[JensGH]

Here is my log.

http://pastebin.com/VaX1bMC3

[squeezer]

Here is my log - http://pastebin.com/Tte9c25N
I'm using version 11.0 - works without any error, but is not supported now.
With versions 12.x and 13 Playback stutter.

[Memphiz]

No more logs needed for this version. I am cooking up another one with even more printouts. Hold on a bit...

This will lead to a lengthy rinse and repeat session with dozens of xbmc versions with printouts. Not sure how tight i can keep that up as i have a ton of real life stuff atm and are short before vacation too ...

[Memphiz]

Here we go - i need at least one new log with this version:

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

[squeezer]

Here we go - i need at least one new log with this version:

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

So, log was very large and pastbin did not accept his. I uploaded his on mediafire and dropbox
http://www.mediafire.com/view/4j0h8y67zg66rgx/log.txt
https://dl.dropboxusercontent.com/u/59276946/log.txt

[Memphiz]

Thx ... will try to analyse that one as soon as possible.

[Memphiz]

We are getting somewhere. Please one more log with this version (if we are lucky this one will be the last test before i have a fix):

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

fyi ... from my calculation.

1. When we have audio with 32byte framesize (8 channels a 4 bytes per channel/float) and 48khz samplerate we will hit the problem after 46 minutes and 36 seconds exactly.

2. When we have audio with 8 byte framesize (2 channels a 4bytes per channel/float) and 48khz samplerate we will hit the problem after 186 minutes and 24 seconds. (we have reports on ios with streaming problems after 3 hours - that might be the cause as we have 2channels48khz there for sure).

3. 8 channels 44,1khz -> 50 minuts and 43 secs
4. 2 channels 44.1khz -> 202 minutes and 53 secs

This is when using LPCM

Didn't check the numbers on passthrough but i guess i will calc something around 1 hour and 2 mins then ...

[MrMario64]

This sounds awesome! Great work!

[squeezer]

We are getting somewhere. Please one more log with this version (if we are lucky this one will be the last test before i have a fix):

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

fyi ... from my calculation.

1. When we have audio with 32byte framesize (8 channels a 4 bytes per channel/float) and 48khz samplerate we will hit the problem after 46 minutes and 36 seconds exactly.

2. When we have audio with 8 byte framesize (2 channels a 4bytes per channel/float) and 48khz samplerate we will hit the problem after 186 minutes and 24 seconds. (we have reports on ios with streaming problems after 3 hours - that might be the cause as we have 2channels48khz there for sure).

3. 8 channels 44,1khz -> 50 minuts and 43 secs
4. 2 channels 44.1khz -> 202 minutes and 53 secs

This is when using LPCM

Didn't check the numbers on passthrough but i guess i will calc something around 1 hour and 2 mins then ...

Log with last version
http://www.mediafire.com/download/7xmeb7...g/log2.txt
https://dl.dropboxusercontent.com/u/59276946/log2.txt

[Memphiz]

No its definitly not! If it is what i think then it is a Regression i have added myself when trying to fix a crash related to livetv channel zapping. If i revert that fix the 46min problem might be solved but the Crash might come back instead.

That is when a non AE dev tries to dev AE *grrr - annoyed*

[Memphiz]

Ok - it is what i thought indeed

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

This Version should fix it (please verify Once more)

Though that fix can't go in without testing for that pvr Crash (which i can't really remember how to trigger)

[squeezer]

Ok - it is what i thought indeed

http://mirrors.xbmc.org/test-builds/osx/...s-i386.dmg

This Version should fix it (please verify Once more)

Though that fix can't go in without testing for that pvr Crash (which i can't really remember how to trigger)

thank you, thank you, thank you!!! I watched a movie without any errors.
log - http://pastebin.com/E2DN5F1h
will you make full version of this fix?

[Martijn]

Darn, I actually liked this bug

[Memphiz]

I need to verify the fix first (as said - it might introduce a crash again i wanted to fix when i introduced that unwanted regression).

[Memphiz]

should be fixed once this pull request goes in:

https://github.com/xbmc/xbmc/pull/3156