Forums ā† News ā† Soundplane version 1.6 update

Soundplane in the lab
randy
Wed, Dec 20, 2017, 11:26

Another Soundplane software update is available. This release is recommended for all Soundplane owners. It's coming out hot on the heels of the 1.5 version in order to fix a problem that some users were having under Mac OS 10.10.

To keep up with Apple's code-signing recommendations I'm now distributing the application as a signed .dmg file. Just drag the Soundplane app to the Applications shortcut to install in typical Mac fashion.

Finally, latency is also decreased with this version, for the first time since the initial release! I hesitate to throw out precise numbers yet, because latency measurements are hard to do right. Roughly speaking, an underlying isochronous USB layer that added 10ā€“20 milliseconds of latency previously is now down to just a few ms. I'm looking forward to doing some accurate tests. Maybe this will make a cool movie?

For those of you who write computer program and are doing things like compiling Soundplane apps for Linux, the C++ API is now settled down, so that you won't have to keep up with any further changes for a long while. As always the code is available on github: https://github.com/madronalabs/soundplane

Direct link: Soundplane 1.6.0.dmg, 5.9MB

version 1.6 changes:

  • improved isochronous USB driver: lower latency, allows restart after pausing app
  • fixed an issue selecting cleanest carriers after "select carriers"
  • now distributing as signed .dmg
  • fixed a possible crash on shutdown

Stay tuned for a maintenance update of Aalto, Kaivo and Virta very shortly.

thetechnobear
Thu, Dec 21, 2017, 06:49

ooh, great news, can't wait to try this...
thanks for the continued development.

look forward to digging into the source again in the new year.

thetechnobear
Thu, Dec 21, 2017, 07:05

oops, bad news... the OSC t3d output seems to be broken...

its not working with Aalto properly, and when i look at the underlying osc message I can see its hanging on to notes that have been released, and its transmitting all 16 touches on every update.

ive tried resetting to factory defaults , recalibating, changing touches, rotate - all make no difference.

note: the touches display looks absolutely fine, so seems to be OSC.
note2: though ive not tested MPE... so could be interface between the touch model and the outputs.

Im going to have to rollback, or rather reinstall 1.5 for now :(

randy
Thu, Dec 21, 2017, 08:28

Oh no! I'll investigate today, thanks for the report.

randy
Thu, Dec 21, 2017, 19:54

What I found out today is that there is an intermittent problem. I can start up the Soundplane app and have it work fine, then start up again in the same way and it gives the symptoms you are describing. This is why I was able to test and think it was functional. I'll work to fix this ASAP.

randy
Thu, Dec 21, 2017, 19:54

What I found out today is that there is an intermittent problem. I can start up the Soundplane app and have it work fine, then start up again in the same way and it gives the symptoms you are describing. This is why I was able to test and think it was functional. I'll work to fix this ASAP.

thetechnobear
Fri, Dec 22, 2017, 09:03

ah, those are the pesky kind of bugs...
hope the fix is not too difficult.

thetechnobear
Wed, Jan 17, 2018, 04:17

Have you had a chance to fix1.6 yet? Iā€™d love to move to it, and also look at porting my code to it.
Thanks Mark

randy
Wed, Jan 17, 2018, 07:53

I ended up taking a deep dive into things and fixing some other issues that made development slow. (Kernel panics when debugging, no more!) I will be wrapping the app up and sending out a beta this week. Since the internals changed so much, I'll be upping the version to 1.7 (1.6, we hardly knew ye) and starting afresh with a proper beta.

In addition to the improvements to lower layers I did find the bug that was causing the intermittent issues with output.

thetechnobear
Wed, Jan 17, 2018, 08:29

he he, i know how that goes ;)

what was the issue with the kernel panics? when i looked at this (on eigend) it was due to completion events being returned to a process that had been stopped/killed ... so all i could do was to ensure the process closed cleanly.
did you find another fix?

randy
Thu, Jan 18, 2018, 08:07

The low latency isoch callback routines are called by the Mac at main interrupt time in order to allow low latency, which makes them kind of dangerous. Being careful about when these are called and comprehensive about their errors are handled seems to have eliminated the crashing.

randy
Sat, Jan 20, 2018, 19:42

Found another couple issues I need to fix before sending out the beta. Should be able to take care of them Monday.

randy
Wed, Jan 24, 2018, 20:50

and done, see hardware forum.