Crashed PLayback Queue

Hi @PG1

I have some crashes with WaveLab 13.0.20

Before these latest crashes, I completely reinstalled WaveLab 13. I also updated the plug-ins used in the Master Section and in the Playback Processing chain, including Perception AB 1.1.15.

All reports show the same pattern: (also a crash from april is similar before 13.0.20)

  • crashed thread: Playback Audio Queue;
  • EXC_BAD_ACCESS / SIGSEGV;
  • invalid read at address 0x8;
  • rax = 0 and r14 = 0;
  • the failing instruction reads from [r14 + rax + 8];
  • the crashed stack contains only WaveLab and Qt code.

The three WaveLab 13.0.20 reports use the same WaveLab offsets and the same instruction sequence. The crash usually occurs within some minutes after reopening the montage.

The main thread was executing different code in the different reports:

  • Perception AB;
  • Dynameter;
  • Sima GUI drawing;
  • WaveLab/Qt window-layout code.

The plug-ins were loaded, but none appears in the actual crashed thread.

Assumption is that WaveLab’s playback queue accesses a null, invalid or already-destroyed internal object. This could be a race condition occurring when the playback graph is rebuilt, stopped or reconfigured.
i’ve copy paster all the master files un a new WL projet and working without pluggin in the master section, hoping to avoid crashes.

The montage contains many plug-in instances, which may make the problem easier to trigger. However, the reports do not show an out-of-memory condition or a crash inside the audio driver.

I am attaching the crash reports.
in case i could sent you in private the project.

Best regards,
Olivie

Thank you for the data. I ignore the origin of the problem.
Have you started to use WaveLab 13 recently?
Next week, there is a WaveLab 13.0.30 update; check with this one.

It’s in WL13.
(The April report is from WL12.) All the others are from today, running WL 13.0.20.
I thought it was meaningful that I had a very similar crash in WL12.

I cleared the Master Section and copied all the clips into a new montage, and it’s back on track.

So we could probably assume that something is happening that corrupts the projects. Most of the time, it happens on long montages (>60 minutes).
I could share the montage if it would be useful to you ?

Thanks, but I don’t think this would help.
Is there anything special in the montage? More than Stereo? high sample rate such as 96k?

nope, basic 48k 24b tracks.
nothing fancy