07-10-2013, 02:58 AM
07-10-2013, 03:00 AM
It is my understanding that all backends experience the issue so I am not sure why the time stretch was needed. delroth or degasus may have to be the ones to tell us that
07-10-2013, 03:24 AM
Time stretch isn't needed at all. The broken part was the resampling in Mixer.cpp. I only try to add time streting because resampling is a free operation with time streting. In fact, both can be done with the SoundTouch library (already used by openAL).
So we either reimplement our resampler, or we include SoundTouch for it and get time streting for free.
btw: SoundTouch could also be used without time stretching. In theory it should be possible without unneeded latency. But time stretching always has latency (eg 50-100ms) which will be added to the backend latency :-(
So we either reimplement our resampler, or we include SoundTouch for it and get time streting for free.
btw: SoundTouch could also be used without time stretching. In theory it should be possible without unneeded latency. But time stretching always has latency (eg 50-100ms) which will be added to the backend latency :-(
07-10-2013, 03:53 AM
Is testing for OSX still needed? I tried to do it yesterday, but wii-mote got disconnected, so when i came back is said reconnect wi-mote.
07-10-2013, 04:06 AM
degasus,
It sounds like the best option would give folks the freedom to select whether they want time stretching enabled or to disable it. Wait...since we brought back OpenAL and it already has time stretching implemented, is there a pressing reason to even add time stretching?
I am just really hoping to have no latency on those sound effects if it is possible
I just thought of a side note. Wouldn't this officially close all those similar issues open about this one problem? I could try to get a list together of all those if it makes it easier (assuming that is even something the devs are interested in).
It sounds like the best option would give folks the freedom to select whether they want time stretching enabled or to disable it. Wait...since we brought back OpenAL and it already has time stretching implemented, is there a pressing reason to even add time stretching?
I am just really hoping to have no latency on those sound effects if it is possible

I just thought of a side note. Wouldn't this officially close all those similar issues open about this one problem? I could try to get a list together of all those if it makes it easier (assuming that is even something the devs are interested in).
07-10-2013, 10:03 AM
Just use the first build he posted with the 1h30m bug fixed but with no timestretch.
07-10-2013, 10:13 AM
I am 
I was just *hoping* that the time stretching wasn't going to make it into master as a requirement for all backends.

I was just *hoping* that the time stretching wasn't going to make it into master as a requirement for all backends.
07-11-2013, 04:30 AM
degasus said something along the lines of "fuck timestretching, I'll just fix the bug for now" earlier today, so I think you'll be fine. :D
07-11-2013, 05:06 AM
Best quote ever
07-12-2013, 01:42 AM
Quick question,
Are there any estimates on when this change will be added to master (assuming it will be)?
I understand it may take time, but once 4.0 comes out, I wanted to make sure it had this type of fix in it
Are there any estimates on when this change will be added to master (assuming it will be)?
I understand it may take time, but once 4.0 comes out, I wanted to make sure it had this type of fix in it
