?Audio Snap?

[murat k.]

On 6/15/2021 at 2:56 PM, Light Grenade said:

Agreed, transient detection definitely needs improving.

Agreed. When I use the command "AudioSnap Go to next Transient Marker", sometimes it works without problem like this:

It is not perfect but it works. But sometimes it is not working as you see below:

[reginaldStjohn]

I am late to the party here but I have had mostly good success with the transient detection in Cakewalk but it depends on the material.  Those example pictures above seem like they would be very difficult to find the transient, matter of fact, I don't even think I would say those waveforms have a transient. What I think you want is where the sound starts which is not technically the same thing as the transient or attack of a sound.

[murat k.]

6 minutes ago, reginaldStjohn said:

I am late to the party here but I have had mostly good success with the transient detection in Cakewalk but it depends on the material.  Those example pictures above seem like they would be very difficult to find the transient, matter of fact, I don't even think I would say those waveforms have a transient. What I think you want is where the sound starts which is not technically the same thing as the transient or attack of a sound.

So, we are not calling them "Transients". "Sound starts" seems OK. 

With a proper algorithm, we can navigate in that kind of situation which we need.

[murat k.]

I've made a research, found that beginning of a sound is called Onset.

So, we need the proper Onset detection in the Cakewalk. ?

 

[David Baay]

On 6/12/2021 at 2:49 PM, Will_Kaydo said:

Cut and nudging the information in time doesn't always work that great, cause now we have to do fades in and dout which also adds to the problem sometimes - like ending the tail to quickly on short notes. 

That's precisely why Audiosnap uses stretching instead of 'moving' audio. Stretching lets you move the transient without creating overlaps. This is more important when you're working with audio that has pitched sounds sustaining across transients at a significant level. Drums, and rhythmic bass/guitar with short sustains are the only types of audio that really lend themselves well to the split-and-move approach.

Stretching audio (especially by variable amounts from one transient to the next) is an imperfect science, and is dependent on the quality of transient detection which is also an imperfect science. I hear tell that some other DAWs do this better in general, but have never really heard side-by-side examples of the same audio stretched in the same way by someone with expertise in both DAWs to be able to judge. But I do know that results from Cakewalk can be pretty wildly variable depending on the nature of the material and how much work you put into massaging the transient markers and choosing the best algorithm for the material.

[Will.]

14 minutes ago, David Baay said:

That's precisely why Audiosnap uses stretching instead of 'moving' audio. Stretching lets you move the transient without creating overlaps. This is more important when you're working with audio that has pitched sounds sustaining across transients at a significant level. Drums, and rhythmic bass/guitar with short sustains are the only types of audio that really lend themselves well to the split-and-move approach.

 

Don't know why you're repeating what I've said. 

[David Baay]

2 minutes ago, Will_Kaydo said:

Don't know why you're repeating what I've said. 

Because your original post said "So why is cakewalk stretching the note instead of moving it leaving you ending up with broken information. "

[Will.]

27 minutes ago, David Baay said:

Because your original post said "So why is cakewalk stretching the note instead of moving it leaving you ending up with broken information. "

Exactly.

[Mark Morgon-Shaw]

22 hours ago, David Baay said:

Cakewalk can be pretty wildly variable depending on the nature of the material and how much work you put into massaging the transient markers and choosing the best algorithm for the material.

Best algorithm is getting right at source.  2nd best is probably Melodyne.  3rd best is getting your hands dirty and chopping up the audio manually

I tried Audiosnap with 8.5 and a wayward drummer.  In the end it was easier to sack the drummer and re-track..never used it since. Or the drummer.