From: Silvan <dmm...@us...> - 2003-02-21 17:03:44
|
We've established that the drum kits for this soundfont are on "bank 128.= " I've been surfing, and have concluded that MSB 1 LSB 0 =3D "bank 128" (http://www.kurzweilmusicsystems.com/html/pc2_tutorial_24.html): =2E.. Controller 0/32=09=09Bank Number Equivalent 0/127=09=09=09127 1/0=09=09=09=09128 =2E.. This soundfont has a drum kit on bank 128, program 0 Bank 128: Program 0: "Drum [Stan1] -SC88P" I'm transmitting MSB 1 LSB 0 Prg. 0 rosegarden: sending bank select 1 : 0 rosegarden: sending program change 0 Looks like I have my ducks in a row, but there ain't no drums, Chief. (I= also=20 tried flipping LSB/MSB, to no avail.) Am I missing something? Any idea why this isn't working? I really want to play with these drum kits. An added advantage is that s= ince=20 these are in their own bank, Rosegarden can provide a logical way to name= the=20 kits without any need for Rich to do extra work, if I can ever figure out= how=20 to get to them.... --=20 Michael McIntyre USDA zone 6b in SW VA, USA Silvan <dmm...@us...> Linux Druid ----------[ registered Linux user #243621 ]--------- http://www.geocities.com/Paris/Rue/5407/index.html |
From: Chris C. <ca...@al...> - 2003-02-21 20:47:19
|
On Friday 21 February 2003 17:02, Silvan wrote: > rosegarden: sending bank select 1 : 0 > rosegarden: sending program change 0 > > Looks like I have my ducks in a row, but there ain't no drums, > Chief. (I also tried flipping LSB/MSB, to no avail.) > > Am I missing something? Any idea why this isn't working? Are you getting any sound at all (i.e. the wrong bank or program) or=20 just nothing? You run SF2PatchExtractor on the .sf2 file and it prints out bank 128=20 program 0 for some drums or other, right? It's probably worth trying=20 it in Swami (which can play the samples direct from the soundfont=20 using an embedded iiwusynth) just in case it's some sort of stupid=20 bug in SF2PatchExtractor that's printing out the wrong numbers. If=20 you haven't tried that already, that is. Swami can also show you=20 exactly which pitches on that program actually correspond to samples:=20 for all we know there might be nothing actually there (though I'll=20 admit that's one of the less likely alternatives). Besides, Swami's fun anyway. Chris |
From: Silvan <dmm...@us...> - 2003-02-21 23:50:51
|
On Friday 21 February 2003 03:44 pm, Chris Cannam wrote: > Are you getting any sound at all (i.e. the wrong bank or program) or > just nothing? No sound at all. Not unless... Hang on... Yeah, not unless I then tran= smit=20 bank 0 0 and a patch, then channel 10 is just another channel again. > You run SF2PatchExtractor on the .sf2 file and it prints out bank 128 > program 0 for some drums or other, right? It's probably worth trying I don't have Swami, but I looked at the soundfont with Smurf, and it was=20 indeed bank 128. I have two more similar soundfonts from completely different sources (sim= ilar=20 in having multiple drum kits in bank 128.) Your little program and Smurf= =20 both agree that the percussion is "bank 128" and neither one of them work= s=20 either. Exactly the same behavior. So probably it's an iiwusynth proble= m,=20 since Rosegarden clearly appears to be doing the right thing. I'm just trying to make sure I am indeed doing the right thing in Rosegar= den=20 before I start looking at iiwusynth as the culprit... > Besides, Swami's fun anyway. I'll go dig it up. --=20 Michael McIntyre USDA zone 6b in SW VA, USA Silvan <dmm...@us...> Linux Druid ----------[ registered Linux user #243621 ]--------- http://www.geocities.com/Paris/Rue/5407/index.html |
From: Silvan <dmm...@us...> - 2003-02-22 02:35:46
|
I never did manage to get CVS. Their server keeps timing out on me. May= be a=20 bug they already know about and fixed... I'll look into it more deeply=20 later. Anyway, for those of us still running the old 0.2.2 official release, the= =20 patch is so simple it's not worth the trouble of doing a patch. From the= =20 root of their source tree... sed 's/msb << 8/msb << 7/' src/iiwu_chan.c > foo&&mv -f foo src/iiwu_chan= =2Ec It was shifting one bit too far, and that's why it was doubling. That's = what=20 it looks like at first glance anyway. There may be some deeper reason wh= y it=20 was the way it was, but all I care about is that my stupid drums are work= ing=20 now. :) Also, the bug wasn't with Rosegarden! --=20 Michael McIntyre USDA zone 6b in SW VA, USA Silvan <dmm...@us...> Linux Druid ----------[ registered Linux user #243621 ]--------- http://www.geocities.com/Paris/Rue/5407/index.html |
From: Silvan <dmm...@us...> - 2003-02-22 00:37:48
|
On Friday 21 February 2003 03:44 pm, Chris Cannam wrote: > Besides, Swami's fun anyway. OK, Swami:Smurf::Fluid Synth:iiwusynth, so basically I more or less alrea= dy=20 had it. I'm getting the new version anyway because it looks cool, and I = may=20 as well learn how to use a current-generation tool. Anyway, I thought to try the -V option to iiwusynth, and here's what it s= ays=20 when Rosegarden transmits MSB 1 LSB 0 Prg. 0 iiwusynth: prog 9 128 0 128 is the bank, so that's all good. 9 is the channel. Are channels 0 b= ased=20 too, so that we send a 9 for "channel 10" and a 0 for "channel 1" ? Anyway, I'll keep dicking around with this for a bit yet and see if I can= =20 figure it out, but if I don't figure it out soon I really should find=20 something more productive to do. --=20 Michael McIntyre USDA zone 6b in SW VA, USA Silvan <dmm...@us...> Linux Druid ----------[ registered Linux user #243621 ]--------- http://www.geocities.com/Paris/Rue/5407/index.html |
From: Mark K. <mk...@co...> - 2003-02-22 00:44:35
|
> 128 is the bank, so that's all good. 9 is the channel. Are > channels 0 based > too, so that we send a 9 for "channel 10" and a 0 for "channel 1" ? > Probably yes. RG used to be this way (Channels == 0-15) but was changed a few months ago to follow the way hardware synths work. (1-16) The value transmitted on the MIDI bus is 0-15, and your soft synth may work this way. Good luck, Mark |
From: Silvan <dmm...@us...> - 2003-02-22 01:08:04
|
Wait a minute... I looked at this again... iiwusynth: prog 9 128 0 That's its initialization state... Putting the "standard" drum kit (bank= 128,=20 program 0) on channel 10 by default.=20 Then we change it to bank 0-0 prog. 0 and get: iiwusynth: prog 9 0 0 (It plays a piano...) Now I flip to the MSB 0 LSB 0 bank, prog. 1 (actually 0) and iiwusynth says: iiwusynth: prog 9 256 0 Bingo. Where's the 256 coming from? Odd that RG could have such a huge problem, seeming on the surface to be=20 doubling the MSB value (2-0 is bank 256). It can't be thus, because my G= S=20 bank works fine (just confirmed), and everything in GS is selected on the= =20 MSB, ergo a simple +1 problem cannot be. Mysterious. So if it thinks 1 0 is 256, it would think 0 63 is 128 then, maybe... iiwusynth: prog 9 63 0 Nope. 0-63 comes out right. But if I try 2-0 (should be 256) I get iiwusynth: prog 9 512 0 I wonder who's doing the wrong math here. 127-127 should be bank number 16,384 It yields: iiwusynth: prog 9 32639 0 Doesn't make any sense at all, that one. Doesn't look like a sign rollov= er,=20 so why isn't it 32768? One of us has a serious bug, and since the GS stuff works flawlessly, I t= hink=20 iiwusynth must be borked. I'll bet _it_ has the +1 problem. I'll go try to find it. Rosegarden benefits indirectly. The other stuff= I=20 was going to do can wait a bit yet... --=20 Michael McIntyre USDA zone 6b in SW VA, USA Silvan <dmm...@us...> Linux Druid ----------[ registered Linux user #243621 ]--------- http://www.geocities.com/Paris/Rue/5407/index.html |