Torturing a Classical Guitar Track

[seanmccoy]

Hey, all.

Working on a project with a flute player who is doing a COVID-style duet with a classical guitarist, recording the parts separately, guitar first. It's an avante-garde piece that's completely unclickable, and though they've performed it together and the guitarist is a very fine player, not surprisingly the guitar part as sent to us has a number of places where the flutist wants pace, spacing and lengths changed. So I'm doing a ton of cutting and time stretching that should never be done to a classical guitar recording—especially one as beautifully recorded in a Paris studio using three pairs of mics as this one was.

My long-winded question is about whether I'm using the best method to accomplish this. I'm good at making believable musical cuts, often using checkerboarding, when spacing needs to be shortened. But the real problem is when trying to lengthen things, especially straight sustains. I'm just setting all the tracks to Polyphonic for quick editing and experimentation, then switching them all to X-Form at the end and waiting (and waiting and waiting and waiting) to see how much of the stretched warbling will go away. For the most part I've actually been impressed with how well it works, but the question remains: is there a better way to go about this?

I'm totally open to suggestions.

[albee1952]

My opinion only I think you ARE doing it in the best way, overall. The only change I would make is that I would stick with X-Form all the way thru. The biggest issue with multiple tracks of the same instrument is that sometimes(all to often), phasing will go sour. If that happens, it can be really hard to solve, and sometimes I will mute some of the extra tracks and simply not use them. Its up to you whether you need them or not, of if you can find other ways around phasing issues if they happen. Of course, you may get lucky

BTW, if you are groaning over using X-Form, its a big help if you cut the audio up into clips that fit across the screen and anchor(drop WARP markers) near each end of the clip being edited. The render happens much faster this way

[seanmccoy]

Originally Posted by albee1952

My opinion only I think you ARE doing it in the best way, overall. The only change I would make is that I would stick with X-Form all the way thru. The biggest issue with multiple tracks of the same instrument is that sometimes(all to often), phasing will go sour. If that happens, it can be really hard to solve, and sometimes I will mute some of the extra tracks and simply not use them. Its up to you whether you need them or not, of if you can find other ways around phasing issues if they happen. Of course, you may get lucky

BTW, if you are groaning over using X-Form, its a big help if you cut the audio up into clips that fit across the screen and anchor(drop WARP markers) near each end of the clip being edited. The render happens much faster this way

Thanks for the insight, Dave, and I'll follow your advice about dropping warp markers. The reason I'm not staying in X-Form all the time is that it's just so dang slow if we're working with longer clips. We're often nudging clips into place experimenting with nine or ten placements, and having to wait several minutes for each one just isn't tolerable.

An update, though: I broke down and upgraded to the full Elastique AAX from my TC/E only version, based on lots of online praise and Mike Thornton's review (and it was on sale), and have been very, very impressed with its ability to do unreasonable stretching of the guitar sustains cleanly. On the shorter ones X-Form is often dead on, but Elastique allows for more radical attempts. And it's lightning fast. So, another potential tool for the arsenal.

On a side note, the irony is that I was never impressed with Elastique at speeding up VO's, which I often have to do because none of my clients have ever figured out how to use a stopwatch. And after doing another fairly extensive comparison using the full AAX version, I still find Avid's built-in TC/E algorithm to be noticeably superior at this.

[m4tune]

The biggest issue with multiple tracks of the same instrument is that sometimes(all to often), phasing will go sour.

I'm using serato pitch'n'time for years now. As AS Plug-in which doesn't support time handles (what a bad joke) it's quality can be stunning. Of course very depending on the given signal!
But it keeps the phase relation between multi-miced tracks and the other advantage is that you can create ritardandi and accelerandi as well as complete free progressions.
Maybe a little overpriced for this job only?

[seanmccoy]

Originally Posted by m4tune

I'm using serato pitch'n'time for years now. As AS Plug-in which doesn't support time handles (what a bad joke) it's quality can be stunning. Of course very depending on the given signal!
But it keeps the phase relation between multi-miced tracks and the other advantage is that you can create ritardandi and accelerandi as well as complete free progressions.
Maybe a little overpriced for this job only?

Yeah, I've had my eye on that one for a long time, but have never been able to quite justify its high cost. Good to know it does seem to live up to its reputation, at least.

[c-tone]

In tough cases such as these where time-stretching is almost always audible to some extent, I usually end up not time stretching the exposed long sustained notes. Instead, I'll spend a lot of time experimenting with duplicating the sustained notes to other tracks, offsetting them, and doing a combo of crossfading and fader rides to make the notes longer without the artifacts.

I also sometimes use regenerating delays to extend the long notes, but in this case with so many stereo pairs, it probably wouldn't work so well, but it's worth a shot.

It will take a lot longer to get right than stretching, but it can be more transparent.

[seanmccoy]

Originally Posted by c-tone

In tough cases such as these where time-stretching is almost always audible to some extent, I usually end up not time stretching the exposed long sustained notes. Instead, I'll spend a lot of time experimenting with duplicating the sustained notes to other tracks, offsetting them, and doing a combo of crossfading and fader rides to make the notes longer without the artifacts.

I also sometimes use regenerating delays to extend the long notes, but in this case with so many stereo pairs, it probably wouldn't work so well, but it's worth a shot.

It will take a lot longer to get right than stretching, but it can be more transparent.

Also really tough with an always-decaying instrument, and the nature of the piece is that nothing repeats anywhere. My attempts at such things have always taken forever and rarely yielded usable results. My compliments on your diligence, though.