[mrfatboy]

@Memphiz

I subscribed to this thread but I just only now got a email notification after the last 3 days I would like to help narrow down the the airplay bugs f possible. Which one are we exactly testing for now (visualizer crash?, refresh rate?, etc) Just let me know what you would like to test for and what version.

[SadGamerGeek]

Which one are we exactly testing for now (visualizer crash?, refresh rate?, etc) Just let me know what you would like to test for and what version.

For clarification, when I talk about "fine" / "breaks" above I'm referring to the bug where the audio starts stuttering (repeat loop) and xbmc completely locks up (needs to be killed from Task Manager).

[Memphiz]

Yep and thats what this thread is about. Because that was the initial bugreport in this thread. Well its really nice that you guys try to "track it down". But tracking down would mean compile each revision of XBMC between jul 7th and jul 13th and check.

I guess nobody of you will do this. There are some serious changes inbetween this days and unfortunatly the buildbot was broken and broken again after each day so we don't have any nightlies in that time range .

So i guess its on me and i'm really too busy for such an time intensive hunt (and i HATE windows - i normally need 2 months vacation once i have digged under windows because its really annoying in a VM - well its annoying outside a VM too *haha*).

What would be helpfull though - if this issue is only present on windows or also on other platforms (linux, osx, ios).

[SadGamerGeek]

Yeah, you're right about the compiling I'm afraid - that sounds like something best not attempted for a non-developer such as myself.

I'm pretty much a Windows household I'm afraid, so I'll get a Linux VM going under VirtualBox and try it on that.

[0wing]

I took some time to investigate.

This problem only exists when you use wasapi, directsound works fine.

It seems like https://github.com/xbmc/xbmc/commit/8ac5...ab24a092eb introduced the problem.

With a build based on 0ed0f26c11, the previous commit, I was able to use airplay for 10+ minutes several times without any problem.

[Memphiz]

0wing - cool thats a attack point for sure. Does the issue manifest when playing mp3 audio too? I guess not?

[Memphiz]

@SadGamerGeek if you didn't already try on linux (or even have to setup the VM first) - i don't think i see that issue on linux ... airplaying music for about 7 mins now with multiple tracks and it seems stable... Not sure if i need to do anything special for triggering it (just pushing music from the ipod and having xbmc in fullscreen with spectrum vizualisation).

[0wing]

0wing - cool thats a attack point for sure. Does the issue manifest when playing mp3 audio too? I guess not?

Mp3s have the same issue apparently, didn't test that before.
Needless to say, without that commit they also play flawlessly.

[SadGamerGeek]

I took some time to investigate.

This problem only exists when you use wasapi, directsound works fine.

It seems like https://github.com/xbmc/xbmc/commit/8ac5...ab24a092eb introduced the problem.

With a build based on 0ed0f26c11, the previous commit, I was able to use airplay for 10+ minutes several times without any problem.

That's great news! I can confirm that directsound works fine for me too - cheers.

@SadGamerGeek if you didn't already try on linux (or even have to setup the VM first) - i don't think i see that issue on linux ... airplaying music for about 7 mins now with multiple tracks and it seems stable... Not sure if i need to do anything special for triggering it (just pushing music from the ipod and having xbmc in fullscreen with spectrum vizualisation).

No, nothing special need to trigger it (though changing the play position in the scrub bar seems almost guaranteed to make it happen immediately). I hadn't had chance to build the VM yet, and 0wing's discovery did make it seem to be Windows only. I might install the VM anyway, but I won't hurry now!