I've done yet more work on the Soundplane code. In addition to fixing an intermittent error that a few people ran into since version 1.5, this 1.7 release adds new Mac system code to get the lowest possible latency.
- allow running app hidden / in background without added latency
- fix redundant calibrate on startup
- fix a stuck notes issue
- add quick recovery after unplug / replug instrument
OSC has been tested here fully, MIDI less well. Please give it a whirl and let me know about any issues.
Full Kyma functionality may or may not be restored with this release. Unfortunately, I can't test that here. I'll try to get some quality time with a recent Kyma in the near future.
ok, the OSC kind of works ;)
but you have something odd going on with the port offset....
the default for t3d is 3123, but you seem to be using 3125?
anyway the upshot is, is you need to set the osc offset in Aalto to 2 to work.
(using aalto 1.8.0, Kaivo 1.3.0)
from logs, look like you are connecting from 3125 onwards, and created an inbound socket for 3124... not looked for why this is yet :)
(also default in the dropdown appears to use 3125)
perhaps a simple change to move 0 back to 3123, and perhaps move your input port if required to 3122.
Starting Soundplane... SoundplaneModel: listening for OSC on port 3124... MLOSCListener: trying listen on port 3124... MLOSCListener::listenToOSC: created receive socket on port 3124. initializing pthread attributes... creating listener thread... MLOSCListener running socket. --------------- SoundplaneOSCOutput: trying connect to ports starting at 3125 connected to port 3125 connected to port 3126 connected to port 3127 connected to port 3128 connected to port 3129 connected to port 3130 connected to port 3131 connected to port 3132 connected to port 3133 connected to port 3134 connected to port 3135 connected to port 3136 connected to port 3137 connected to port 3138 connected to port 3139 connected to port 3140
btw: any chance you could move the forum software over to Discourse, it really is much better forum software... and formatting options (for code) are much better ;)
anyway, this was just a quick sanity check... will hopefully have time to sit down for some quality time with the soundplane (on 1.7.0) later today :)
Great, I look forward to further results form your end. I did get one stuck note here playing last night.
Sorry I forgot to mention the port number change. I will change it back for the time being.
Soundplane listens for Kyma on port 3124, I was testing a Kyma fix.
Work is currently being done on the website with a partner and we're considering what to do with the forums in the future.
I posted 1.7.0b2. This restores the port numbers.
It's the same link as last time, however I am updating the beta version, which you can see in the app.
cool, much better :)
had a play over the last few days, and the tracker does feel better and in particular the top and bottom rows are much more reliable... i can now use them, before I found them too likely to trigger the row above - so this is excellent.
a couple of small bugs:
first, now only 3123 works , if you select another port (e.g. 3) in aalto etc, then SP app shows the new port (aalto(3) , but when you select it, you dont get any sound, and in fact its still sending OSC messages on 3123
second, i think your sending empty osc packets continuously?
(or the frame bundle has a bunch of empty messages)
if you do an oscdump you will see
/t3d/frm ii 10209708 65590 / / / / / /t3d/frm ii 10209724 65590 / / / / / /t3d/frm ii 10209740 65590
Thanks for the report, I appreciate the help.
I'm back on working on my projects involving Soundplane and thus am testing 1.7b2. First of all, thanks for your work on this !
Beside the lonely / that get sent as mentioned above, I also notice that now touch message are not get sent when there aren't any. I like this ! So I dont have to filter out zeroed Z touches in my Max patch.
Just in case, is it now the expected behavior ? I just thought I'd ask for confirmation before changing all my Max patches. :-)
This is the expected new behavior, more or less. One issue is that the present code can result in rare stuck notes due to UDP dropouts. I think it may make sense to fix this in the sender as well as the receiver. So I may send intermittent (but far less frequent) zero z frames if a voice is not active.
NOTE: the 1.7.0 release is final and available from the "STICKY: Soundplane client for Mac" thread.