Virtual Theatre Pipe Organs and Puppy Linux 4.0

[Miditzerman]

I have a question for Graham regarding the splitting of the audio output from LinuxSampler or maybe someone else if he doesn't know yet. First I will explain what I am doing now.

The standard Miditzer output is from a fluidsynth to a stereo output and 2 speakers. This means that Diapason, Flute, Viol, Celeste and Clarinet are on the left audio channel, and the rest are on the right channel. This means that the signals are being mixed electronically and the results are not the same as mixing them acoustically.

I have a Delta1010LT soundcard which has 8 audio outputs channels, or if you like 4 stereo outputs. These feed 4 computer 3-way multimedia amp/speakers. I want to send these 5 ranks down separate channels and send the Tibia, Vox and Oboe down three more channels.

This is how I am doing it now. I am running 4 instances of Qsynth each loaded with the full soundfont because the soundfont comes as a package. These show up as Qsynth1, Qsynth2 etc. In Jack I connect the O/P of Q1 to ch 1 and 2 of the Delta1010LT, Q2 goes to D1010 ch3&4 etc. Ch1&2 feeds the first Stereo Amp and CHs3&4 feed the 2nd Stereo Amp etc.

The Flute is in the Left Chamber, so is LH audio channel and is output as MIDI ch1 from Miditzer so I send it to MIDIch1 of Qsynth1 which goes to ch1 of the soundcard and to the lh channel of the first pair of stereo speakers. Diap is LH and MIDIch2 so I send the MIDI ch2 to Qsynth2 etc. This means ch3 of the soundcard and ends up at the LH ch of the 2nd pair of stereo speakers. So not only are they coming out of seperate Audio channels but they are being produced by different Synths. Plus each Qsynth can easily manage 256 note polyphony. I hope this makes sense.

I want to get the same result with LinuxSampler and I tried to start a second LinuxSampler but it doesn't show. I only get one. Maybe I don't need to do this with Sampler. Can I route them some other way?

Regards

Russ

[Alex]

Russ,
I'll assume here that you're going to use gig files, and proceed with a suggestion from there.

Open an instance of Fantasia, create an audio device, and a midi device.
Once created, go to the audio device, and increase the number of ports to reflect your wish, and in this instance we'll say 4 stereo pairs, or 8 mono.

Create new channels strips, one for each section, and on the right hand side of the section, click on the little arrow that expands the channel to reveal routing and assignment options. Assign a midi channel to each strip, (Fantasia will do this for you, at least for the first port), then go to the bottom right of the channel strip and assign your audio outputs. They come in pairs in the assignment box, so simply label them both the same if you wish a mono output, for each channel.

Do the same for the remaining channel strips, assigning the relavent midi and audio. The outputs will appear in Jack, same as Qsynth, and you can route them as normal.

I gather from what you've written, that you only use 1 midi channel per section, Flute=1, vox=2, etc.

You won't need 4 instances of your miditzer package (gig format), as unlike qsynth, Lsampler will handle a large voice count according to your compile time options when you built LSampler.

If i've misunderstood your original post, and you in fact require 4 ports, each with 16 midi channels, then you can increase the port count in the midi device.


Alex.

[ggoodesa]

Hi Russ,

Alex is right about this, but I have never had the chance to try it as I only have 2 channels available on my soundcard. Once you've set the Audio device to 4 or 8 channels you'll find them in the 'Connections' screen in Qjackctl. You can do this in Qsampler as well, so if you need instructions just let me know.

GrahamG

[Miditzerman]

Thanks Alex,

Graham has got us using Qsampler at the moment but I will see if I can find the bit that I can change or transfer it to Fantasia. The next Alpha will use Fantasia though I think, along with some reverb. Yes we are using GIG samples.

I guess what I need is 4 groups of stereo audio outputs that Jack can recognize so that I can route each to a stereo amp. With the Qsynth, the Miditzer soundfont came as a package so I loaded the whole thing in each and just took out the ones I wanted, or rather sent the MIDI signal to the Qsynth that I wanted the output from.

Now we are using GIG files of individual voices. If I can put each "Rank" in the appropriate Sampler "box" that is connected to its amp by Jack then I can just load the sample that I want into the MIDI device. I think.

I will experiment over the weekend and let you know how it worked out.

Thanks

Russ

[Miditzerman]

Hi Russ,

Alex is right about this, but I have never had the chance to try it as I only have 2 channels available on my soundcard. Once you've set the Audio device to 4 or 8 channels you'll find them in the 'Connections' screen in Qjackctl. You can do this in Qsampler as well, so if you need instructions just let me know.

GrahamG

Hi Alex and Graham,

I got it working with Qsampler on UbuntuStudio.

Actually I used 3 devices with 4 amps. One amp is on the "console" and is connected to all 3 devices, it also has the headphones so I use that one for practice. The other 3 amps and speaker sets are up on a shelf above my head and are each connected to a device so that the voices are split. This one is louder, sounds better and is for when I am feeling bolder and there isn't anybody else around.

Next problem is how to control the volume from the swell pedals using the output from Miditzer.

I have also sorted out the GIG samples problem. Use the stereo samples listed on Joe Hardy's sites
http://www.golfcap-rouge.qc.ca/mdz/ugsf.asp not the mono ones unless there isn't a stereo one. The other mono samples seem to be one end of a stereo set of a rank and they alternate end to end.

Now I have to do it all again with Puppy.

Russ

[ggoodesa]

Hi,

I have jconv working in Puppy, but I want to get a collection of IR wav files together and create some example .conf files with them before I release a .PET install file. Any favorites being used? Any hints and comments on configuring conf files, etc would be welcome...

GrahamG

[Miditzerman]

Hi Graham,
I notice that the swells aren't working, and I am guessing that they aren't hooked up yet.

Miditzer uses CC11 and CC7 for the Main and Solo chambers to control both volume and high frequency cut-off. Also the Crescendo uses CC1. Is there any way to do this with LinuxSampler?

Russ

[bartonbrass]

I had the swells working under winXP with Linuxsampler, Fantasia and midiox. I don't know if Midiox will run under linux?
Havn't been abel to play with my MidiTzer for some time, Been watching grandmother, I have had "very" little time at home..

[ggoodesa]

Yes, MIDI-OX will work in Linux (though WINE)... did you have to map signals or re-route MIDI data?

GrahamG

[bartonbrass]

My brain is fried at the moment, let me see hmmm. I had Midiox map from cc7 to cc11 or was it cc11 to cc7. I'll have to look at Midiox and see If I can rehember what I did.