Grid timing guitars

[Mr. Torture]

So I have been playing around with audio snap, I assume this is what you use to perfectly align audio to the grid. Some things sound a bit wacky and stuttering, I assume it because I am snapping the wrong transients. 
 

Maybe it’s because I am trying to snap distorted guitars? The transients rarely land right on the beat and I know my timing isn’t that bad lol.

How do you guys normally time guitars?

Thanks!

[Byron Dickens]

1 hour ago, Mr. Torture said:

How do you guys normally time guitars?

Play in time.

[CSistine]

3 hours ago, Mr. Torture said:

Maybe it’s because I am trying to snap distorted guitars? The transients rarely land right on the beat and I know my timing isn’t that bad lol.

How do you guys normally time guitars?

Except for instruments like piano, for most instruments the transients are very difficult to find. Very often there are several places where a note/hit could have its transient and depending which one you choose it sounds completely different. With guitars, winds and vocals it's especially challenging! Guitar transients can be at the pick/finger click, but also at the real tone start or somewhere in between. And this varies often in the same playing whether rhythm or lead, but it is usually a recognizable pattern.

That's why DAWs and other programs (like Melodyne) very often fail to find the transient "felt" by the musician! This can be proved, if you take the same audio clip and let several programs find the transients. You will be surprised how different the results will be!

That's the reason why I check through all the transients manually and most of them won't be placed at the DAWs propositions! Then there is also the swing feel of the musician that plays in. Usually I try to find out what was the intention, before I quantize the transients to this feel, i.e not to the exact grid (in Reaper this is easier, because you can define a swing into the grid, but in Sonar I do this manually).

By the way, because of the poor automatic transient detection I don't think it will be realistic with a lot of A.I. ideas in the near future! ?

[Mr. Torture]

7 hours ago, Byron Dickens said:

Play in time.

I came here to ask a legitimate question about a feature the software offers, just looking for some pointers and maybe some insight. I don’t know why you felt you needed to post this. Not only does it add no value to the thread, it’s disrespectful. Maybe you think you’re being funny or you are bitter. Don’t know, but please just avoid comments like this in the future. It’s old and not productive.

 

4 hours ago, CSistine said:

Except for instruments like piano, for most instruments the transients are very difficult to find. Very often there are several places where a note/hit could have its transient and depending which one you choose it sounds completely different. With guitars, winds and vocals it's especially challenging! Guitar transients can be at the pick/finger click, but also at the real tone start or somewhere in between. And this varies often in the same playing whether rhythm or lead, but it is usually a recognizable pattern.

That's why DAWs and other programs (like Melodyne) very often fail to find the transient "felt" by the musician! This can be proved, if you take the same audio clip and let several programs find the transients. You will be surprised how different the results will be!

That's the reason why I check through all the transients manually and most of them won't be placed at the DAWs propositions! Then there is also the swing feel of the musician that plays in. Usually I try to find out what was the intention, before I quantize the transients to this feel, i.e not to the exact grid (in Reaper this is easier, because you can define a swing into the grid, but in Sonar I do this manually).

By the way, because of the poor automatic transient detection I don't think it will be realistic with a lot of A.I. ideas in the near future! ?

Thank you for your response, it’s very helpful!

[petemus]

I used Audio Snap until I didn't. Got tired of the artifacts it tends to introduce. Maybe I just don't know hot to use it properly.

These days I tend to just split the audio (with short auto crossfades) and move/stretch/shrink where needed. It seems to retain the quality if you don't have to do much stretching/shrinking. Sure, it's a lot of work but you usually get good results and there's no need to work around tool peculiarities. When done, just bounce to clips to make it nice one piece of audio.

[CSistine]

22 minutes ago, petemus said:

It seems to retain the quality if you don't have to do much stretching/shrinking.

It's the same with AudioSnap, it uses stretching/shrinking, too. Thus, the result fairly depends on the algorithm (stretch method) that you have chosen for the track´s Offline Render! I prefer the different Radius algorithms (depending on the source), they are superb (only available in Sonar/CbB)!

28 minutes ago, petemus said:

These days I tend to just split the audio

Though I use AudioSnap a lot, in some cases I use the same method as you do. Especially if there would be too much stretching/shrinking involved. Sometimes if I have to shorten a long note a lot, then I don't use xfades, but I cut out the note in the middle and align the waveform with the biggest zoom, so that the waveform continues in a harmonic way. Works great!

[John Vere]

Byron’s answer is possible the closest to the truth. 
Distorted guitar is normally a huge sustained wave form and more or less the worst possible audio you could use with audio snap. 
There’s no Auto Fix for timing issues with that type of material. 
 

As said, if I hear an out of time part I highlight and drag. Once done I bounce to clip. 

[Byron Dickens]

4 hours ago, Mr. Torture said:

I came here to ask a legitimate question about a feature the software offers, just looking for some pointers and maybe some insight. I don’t know why you felt you needed to post this. Not only does it add no value to the thread, it’s disrespectful. Maybe you think you’re being funny or you are bitter. Don’t know, but please just avoid comments like this in the future. It’s old and not productive.

And I gave a legitimate answer. Play in time. That's not disrespectful. Man,  you would have had a rough time with my guitar teacher, that's for sure....

[Bristol_Jonesey]

You should also consider whether the timing needs correcting or not.

If it's not audible and just looks off, leave it alone.

[Byron Dickens]

^^^^^^^^^^

This!

Listen with your ears and not your eyes.

[David Baay]

In addition to CSistine's comments about how transient markers are located relative to the initial onset or peak of the transient, you need to be aware that by default Audiosnap is typically going to place a lot of superfluous/duplicate and 'false postive' markers. This can be a problem because when you drag a transient marker, the range over which stretching/compressing  occurs is only back to the previous marker and up to the next marker which may or may not be accurately on a beat or may be on the next 16th when what you actually want to do is stretch everything proportionally across a whole beat, several beats or a whole meaure. You can select several markers and do a  proportional stretch, but the better approach is to raise the Threshold slider in the Audiosnap palette until approximately the right number of transient markers remain enabled and then manually enable/disable any additional ones you need. or don't need to be active. Then, as you work through the part, adjust the position of any poorly placed marker by dragging  its 'head' (the diamond at the center of the marker) before dragging its 'tail' to stretch the audio.

All that said, my preferred approach  to improving the timing of any audio is to use Set Measure/Beat At Now to get the timeline aligned to the audio with tempo changes only where needed, enable Audiosnap Follow Project Tempo in Auto-stretch mode on the clip, and then reduce or flatten the variability in the resulting tempo map. If the project already has multiple audio tracks  you can do this only with the 'wonky' one copied to a temporary project to do the flattening, then bounce down the result and re-import it to the project with the fixed tempo. The advantage of this approach is that you can let the part run for several measures at a time where the timing is good, and only 'Set' beats (or fractions of a beat) where there's a problem that audibly needs to be addressed so you do an absolute minimum of audio-distorting stretching. Also, with Set Measure/Beat At Now you're not beholden to where Audiosnap places markers. You can put 'Now' anywhere your want and just use the markers as a guide or Tab to them when they look right.

Whatever technique you choose to use, you're going to have to work with the tools for a while to understand exactly how they work, what their limitations are, and how to make them bend to your will.  And, as has been suggested, sometimes re-recording is the best solution if it's your own performance.