Bug: Warp and latency correction

mr.ergonomics · Post by **mr.ergonomics** » Fri Jan 02, 2009 11:41 pm

I think I found a bug.

1. create two tracks with the same audio file.

2. drop an "invert" utility in one channel

3. disable WARP in one audio file

4.take care that the master tempo is equal to the tempo from the audio file that no time stretching occurs!

5. play back both files in session view, they should wipe out each other, but they don't. If you set it up in the arrange view there is no problem.

6. add 1 sample delay (the delay in lives mixer, click on "ms" to change it to "smp") to file with WARP off and you get 100% silence.

should look like this...

can you reproduce it?

mr.ergonomics · Post by **mr.ergonomics** » Wed Jan 07, 2009 3:32 am

c'mon that takes about 2 minutes to test it

or is something unclear?

Bunky Freaks · Post by **Bunky Freaks** » Wed Jan 07, 2009 1:16 pm

clips cancel perfectly here. which version of live are you using?

try to type in the "seq. BPM" value in the sample info manually. i found that sometimes live automatically puts in a value slighty different from the displayed one, like 120.00001, which then is displayed as 120. This obviously prevents the clips from canceling 100% ...

also remember that the complex warp mode changes the clip even if played back at the original pitch.

what about saveing and uploading your liveset for others to test this with your exact settings?

mr.ergonomics · Post by **mr.ergonomics** » Wed Jan 07, 2009 8:02 pm

Thanks for your answer bunky! I use live 7.0.10 but also tested it in 7.0.14

also remember that the complex warp mode changes the clip even if played back at the original pitch.

Yeah, I used repitch because of that. Also tested beat 1/16.

try to type in the "seq. BPM" value in the sample info manually. i found that sometimes live automatically puts in a value slighty different from the displayed one, like 120.00001, which then is displayed as 120. This obviously prevents the clips from canceling 100% ...

As far as I remember I did it that way.

Upload the liveset: good idea, I will upload it tomorrow!

mr.ergonomics · Post by **mr.ergonomics** » Sun Jan 11, 2009 7:03 pm

update!

I think it has do with "fade in" and is not related to " warp on" or off.
I played around a bit and couldn't reproduce this problem. After I load the "old" test project it was no problem to reproduce the problem again. I searched for a difference and the only one I found was "fade in". It seems so that "fade in" not Warp cause the problem.

I have done a test project. Maybe you could take a look?

* the first scene with fade in don't cancel 100% without 1 Smpl delay
* the second scene cancel 100%

Note: it's important to trigger the scene multiple times (just click the scene multiple times..), because the first time the problem don't occur.

http://www.zshare.net/download/539950984d0aaae6/

mr.ergonomics · Post by **mr.ergonomics** » Sun Jan 18, 2009 7:11 pm

c'mon. 180 views and no one wants to test it? should I provide better instructions?

I'm a bit disappointed, because if it's a bug it's a serious problem and therefore I don't get it why nobody is interested in.

Palmer Eldritch · Post by **Palmer Eldritch** » Sun Jan 18, 2009 8:03 pm

Hi mr.ergonomics,

I can reproduce the bug with your project.
In my tests one of the clip is one sample off only when one is warped and the other is unwarped and at the same time fade is on and at the same time when I start the just playing clips and at the the same time the starting of the clips is quantized.

Just curious, why this is so a serious problem for you?

cheers p.e.

Bunky Freaks · Post by **Bunky Freaks** » Sun Jan 18, 2009 10:28 pm

i can also reproduce this here.

seems like this issue is tied to "global quantize". if its set to "None", the clips always cancel perfectly.

the best is to write to ableton support directly and point them to this thread. i can not imagine a scenario where this might be an issue, but it definitly does not behave as it should.

cheers!

mr.ergonomics · Post by **mr.ergonomics** » Mon Jan 19, 2009 1:05 am

thanks for checking it!

1. it seems like the latency correction has a flaw, maybe there is another issue related, which I/we don't found.

2. I can think about an situation. For example: You play two clips with similar frequencies which adds perfectly. If you delay one of the files or sample you will get comb filtering.