After a soundtest YAM can't quit.

The YAM Warning Requester

I forget to mention:
Also after 20 minutes waiting and click 'Nochmals versuchen'
i can't quit YAM.

Maybe disabling the button is not the solution?
What happen if a user configure YAM like it play a sound for every
filter action? Then it may be possible that YAM also try to play
some sounds at the same time. but i don't (yet) test it.

(In [5215]) * Threads.c, YAM_UT.c: each thread now allocates a signal which can be used

by the main thread to abort the thread's current action. This closes #212.

I must add that disabling the play button is not possible, because it would not be enabled afterwards as the playback might take an arbitrary long time and the main thread cannot get any specific feedback when the playback has finished.

Additionally, and that's even worse, there are many 3rd party implementations of sound.datatype and each of them handles the playback in a different way. Some use a per object playback task, some others seem to use one global task only. The latter is not able to signal the finished playback if a second sample is to be played in the meantime. Although the playback finishes correctly that initiating task is not woken up. This is now worked around by the abort signal, but the truth is that this sound.datatype is simply broken and there is no way for YAM to tell working and broken versions apart.

I did not yet check how all this works with OS4, but I know that OS4's sound.datatype definitely uses a per object playback task and hence should not suffer from this issue.

I just verified that playing a sample multiple times in parallel is absolutely no problem on AmigaOS4. So this is a bug of the specific sound.datatype used on AmigaOS3.

You might want to try this one ​http://aminet.net/util/dtype/soundDT41.lha

According to the source it uses one playback task per object and hence there should be no interference. I will try tomorrow on my OS3/WinUAE setup.

I also had some thoughts about disabling the play button. This would only work for synchronous playback. I'll see what I can do about that. But asynchronous playback in a thread is something I won't drop again. Playing a sound after receiving some mails should not halt any other action. That's why the threads are necessary.

Playing a sample multiple times in parallel is here also no problem.
But as i said before with YAM it plays only on the left channel if
i get mails.

If the left channel is still playing and i click the test button
again the second sound (the same sound) is on the right channel.
If i click very fast the 3th, 4th (and so on) is also playing.
Some left and some right...

I don't know if it's YAMs fault to play the sound only on the left
channel. All i know is with play16 the same sound is playing on
booth channels.

I just try your suggested Datatype and with them i can play only one
sound and must wait before the second sound start playing. Here the
sound is only on the left channel too.

So i switch back to my old (new) sound.Datatype. The Version i use is:
sound.datatype 44.25 (18.10.2001) i think it's the version from
AOS 3.9 or any BB...

Replying to opiopi:

Playing a sample multiple times in parallel is here also no problem.

Ok, that's good to know.

But as i said before with YAM it plays only on the left channel if
i get mails.
If the left channel is still playing and i click the test button
again the second sound (the same sound) is on the right channel.
If i click very fast the 3th, 4th (and so on) is also playing.
Some left and some right...

It is up to sound.datatype to decide how mono samples will be played. YAM does in no way instruct sound.datatype to use either left or right channels only.

I don't know if it's YAMs fault to play the sound only on the left
channel. All i know is with play16 the same sound is playing on
booth channels.

Maybe Play16 converts mono samples to "pseudo" stereo by playing the same sound on the left and the right channel instead of the first free channel only.