Hi guys,

I'm quite new to PD and especially with dealing sounds, and I'm trying to deal with multi-input sounds with mics on multi-outputs speakers.
Here's my installation :

• 2 mics connected to a sound card 1
• sound card 1 connected to the linux pc with PD (pd-l2ork-2.4.2)
• linux pc connected to an other sound card 2 that deals with 8 speakers.

What I managed to do so far :

• With one mic connected to the first input I can get the input signal and read it in PD with {adc~]
• I managed to throw this sound live on one of my 8 speakers with a {throw~], this part works fine I can hear myself live on any particular speaker.

Now I would like to implement a kind of switch that allows me to say at any point :

• Mic 1 or 2 is getting sound, throw it to speaker 1 to 8
• Both Mic should be able to speak at the same time (not on the same speaker though)

I tried some stuffs with [spigot~], {*~], [route] but can't figure out the logic of it. I would say that I need to send info with a toggle that this mic and not this one is getting sound, an another info that the sound should be heard on this speaker, but so far can't get it right.

If you have any help it would be welcomed

Thanks !
Florian

@flotac Try [audience~] ......... http://puredata.info/downloads/unauthorized
Here is a patch using it. It will need some other libraries....... audience.zip
There are more up to date solutions here including for 64-bit Pd...... https://forum.pdpatchrepo.info/topic/12280/poorman-s-360-multi-channel-pan
..... and I remember that the grambian library allows to to program pans with messages.
Probably the Audiolab solution can as well.
David.

Ok thank you David I'll have a look at all this. I thought I would have been able to do it without external library but not that easy

@whale-av I tried to follow my first idea for another hour before trying audience and I come up with something working but maybe you will tell me it's absolutely ugly :p
What dou you think ? In this exemple I can have both mics together (even if right now I just have the one mic for my tests) and can manage the switch between 3 different speakers by activating/deactivating the toggles.

@flotac Wow....... that was a big picture that took fifteen seconds to download.
This will switch each mic to one of the eight speakers (I think...... untested).
speaker-select.pd
I have a feeling it might not be exactly what you want though.
If you want to stop both mics talking to the same speaker.......
speaker-select-no-same.pd
David.

@whale-av Sorry for the late reply it's been a busy week I managed to try only yesterday. Thank you so much I just to change a few details and it's working fine!

Quick additionnal question about 2 audio errors that I see in the logs :

First error on startup :
, when puredata is launching at linux startup, I have an error "error: audio I/O stuck... closing audio".
It doesn't appear when I just have just the one audio card connected (the one that deals with out to the speakers). But as soon as I put connect the other audio card (the one that links the mics to the computer), everytime on startup puredata gives me this error. Then if I go into the audio settings, all the params are already ok, but I have to "apply" them again to be able to here my voice and sounds. I would like to avoid this manipulation, would you have any idea?

Second error :
The delay param was set to 25ms (and I have blocksize 64), I had sometimes the error "[2] Opened Alsa Client 128 in:1 out:1
[15] error: audio I/O dropout". I looked it up and people suggested to increase this delay, I put it to 60ms and I hadn't got this error again, but I'm a bit scared it could happen again... any idea?

Thanks again!
Florian

@flotac You can save your audio settings in the patch like this..... set_audio_parameters.pd
click the [audio-properties( message at the bottom right and then set the audio parameters for your soundcards. Then click ok in the audio settings window to close it.
The message below [list trim] will fill with the dialog you need.
Then...... in your patch....... copy that message and connect a [loadbang] to its inlet.
Connect its outlet to an object [send pd]
As your patch starts it will set the audio settings again.
Put a small delay [delay 5] between the [loadbang] and the message if there is still a problem.

Block size 64 is normal. the delay parameter is actually what is called the buffer in the soundcard drivers...... and a larger buffer avoids problems where the audio could be interupted...... but makes for increased latency.
If it is working well at 60ms then it should continue to do so.
David.

Thanks again so much for all this precious info! I'll try that this afternoon

Have a nice sunday,
Florian

@whale-av Thanks a lot again the audio-dialog message with all the params and the loadbang works fine, I don't have to reload the params anymore

@whale-av Hi! If you are still a bit available I still have one bothering issue...

Since I added the new audio card as input device (it's a Tascam us 4x4), I have a lot avec audio I/O dropout. It doesn't bother me that much (don't hear the "audio click" that I've read everywhere), but at some point I think those drop outs generate an error "tried but couldn't sync A/D/A" and "audio I/O stuck... closing audio". At this point when I get these errors I have to reset the audio params (or restart PD) and start my dmx software (QLC+) again so it's a bit tricky.

Would you have any idea how to avoid those errors ? Or is it just to heavy for purr data to have 2 audio cards (one input and one output)

I read this https://forum.pdpatchrepo.info/topic/11806/i-o-errors-in-pd/8 but couldn't figure out an answer.

Thank you very much!
Florian

@flotac I don't use Purr Data but I don't think it should be a problem.
I suggest that you try to set the two sound cards to exactly the same settings in their control panel or config file...... buffer (delay), samplerate, and latency compensation etc.
And then set Pd to match those settings and maybe allow callbacks.
You will almost certainly also need to set one of the soundcards as the "master" for the samplerate clock......... or the computer on-board soundcard maybe?
I don't know how to do that for Linux but there will certainly be some help for that somewhere.
David.