Bitwig

jonljacobi · Post by **jonljacobi** » Tue Feb 26, 2019 8:13 pm

Trying to import drum hits in a single file that are well-spaced to preserve the tails, and no onsets appear. I put more rapid hits at the front of one file, and those were detected, but the well-spaced ones were not. I'll refrain from comment and just ask if anyone else has had this issue? It would appear their detection algorithm gives up too easily.

And yes, they were imported as stretch, stretch was on, and I was in the onset view. I sent screen caps to support. Ableton picked them up just fine.

~Jon

bolba · Post by **bolba** » Tue Feb 26, 2019 8:21 pm

Maybe it is worth uploading these samples and contacting support directly?

jonljacobi · Post by **jonljacobi** » Wed Feb 27, 2019 12:37 am

I sent screens and a description to support before I posted. But there's always a chance I'm doing something wrong, so I post here as well. If they request the samples, I'll provide them.

antic604 · Post by **antic604** » Wed Feb 27, 2019 7:51 am

Have you tried different stretch modes? I've no idea if that impacts onset placement, but I guess it's worth trying.

jonljacobi · Post by **jonljacobi** » Wed Feb 27, 2019 11:53 pm

I actually didn't want to stretch anything, merely split it. But I tried all the modes. Support got back to me and they are able to replicate issue, so I'll expect a fix one of these first days. I'm not the first one to notice this.

M-Prod · Post by **M-Prod** » Mon May 24, 2021 7:56 pm

I still often get no transient detection at all. Even with a simple loop with single well spaced hihats on clipping volume. No transients detected whatsoever. Anyone else still have issues with this?

MuzikFreq · Post by **MuzikFreq** » Mon May 24, 2021 9:20 pm

I haven't gotten transient detection at all even on super simple loops.
I just gave up on it.

antic604 · Post by **antic604** » Tue May 25, 2021 8:33 am

MuzikFreq wrote: Mon May 24, 2021 9:20 pm I haven't gotten transient detection at all even on super simple loops.
I just gave up on it.

Weird, that never happened to me. Can you try narrowing it down to file type, bit depth and audio rate? Or maybe try different settings for how Bitwig is treating long & short samples (I think I'm using Bitwig's native Stretch Pro for both long & short samples)? And contact support with your findings.

cfernn · Post by **cfernn** » Mon Feb 06, 2023 1:06 am

Sooooo anyone else completely baffled by the awful onset detection after so many years? Sometimes it works, but most of the time it looks like my attachment. I can't...detect...any rhyme or reason why it behaves this way.

My audio is standard 24 bit 48khz, tried both Raw and Stretch modes, tried boosting and cutting the volume of the clip. No luck.

onsets.jpg

liquidsound · Post by **liquidsound** » Mon Feb 06, 2023 1:09 am

cfernn wrote: Mon Feb 06, 2023 1:06 am Sooooo anyone else completely baffled by the awful onset detection after so many years? Sometimes it works, but most of the time it looks like my attachment. I can't...detect...any rhyme or reason why it behaves this way.

My audio is standard 24 bit 48khz, tried both Raw and Stretch modes, tried boosting and cutting the volume of the clip. No luck.
onsets.jpg

I guess when they left Ableton they forgot to ask to the Transient technician how to do it right

Dionysos · Post by **Dionysos** » Mon Feb 06, 2023 7:50 am

cfernn wrote: Mon Feb 06, 2023 1:06 am Sooooo anyone else completely baffled by the awful onset detection after so many years? Sometimes it works, but most of the time it looks like my attachment. I can't...detect...any rhyme or reason why it behaves this way.

My audio is standard 24 bit 48khz, tried both Raw and Stretch modes, tried boosting and cutting the volume of the clip. No luck.
onsets.jpg

Might be worth sending them this project/audio file, as it looks like a good test case. support@bitwig.com

Trancit · Post by **Trancit** » Mon Feb 06, 2023 9:35 am

It´s not just many missing onsets... often they are so far off that one has to ask what this algorhtythm is actually trying to detect...

Onsets.jpg

I mean this wouldn´t really be hard to detect... this is a very classical situation...
a zero part before and an immediate jump of the amplitude nearly (if not exactly) on the grid...

Why does this algo think it would be a good idea to turn over all proved methods and put this thing completely off the grid into the flesh of the audio...
I don´t get it...

EDIT: this is how Ableton detects the same loop:

Ableton.jpg

.
Bitwig has by far the worse onset detection of the whole industry... I never had so many issues in any other product...
9 out of 10 times the onsets are completely off in any loop... perfect example of how it shouldn´t work...

If they would at least allow reading of in the audio file included markers as onsets if they can´t manage this themselves... but nope...

For a loop based DAW it´s actually ridiculous!

liquidsound · Post by **liquidsound** » Mon Feb 06, 2023 9:47 am

Pathetic is a better definition.
I’m beginning to think that they got stuck.
Maybe they built a structure from the beginning that slowly went on another path…
perhaps making it difficult to address core functions as this one, individual groove pool, automation issues reported in another thread, zooming stuff and the list goes on.

They started nicely…,

JHernandez · Post by **JHernandez** » Mon Feb 06, 2023 6:10 pm

Yeah it's pretty crap from the times I tried just to try. Luckily I don't do loops so it's someone else's headache.

PhilipVasta · Post by **PhilipVasta** » Mon Feb 06, 2023 6:36 pm

I'm also perplexed by how bad transient detection is in Bitwig. Related is the omission of audio quantize. Maybe the dev team is waiting on something else to make improvements, but I would hope that this is a top priority.

Bitwig

KVR Audio

Onset/transient detection issues still?