VOGONS


First post, by zaphod77

User metadata
Rank Newbie
Rank
Newbie

The TTS-1 from current Cakewalk is a 64 bit DXI, and seesm to be legitimately the only free roland option available (i know of no way to get the VSTi hypercanvas for free legally.

Naturally i'd like this as a system midi driver, selectable with coolsoft's midi mapper.

While it's easy enough to acquire other VSTis, and install those as system midi drivers, i know of no legit free vSTi for roland midi playback. i think at one point band in a box had it but i don't think band in a box is free either. and now it has the very ironically named CoyoteWT (laugh). For those who don't get the joke, the Acme Coyote Wavetable is a hypothetical card mentioned in a ancient magazine review of wavetable cards as one that screws up general midi composed for sound canvas very badly.

For XG we have the famous SYXG60 portable version VSTi. For XP we have the MS update cab. Neither is actually legal, but yamaha no longer seems to care. I consider this solved, unless someone has a better way.

For FM we have a lot of options. Nukedyt seem to have system driver that does win95 2 op, and doom and apogee. But I'm also guessing that that has the same polyphony limit as the original. this isn't at all a deal breaker, but it would be nice to have a version without the limit. libadlmidi supports a lot of other opl2 and opl3 banks, but not sure how to make that a system synth. And there are a number of other options. Libadlmidi seems like it should be able to handle pretty much any opl2/3 bank, and would be nice to have as a system midi driver. it seems to cope very well with my ancient iNES midi rips with a bajillion channel aftertouch and pitch wheel messages to actually replicate the volume envelopes for NES music.

For mt-32, we have munt, which i'm pretty sure is installable as a system midi driver. I believe it was determined that munt was legal, due to the original mt-32 rom technically having no copyright because it was not registered. But mainly this is for dosbox use, which doesn't really need it as a system driver. Even if the original mt32 rom is copyright free, that doesn't cover the cm-32l or the cm64, and probably not the lapc-1. but munt is still a very good solution, and i don't think it's been beaten.

My standard for XG is plays every midi composed for mu-50 properly. Not even the later yamaha xg pc cards do this (they mess up the delay on Children.mid) but syxg60 DOES. getting the final fantasy 7 pc midis working also is a plus.

My standard for GS midi is plays back everything composed for sc55mk2 correctly. I'm aware that hypercanvas/TTS-1 doesn't meet this, but it's as close as i'm probably gonna ever get without violating copyright. But i know no way to use it outside of Cakewalk without buying a DXi host.

S0 here's my questions.

1) is there a legit way to playback all sound canvas composed midi files without paying anything using a system driver? This includes stuff using the time variant filter (included in most versions of VSC).
2) is there a way to put libadlmidi and the like into a system driver, and configure easily to replicate various midi opl banks?
3) if system drivers are impossible, what is the most stable cost free way to software midi cable everything up, including a full GS softsynth?
4) is there anything i don't know about yet?

Reply 1 of 6, by Shponglefan

User metadata
Rank l33t
Rank
l33t
zaphod77 wrote on 2023-11-26, 20:45:

I believe it was determined that munt was legal, due to the original mt-32 rom technically having no copyright because it was not registered.

Registering a copyright isn't required for something to be protected by copyright. I would assume the MT-32 ROM is copyrighted unless it was explicitly released into the public domain.

Pentium 4 Multi-OS Build
486 DX4-100 with 6 sound cards
486 DX-33 with 5 sound cards

Reply 2 of 6, by darry

User metadata
Rank l33t++
Rank
l33t++
Shponglefan wrote on 2023-11-26, 20:59:
zaphod77 wrote on 2023-11-26, 20:45:

I believe it was determined that munt was legal, due to the original mt-32 rom technically having no copyright because it was not registered.

Registering a copyright isn't required for something to be protected by copyright. I would assume the MT-32 ROM is copyrighted unless it was explicitly released into the public domain.

I am not a lawyer and am going from memory, so I could be wrong. Please seek qualified legal council if you want any of the assertions below confirmed or infirmed.

AFAICR, it used to be required, laws changed and there was some legal legwork needed for works that were in a grey zone of sorts. This may or may not have taken place for MT-32 ROMs, so their status is, at best, unclear. There was a discussion about this here 20ish years ago. Search for posts by canadacow and you should find it.

Munt itself, AFAIK, is in the clear, because

a) no potentially copyrighted ROM data is included with it.

b) any theoretically relevant patents that might have covered something regarding the MT-32's hardware design are likely expired by now. I have no idea if software patents were even a thing back then (and that might vary according to one's jurusdiction), but I would guess they would probably be expired as well. In doubt, consult a copyright and/or patent specialized attorney.

Reply 3 of 6, by zaphod77

User metadata
Rank Newbie
Rank
Newbie

there was a big kerfluffle with roland over the issue if whether the original mt32 rom was actually protected by copyright, having to do with the laws at the time it was made.

The original mt-32 was made and released in the USA BEFORE the requirement to register a copyright to enforce it was dropped. So by failing to actually register the USA copyright it was by law not protected at the time.

While it was registered in Japan, it wasn't registered in the USA, and predated adoption of the Berne Convention there, which is what makes registration not required for modern copyrights.

That said, the lawmakers realized that this was not fair to copyright holders, so there is a way for such orphaned copyrights to be restored, and regain the protection they should have under the Berne Convention. There are four tests. here they are in simple english

1) Copyrighted in Berne Convention Country not the USA. Passed. Was registered in Japan
2) Copyright not expired. Passed. (yup. still hasn't expired.)
3) Does not currently have copyright in the usa because the legal requirements were not filled. (true, and the reason for the entire problem. the copyright did actually need to be registered in the USA at the time for it to be valid!)
4) must have been published in other Berne Convention country 30 days or more BEFORE it was published in the USA. This one is the kicker, and the true intent test. If it was published later the 30 days, failing to register is considered an honest mistake, and the work deserves restoration. But if it was published within the 30 days, then it's presumed that they were aware registration was required, and the previous law holds due to the US prohibition against ex post facto laws.

The issue was apparently that it was released in the USA within 30 days. The emulator writer said "sorry, but you can't stop us and reclaim the copyright unless the MT-32 was brought to the US 30 days later. prove that and we'll remove the ROM from our downloads." Roland went and checked their records, and found that they couldn't prove requirement number 4, and said GOD DAMMIT!

Hence the argument that the mt-32 rom is under no copyright in the USA, because Roland failed to register it at a time where registration was, and they were unable to restore it.

This is a really fiddly argument, and most definitely against the spirit of the law, but the emulator developer that advanced it seemed to have won at the time.

Regardless of the status of the original rom, the later ones are indeed protected by copyright in the USA for sure.

Any patents are expired for sure.

Roland seems to have eased up on things somewhat, but if anyone actually tries to make their own mt-32 clone outside of software emulation for playing old computer games in DOSBOX, they are still gonna get mad and take action, and despite the issue with registration would probably win in that case. Witness the situation with the dreamblaster cards.

Reply 4 of 6, by midicollector

User metadata
Rank Member
Rank
Member

It’d end up going to court, and Roland would probably win, so just depends on if anyone wants to spend a bunch of money on lawyers just to lose in the end. That’s assuming Roland notices, but once they send a cease and desist, the game would be up unless you want to spend a bunch of money on lawyers.

Reply 5 of 6, by zaphod77

User metadata
Rank Newbie
Rank
Newbie

they pretty much aren't bothering now, but only because they aren't selling a solution themselves for dosbox (Roland Virtual MT-32 doesn't exist at all yet), so as long as nobody is actually profiting from this except gog.com (who is selling those old games still), they seem to be willing to look the other way for now. And while there do seem to be MT32-pis that are munt based around, no one is actually using them as a real instrument, which is the same condition attached to the dreamblaster wavetable daughterboards, and nobody is selling them that I know of, because THAT would actually get Roland's attention, and make them rightfully demand a license at least.

Reply 6 of 6, by zaphod77

User metadata
Rank Newbie
Rank
Newbie

okay so tts-1/hypercanvas is not actualyl GS compliant.

So we are back to square one on the GS it looks like. there is no legal option except to buy sound canvas VA (which is the only one that doesn't suck) or a hardware sound module. And the only affordable one of the latter is MAYBE the new gs dreamblaster wavetable boards, and i'm not going to buy one just to test rtft.mid on.

That said, if someone were to write a fully gs compliant softsynth driver or vsti for windows that loads GM.DLS for the sample data, that would be legal, and probably good enough for my purposes. The license for gm.dls literally only lets you use it in conjunction with microsoft operaing systems, and i know of two. actual ms-dos and windows.