BUG 2024.3 - Avid Time Shift TCE Algorithm Doesn't Snap to Grid, Unless...

[skylar_battles]

Found a major bug in 2024.3 that will affect many users especially since the default algorithm for the TC/E tool is Avid Time Shift. And yes I did try deleting preferences, but no luck :)

When I try to use the TC/E trim tool with the default Avid Time Shift algorithm and Grid mode enabled, the resulting rendered file is always 1 sample too long at 48 kHz (0.02083333ms). This is a massive issue when tryin to use that method to stretch loops, as if the user uses, Duplicate, Repeat, Duplicate and Extend Edit, or Clip Looping features, the loop will not land properly to the downbeat.

Oddly, if I enable elastic audio on the track and choose any algorithm at all, the issue does not occur, with Avid Time Shift still selected in preferences. I thought the only factor of which algorithm is used when using TC/E was what was selected in preferences?

Also, I tested every other algorithm I have access to and they were all sample accurate. Only Avid Time Shift has this problem. And samples vs. ticks makes no difference as expected.

Working algorithms I have tested:

Waves Sound Shifter
Avid X-Form
Avid TC/E
źplane ÉlastiqueAAX

I don't have access to either Serato's options or Soundtoys' and Synchro Arts' so I'm not able to test those for sample accurate rendering.

One more thing I should mention is that I'm having some issues editing at the sample level. I've noticed this before. When I try to use B or Command E to seperate the clip that is going 1 sample over the bar line, Pro Tools trims the end (S key) instead of separating. Anyone else experiencing that as well?

[skylar_battles]

Can someone attempt to reproduce this bug? Pretty easy to test. Thanks!

[original_min]

Yep, same here.

Tested on PT Intro 2023.12.1, Intel Mac Mini, Catalina.
Waves Sound Shifter
Avid Time Shift
Avid TC/E

Avid Time Shift adds a sample

[Digi_User.]

Hey,
Just to add I was also able to recreate this bug on a test machine running Mac OS X 10.15.7 with Pro Tools 2020.11.0 with both the Avid Time Shift and Avid TC/E plug-ins.

[MDGenerato]

Originally Posted by skylar_battles

Found a major bug in 2024.3 that will affect many users especially since the default algorithm for the TC/E tool is Avid Time Shift. And yes I did try deleting preferences, but no luck :)

When I try to use the TC/E trim tool with the default Avid Time Shift algorithm and Grid mode enabled, the resulting rendered file is always 1 sample too long at 48 kHz (20.833333333333332 ms). This is a massive issue when tryin to use that method to stretch loops, as if the user uses, Duplicate, Repeat, Duplicate and Extend Edit, or Clip Looping features, the loop will not land properly to the downbeat.

Oddly, if I enable elastic audio on the track and choose any algorithm at all, the issue does not occur, with Avid Time Shift still selected in preferences. I thought the only factor of which algorithm is used when using TC/E was what was selected in preferences?

Also, I tested every other algorithm I have access to and they were all sample accurate. Only Avid Time Shift has this problem. And samples vs. ticks makes no difference as expected.

Working algorithms I have tested:

Waves Sound Shifter
Avid X-Form
Avid TC/E
źplane ÉlastiqueAAX

I don't have access to either Serato's options or Soundtoys' and Synchro Arts' so I'm not able to test those for sample accurate rendering.

One more thing I should mention is that I'm having some issues editing at the sample level. I've noticed this before. When I try to use B or Command E to seperate the clip that is going 1 sample over the bar line, Pro Tools trims the end (S key) instead of separating. Anyone else experiencing that as well?

Hello! Already gone ahead and logged a case for this: 04778784

Thanks!

[Studio54]

At 48k, one sample is 0.02083333ms, not 20.83333ms as stated. Still an issue, but a much smaller one..!

[skylar_battles]

Studio54, thanks for the correction! Can you post the calculations used for determining? I used an online sample calculator for this purpose and I should've realized the result was incorrect. I'd say it's a bigger problem though because it's such a small error that it will be missed until your loops are evidently out of time later in the session.

Digi_User., could you test Avid TC/E algorithm once more? I'm not having the issue with that one. Many thanks!

Sticking with Waves Sound Shifter in the meantime...

[Studio54]

Sure!

It’s 1000 (ms) divided by your sample rate, so at 44.1k:

1000 / 44100 = 0.02267574 ms per sample.

And each millisecond of audio consists of 44.1 samples.

At 48k:

1000 / 48000 = 0.02083333 ms per sample.

Each millisecond consists of 48 samples (0.02083333 x 48).

[skylar_battles]

Thanks for the formulas! Any chance you could re-test the Avid TC/E algorithm, since I can't recreate the issue with that one? Thanks!