i've finally tracked down some of the infamous yamaha softsynth hacks.

i now have vstis for
s-yxg20 (the basic gm only version of sy50) (didn't have it before, kinda pointless)
s-yxg2006LE (the xglite synth extracted from the midiradio player, already known, better than the above)
ymf-754 HiEnd (sy50xg hacked to use the ymf-754's samples, with a 9 megabyte sample table. plays back xg midis fine)
"Vampire" (seems to the the same as the above, and even calls itself ymf-754 HiEnd still, what's the difference? it's EXACTLY the same sample table, just with a different name)
"Tyrus" (has a gigabyte sample table, presumably ripped from a tyros keyboard? doesn't seem to be as nice on mu-50 composed midis as the others, but does seem to understand the effects. think of sc-55 vs sc-88 pro. the patches are in the right banks for xg playback)

still waiting for an actual mu-50 sample table rip to try and plug into one of these or the s-y50xg portable with external tables version.

zaphod77 wrote on 2024-04-29, 20:59:

i've finally tracked down some of the infamous yamaha softsynth hacks. […]
Show full quote

i've finally tracked down some of the infamous yamaha softsynth hacks.

i now have vstis for
s-yxg20 (the basic gm only version of sy50) (didn't have it before, kinda pointless)
s-yxg2006LE (the xglite synth extracted from the midiradio player, already known, better than the above)
ymf-754 HiEnd (sy50xg hacked to use the ymf-754's samples, with a 9 megabyte sample table. plays back xg midis fine)
"Vampire" (seems to the the same as the above, and even calls itself ymf-754 HiEnd still, what's the difference? it's EXACTLY the same sample table, just with a different name)
"Tyrus" (has a gigabyte sample table, presumably ripped from a tyros keyboard? doesn't seem to be as nice on mu-50 composed midis as the others, but does seem to understand the effects. think of sc-55 vs sc-88 pro. the patches are in the right banks for xg playback)

still waiting for an actual mu-50 sample table rip to try and plug into one of these or the s-y50xg portable with external tables version.

Hi,
I know/have many of them except "Vampire". It somewhat surprises me that it resembles mostly ymf-754. I thought it was even bigger than "Tyrus" with many samples/Midi systems integrated.
Here Roland User's post suggests that "Vampire" has a sample size around ~2GB:
Re: Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi

At one point "tyrus" was switched to the "vampire" engine.

i'm still not sure what that actually means, and all the discussion is in russian, so i can't really join in and ask questions.

Falcosoft wrote on 2024-04-29, 21:14:

Hi, I know/have many of them except "Vampire". It somewhat surprises me that it resembles mostly ymf-754. I thought it was even […]
Show full quote

Hi,
I know/have many of them except "Vampire". It somewhat surprises me that it resembles mostly ymf-754. I thought it was even bigger than "Tyrus" with many samples/Midi systems integrated.
Here Roland User's post suggests that "Vampire" has a sample size around ~2GB:
Re: Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi

Yes , in YAMAHA Vampire Samples size 2 GB , however , because Wlad Rever not want free distribute this synthesier , instead of this Wlad Rever want distribute this synthesizer per donate basis , I as tester was be forced remove Vampire form my SynthPack , however , because the I do SynthPack for me and then share , I replace Vampire to YMF-754 , this was do what not have port without sounds. If you want have Vampire , you can contact me and I remapin you to Wlad Rever )

zaphod77 wrote on 2024-04-30, 03:01:

At one point "tyrus" was switched to the "vampire" engine.

i'm still not sure what that actually means, and all the discussion is in russian, so i can't really join in and ask questions.

If you want , you can ask me , I as much as I can answer to you ) this can do from this topic if FalcoSoft do not mind , or this can do in personal messages )

so yeah. there seems to be no difference between ymf-754 HiEnd and Vampire in that pack, as they both have the exact same wavetable, confirmed by fc /b. is this an error?

No this not error , both synth same
If was be so what I remove Vampire but leave port for Vampire , then you not wouldn't have heard sound on port for Vampire , that's why I maked stub )

Has the topic of pitch bend range sensitivity been brought up regarding the OPL3GM VSTi plugin? So far I've tried 4 different plugins and none of them (OPL3GM included) supports the 14-bit MIDI RPN for adjusting pitch bend sensitivity. At least, not in the traditional method. ADLPlug doesn't, JuceOPL doesn't, and Datajake's OPL3 VSTi doesn't either. Is this something that can be added to OPL3GM or is it already a feature that's accessible in some other way somehow?

It's supposed to work a least in libadlmidi. are you sure you are sending the correct data to do it? you are correct that opl3gm doesn't do it, as near as I can tell.

You know what, I am sending the correct data but as I was typing this response I remembered that Cakewalk stupidly does not send sysex to VSTi plugins. This appears to also be the case with RPNs. Fricking Cakewalk...

yeh. drag and drop the midi onto chiptune.app and select libadlmidi to test.

MusicallyInspired wrote on 2024-05-06, 19:08:

Has the topic of pitch bend range sensitivity been brought up regarding the OPL3GM VSTi plugin? So far I've tried 4 different plugins and none of them (OPL3GM included) supports the 14-bit MIDI RPN for adjusting pitch bend sensitivity. At least, not in the traditional method. ADLPlug doesn't, JuceOPL doesn't, and Datajake's OPL3 VSTi doesn't either. Is this something that can be added to OPL3GM or is it already a feature that's accessible in some other way somehow?

OPL3GM VSTi definitely supports RPN(0,0) pitch bend sensitivity. Valid values are even broader than with most other synths from Yamaha/Roland. Roland usually only specifies 0-24 semitones as valid values.
But you do not have full 14-bit resolution since it is not needed. The LSB part is ignored.

Here is a Midi file for testing that uses 36 semitones as Pitch bend sensitivity right from the beginning on channel 11 (Saw Wave):

apogee driver does support it, but win9x driver does not. I looked in the source.

zaphod77 wrote on 2024-05-07, 07:10:

apogee driver does support it, but win9x driver does not. I looked in the source.

You can also look at the source of the engine of OPL3GM VSTi:
https://github.com/Falcosoft/OPLl3emu4v/blob/ … rc/OPLSynth.cpp
The relevant lines are from 627.

MusicallyInspired wrote on 2024-05-06, 20:31:

You know what, I am sending the correct data but as I was typing this response I remembered that Cakewalk stupidly does not send sysex to VSTi plugins. This appears to also be the case with RPNs. Fricking Cakewalk...

About Cakewalk with use embeded VST Host in Cakewalk I can say what this solution very curve and very bad , this VST Host can not work with XG if need use any banks which different on GM , embeded host easy ignore all cc00 / cc32 messagegs , also embeded host ignore all messages for insert effects group. This I learned from one musician wich try make XG music in Sonar use embeded VST Host for YAMAHA Vampire and YAMAHA YMF-754. That's why I think what you have same problem )

However if use Cakewalk and output MIDI to any other port for example LoopMIDI + SAVIHost , Cakewalk will process all MIDI messages correctly )

Falcosoft wrote on 2024-05-07, 07:17:

You can also look at the source of the engine of OPL3GM VSTi: https://github.com/Falcosoft/OPLl3emu4v/blob/ … rc/OPLSynth.cpp Th […]
Show full quote

zaphod77 wrote on 2024-05-07, 07:10:

apogee driver does support it, but win9x driver does not. I looked in the source.

You can also look at the source of the engine of OPL3GM VSTi:
https://github.com/Falcosoft/OPLl3emu4v/blob/ … rc/OPLSynth.cpp
The relevant lines are from 627.

as i recall you intentionally did this because the real win9x driver also ignored it,

i based this of of

https://github.com/datajake1999/OPL3GM_VSTi/b … 9x/opl3midi.cpp

which doesn't seem to have parser for that RPN

and

https://github.com/datajake1999/OPL3GM_VSTi/b … gee/al_midi.cpp

which does have a parser for it.

zaphod77 wrote on 2024-05-07, 18:36:

Falcosoft wrote on 2024-05-07, 07:17:

You can also look at the source of the engine of OPL3GM VSTi: https://github.com/Falcosoft/OPLl3emu4v/blob/ … rc/OPLSynth.cpp Th […]
Show full quote

zaphod77 wrote on 2024-05-07, 07:10:

apogee driver does support it, but win9x driver does not. I looked in the source.

You can also look at the source of the engine of OPL3GM VSTi:
https://github.com/Falcosoft/OPLl3emu4v/blob/ … rc/OPLSynth.cpp
The relevant lines are from 627.

as i recall you intentionally did this because the real win9x driver also ignored it,
...

Yes, of course. I had no intention to replicate all the bugs and deficiencies of the Win9x driver.
The plugin also supports Modulation wheel controller messages and GS/XG/GM2 style multiple drum channels among others that you cannot find in the Win9x driver.

Hi Zoltan!
You have no desire do expand your plugin WinMultMIDI from 4 parts to 16 parts ? In teory this will give users use up to 16 different synthesizers with use one synth one part and this as I think will be verry cool ) because so can to scale polyphony of synth or we can use one synth per one part and one instrument )
If this do , you create full-fledged MIDI sequencer in one plugin )
Also this give up to 256 MIDI tracks if use per 16 tracks on everyone of 16 ports

libadlmidi supports large numbers of chips. 😀

Yes , but working this very slow

If you sayd about my post to Falcosoft , that you not understand my idea ) I sayd not more chips in OPL3 synthesizer , instead of this I offered make VST MIDI Sequancer in which user can custom route output per channel with select any synth for out )