SOUND ON SOUND

[Elephone]

Hello. I was on the subject of chopping up bits of audio in a DAW and wondered... have software developers ever considered using a visual method rather than (just) amplitude level to decide on gating (or splicing) audio...?

I mean, I can do it pretty much by eye because I recognise what drum attacks look like from start to end and where the signal returns back to approach 0dBFS.

And yet, if I want to transparently gate or splice audio without worrying about strange choices, I still have to do manually it seems to me. I can't trust a plugin not to chop it somewhere I don't want it to.

In an era of facial recognition, I'd expect someone could create something similar to provide more transparently/rounded gating and splice points.

Is there any software that can do this more intelligently now? I did ask a few years back, and the answers were really about there not being much demand for splicing software, as it's more for creaters of sample libraries than users.

Thanks

[hobbyist]

Hello. I was on the subject of chopping up bits of audio in a DAW and wondered... have software developers ever considered using a visual method rather than (just) amplitude level to decide on gating (or splicing) audio...?

I mean, I can do it pretty much by eye because I recognise what drum attacks look like from start to end and where the signal returns back to approach 0dBFS.

And yet, if I want to transparently gate or splice audio without worrying about strange choices, I still have to do manually it seems to me. I can't trust a plugin not to chop it somewhere I don't want it to.

In an era of facial recognition, I'd expect someone could create something similar to provide more transparently/rounded gating and splice points.

Is there any software that can do this more intelligently now? I did ask a few years back, and the answers were really about there not being much demand for splicing software, as it's more for creaters of sample libraries than users.

Thanks

I would think any good DAW would let you diy without problems and no plugin needed.

As to AI and automating it I have not seen such a program yet.

[Tim Gillett]

I think when using my old Wavelab mastering programme to edit out sections, it automatically makes the cuts at identical amplitude points to avoid audible clicks and pops. It can sometimes be frustrating as the closest identical points can be slightly further into the material than I'd hoped and it's not the edit I'd planned.

[Elephone]

I think when using my old Wavelab mastering programme to edit out sections, it automatically makes the cuts at identical amplitude points to avoid audible clicks and pops. It can sometimes be frustrating as the closest identical points can be slightly further into the material than I'd hoped and it's not the edit I'd planned.

Yes, there are splice points added within the envelope of (what are to us) obvious events (drum hits, etc). I think it would be quite easy for a program to 'look' at waveforms from a sideways viewpoint and register these envelopes as clearly part of a events. I know gates have 'hold' settings but this could easily extend beyond some of the shorter events.

[Elephone]

I would think any good DAW would let you diy without problems and no plugin needed.

Yes, taking up hours of time (doing what a 6-year old could probably do) and which could be put to better, more productive use.

[Kwackman]

have software developers ever considered using a visual method rather than (just) amplitude level to decide on gating (or splicing) audio...?

I mean, I can do it pretty much by eye because I recognise what drum attacks look like from start to end and where the signal returns back to approach 0dBFS.

And yet, if I want to transparently gate or splice audio without worrying about strange choices, I still have to do manually it seems to me. I can't trust a plugin not to chop it somewhere I don't want it to.

I can't grasp what you mean, which probably says more about me than you!
I don't see gating and splicing to be the same thing, but maybe that's not important here?

How would you expect the software to know where you wanted to cut?

Not sure if this is relevant but before DAWs, when editing was done on tape, deciding where to splice was done purely by ear (and razor blade!) with no visual clues.
But you still had to manually decide where to cut e.g.on a word, on a beat, on a breath, in room atmos. Software can still do the razor-blade bit, but I don't know how software could know where you wanted to cut?

[Elephone]

How would you expect the software to know where you wanted to cut?

If you take something like a drum 'event', it's clear that the envelope has a certain shape. Different audio events have distinct shapes that we can recognise by eye, i.e. where an event starts and ends by the envelope. But even when an envelope is as clear as this:

https://dt7v1i9vyp3mf.cloudfront.net/st ... _6E456.jpg

...software can still insert markers or splice points within the envelope because it's going by amplitudes (not seeing a bird's eye view). Obviously, it can get tricky, but they're not even doing a basic job without chopping attacks or decays or sometimes placing divisions right within an event.

[hobbyist]

I would think any good DAW would let you diy without problems and no plugin needed.

Yes, taking up hours of time (doing what a 6-year old could probably do) and which could be put to better, more productive use.

Depends how slow you work and how good that AI software would be whether you DIY or not.

It never takes me hours to do it.

Why dont you write your own software. If a six year old can do it then the software would be pretty simple to create.

[blinddrew]

If a six year old can do it then the software would be pretty simple to create.

Well there's a logical fallacy for you right there...

[Hugh Robjohns]

IDifferent audio events have distinct shapes that we can recognise by eye, i.e. where an event starts and ends by the envelope....

Yes... you are visualising the amplitude, and deciding on the threshold displayed graphically the point at which to edit... it's exactly the same as a gating system i you think about it.

...But even when an envelope is as clear as this ...software can still insert markers or splice points within the envelope because it's going by amplitudes...

If it's putting them in the 'wrong' places, then it simply doesn't have the correct threshold values to work with!

Most DAWs have a 'strip silence' function of some sort to isolate wanted audio above a set threshold into separate clips or events. And the good ones let you adjust the action threshold so that you can get it to make exactly the same decisions as you would visually. If you DAW won't let you adjust the threshold, you may need to investigate a better DAW!

And, of course, being non-destructive, you can always go in and tweak any which need it. I use that function a lot -- let the DAW do the donkey work and then I fine-tune and finesse...

H

[Kwackman]

Right, I think I've understood better, thanks!

I'm not a great user of hit points etc.
My DAW is Cubase.
It has some variables for picking hit points.
It has a Threshold, Intensity and a Minimum Length setting.
Threshold and Minimum Length are obvious, I had to check the manual for "Intensity"!
"(This) filters hit points by their intensity. Drag the slider to the right to discard less intense hit points"
I'm not sure I'm any wiser, TBH!

Anyway, maybe your DAW has settings that might help you achieve what you're after?
Good luck.

[Elephone]

let the DAW do the donkey work and then I fine-tune and finesse...

H

I've been using Recycle to divide drum events, which is old, but Propellerhead re-released it as it is, so I don't know what's better. It often puts splice points in unwanted areas and the solution is always to "just adjust them manually". I can do this entirely by eye and easily see where the audio begins to attack and finishes decaying. Manually adjusting these markers is fine for short loops, but entire drum stems (for re-purposing)...?

If it's putting them in the 'wrong' places, then it simply doesn't have the correct threshold values to work with! H

Could the problem be that, real (played) drum stems might not have a uniform threshold (i.e. that can be set ideally for the whole stem) but the eye can make better judgments, perhaps?

Thanks

[Hugh Robjohns]

Given the very crude graphical display, I doubt that... but it does all come down to how well you can adjust the thresholds and how intelligent the gating process is...

I don't do much like this with drums, but I used to do it a lot with speech, and the amount of manual tweaking was only ever a small proportion of the total number of edits if the threshold was optimised.

H

[hobbyist]

If a six year old can do it then the software would be pretty simple to create.

Well there's a logical fallacy for you right there...

If you think that then tell me what the fallacy is.

Any six year old can program these days:)

And if the approach is so simple a six year old can do it then even an older person should be able to program that easily.

[blinddrew]

It's a complete non-sequitur. Just because complex pattern recognition is easy for humans doesn't follow that it's easy to program a machine to do it.
B does not logically follow A because they are unconnected actions.

[Mike Stranks]

Any six year old can program these days:)

Not quite. Any six year old may be able to write contextually correct code, but I'd argue that there's a whole lot more to programming than that. You have to understand the application, determine logical action sequences and know where and when the system has to test parameter values and branch accordingly.

The most important part of that is understanding the whole application. Some of the stuff for which I designed systems and programs took me many weeks to understand, such was the complexity. I know some pretty bright under-10s, but...

[hobbyist]

Any six year old can program these days:)

Not quite. Any six year old may be able to write contextually correct code, but I'd argue that there's a whole lot more to programming than that. You have to understand the application, determine logical action sequences and know where and when the system has to test parameter values and branch accordingly.

The most important part of that is understanding the whole application. Some of the stuff for which I designed systems and programs took me many weeks to understand, such was the complexity. I know some pretty bright under-10s, but...

In theory yes. In practice it is still a program and if it works correctly then the frills of a salable program are fluff.

Oh, wait. Most programs sold are full of bugs and alpha versions that the users are debugging for the adult programmers writing them.

[Mike Stranks]

I'm done... life's too short...

:wave:

[Hugh Robjohns]

Don't worry Mike, James has escorted hobbiest from the building for expressing some unpleasant and inappropriate views in other threads.

[Tim Gillett]

On gating generally, I feel it's so easy to set gating thresholds wrongly by, as Hugh says, not fine tuning manually, or even using gating when it's unnecessary and harmful. I was just watching this short historical Wimbledon clip from British Pathe. Around 0:50 the audio gets interesting. No substitute to actually listening to the results of our work...

https://youtu.be/GMwcmWnB80I