clueless1 wrote:

I always felt that game developers wanted to push the existing technology to its limits, that's why they often pursued higher-end music options for their games.

On most platforms, they did.
On the PC platform, they basically didn't push the AdLib at all. Only a handful of games actually do more with the AdLib than just using it as a generic MIDI playback device. Almost as if the AdLib was an afterthought, and they just tried to 'backport' the music they did for the MT-32.
At least for PC speaker, there were various games that tried various techniques. Most of them would at least try to multiplex 3 or 4 channels by a form of arpeggiating the different notes played on the channels. This also allowed music and sound effects to play at the same time.
Then there were games like Digger and Budokan, which used rather unique techniques to make the PC speaker sound like more than just a single-channel square wave device.
And of course there was Access Software's RealSound, to play back sampled audio using PWM.

On Tandy/PCjr, again, not many people seemed to actually push the device.
The main example I know of pushing the device is the 4 games that Rob Hubbard did for EA: Skate or Die, 688 Attack Sub, Kings of the Beach and One on One.
He used sample playback combined with the regular noise and square wave channels, to make the SN76489 sound pretty damn amazing.
Then again, he had been doing basically the same on C64 (his main platform) for years.
But you rarely saw this kind of pushing the hardware in PC games, neither with audio nor with graphics.

Scali wrote:

That doesn't make sense. MIDI isn't compatible with any other common audio devices either on PC or on other platforms such as C64, NES, Atari, Amiga etc.
.

That wasn't' my point, it wasn't a conversion tool but a composing tool. The professional musicians brought in by the like of Sierra would have been at home on the midi based MT32 and not the adlib which came later anyway. In many cases the adlib port was likely to have been done by more junior people.

lvader wrote:

That wasn't' my point, it wasn't a conversion tool but a composing tool.

Game composers generally don't compose songs on 'regular' instruments first. That makes little sense, because the hardware they have to translate it to, is very limited, and has a very specific sound to it.
Rob Hubbard explained that he basically composed right on the C64. He would get an idea in his head ('thinking' in C64 audio already), quickly write it down, and then program it in, one chunk at a time. So he'd have an idea for a few bars, program that, and then he'd have those bars playing while he tweaked the instruments in realtime to try and get it to sound right.

The reason why AdLib sound sucks on most PC games is that PC developers generally didn't seem to understand how to write games and music (unlike the developers for home computers and consoles, which turned out lots of stunning soundtracks with far less limited hardware than the OPL2).
You can't just use a regular composer. You'll need someone who is both a composer and a programmer (or a team of composer and programmer who can work together productively). Rob Hubbard was one of the first of its kind: he was originally a session musician, but got into programming, initially just to use the computer as a composing tool, but then started getting assignments to write music for software. The rest is history.
Dune is one of the few examples on PC which has great AdLib music. This is because there were two guys working on it, one writing the HERAD tool, the other working with him to compose music with that tool. It shows just how much most other PC devs sucked at using the AdLib.
Which is especially crappy since most PC gamers had an AdLib, if they had a sound card at all. So they either heard the PC speaker, which was crap no matter how skilled a composer/programmer you were, or you heard the crappy AdLib music, despite having a very nice Yamaha FM synth on your card, which could easily have blown most home computers' sound chips out of the water, if the composers/programmers just knew what they were doing.

Most stuff was composed on keyboards, even Rob Hubbard used a Casio in the C64 days.

Back on topic, when I get a chance, I'll do a comparison between the MT-32 and GM versions of that song in Pacific Strike. I don't have a nice capture setup on my DOS systems, so I'll have to make due with phone video. I could not find any reference to this song difference on the web, it seems to be undocumented. Would love for Matt Barton to ask that question to someone from Origin.

@lvader and Scali:
Matt Barton has an incredible set of interviews with George Sanger (aka The Fat Man) where this topic is discussed quite a bit.
https://www.youtube.com/playlist?list=PL154C83C19E37106C

You could try asking on Twitter 😀
For example Richard Garriott often answers on Twitter and might point to the right guy to ask this.

lvader wrote:

Most stuff was composed on keyboards, even Rob Hubbard used a Casio in the C64 days.

He may have used one on occasion, but it was not his primary tool for composing C64 music (and certainly not via MIDI, he coded things by hand).
You can hear it from the man himself in this interview: https://youtu.be/ao92PVEHG3c?t=25m49s

And I believe George Sanger composed on keyboards or other real instruments mostly. I guess it depends on how far back you go. Back when games fit on a single floppy and the (one) programmer also did the music themselves, that's one thing. But composers are not always coders, and the more complex games got, the more there were teams involved, which freed up developers to bring in composers who weren't coders.

Hey, what do you know? Found the two different songs with some YouTube trolling:

MT-32 Version: https://youtu.be/Gchs3opURZA?t=105

General MIDI Version: https://youtu.be/FXq6OouGzbw?list=PLhT9gKHUZZ … M8M0-lj2K&t=125

clueless1 wrote:

And I believe George Sanger composed on keyboards or other real instruments mostly.

But he never did anything on 3-4 channel devices, did he? I only know him from AdLib and newer PC hardware.

clueless1 wrote:

I guess it depends on how far back you go. Back when games fit on a single floppy and the (one) programmer also did the music themselves, that's one thing. But composers are not always coders, and the more complex games got, the more there were teams involved, which freed up developers to bring in composers who weren't coders.

I already covered that in my blog: trackers.
Rob Hubbard developed his own music routine, which was revolutionary on the C64, and in game dev in general, I suppose.
Other developers would reverse-engineer his routines, and found that he basically used a system that we now know from trackers: music data stored in rows of 'patterns', which could be sequenced together to form a longer song.
One of the first programs we could consider a 'tracker' as we know it today, is by another famous C64 musician: Chris Huelsbeck. He developed his own editor and released it for free: Soundmonitor: http://csdb.dk/release/?id=59929
Eventually, C64 people would upgrade to Amigas, and Karsten Obarski introduced us to The Ultimate SoundTracker (which is probably also one of the first, if not the first time the term 'tracker' is used for a music editing tool):
http://www.pouet.net/prod.php?which=34731
This tool was not free, so demoscene people started cloning it and making improved versions. The most important one is probably Mahoney & Kaktus' NoiseTracker: http://www.pouet.net/prod.php?which=47996
This then evolved into ProTracker, arguably the most popular tracker of all time: http://www.pouet.net/prod.php?which=48005

Point being: these programmer/composer people started building tools for writing music for specific chips, and made them available to everyone for free.
So you no longer had to be a programmer, you now had tools that allowed you to do what Rob Hubbard did by hand: write music and modify the parameters in realtime on the actual hardware. These tools were quite low-level, and allowed you to optimize the allocation of notes to the different hardware voices, and make use of all sorts of special capabilities of the hardware. Yet, they were easy enough for any determined composer to use.
You had to have had the coders first of course. And that's what's sorely lacking on the PC side: not a lot of AdLib trackers were released (or even made, by the sound of most games), and the ones that were, were released very late in the game, like EdLib being from 1994. By which time AdLib was starting to be replaced by wavetable-based soundcards in the mainstream, or just by software mixing on 386/486 systems.
If such tools had been available around 1987-1988, we may have had a way different AdLib experience. But it seems that all the skilled coder/musician-people weren't bothering with PCs, they generally worked on home computers and game consoles. The PC got leftovers.

I have read interviews in which Rob Hubbard talks about composing music on a Casio and writing the score on musical score paper before coding it on a C64. By the time MT32 became popular composing music using midi keyboards, synths and sequencer sortware was common place.