Vocal Comping - Need clarification about Operation Guide

I’ve completed vocal tracking and selected all my favorites takes using the Comp Tool. (I’m only dealing with ONE track version). However, in some spots, the “cut” isn’t clean. The singer didn’t breathe the same everywhere so I would like to smoothen out the cuts, and then assemble this into ONE clean part. (not sure if “part”, “clip”, or “event” is the right word to use).

On Page 150 of the Operation Manual, it says:

AFTER COMPLETING THIS TASK
After assembling your perfect take, you can improve your take.

•To automatically resolve overlaps and remove empty lanes, right-click the track and select Clean Up Lanes.

For audio, proceed as follows:

•Apply auto fades and crossfades to the comped takes.
•To put all takes on a single lane, and remove all takes in the background, select all takes and select Audio > Advanced > Delete Overlaps.
•To create a new and continuous event of all selected takes, select Audio > Bounce Selection.

So does this mean that “Clean Up Lanes” should be creating crossfades on all my overlaps in one fell swoop? Doesn’t seem like it and quite frankly not sure how it would do that without my input. Sounds too good to be true LOL… Which makes me wonder then what does “automatically resolve overlaps” mean??

And more importantly, is this in reality a 4-step process that I should follow to assemble a perfect take? (starting with comping, all the way to the last step of doing a Bounce Selection)?

Help anyone?

Hi,

The Crossfade is applied between the comped lines, but the time of the Crossfade is very short. This Crossfade takes care about a peaks, which could happen while cutting. If you want to take care about the breathe, you need longer crossfades, and you have to do this manually.

thank you Martin - can you also please clarify what is the difference between “Delete Overlaps” and “Close Gaps” functions? It is not clear to me exactly where the differences are or when you might use one instead of the other…

In fact, these 2 functions are exactly the opposite.

If you have some overlaps in the signal, you can delete it by the “Delete Overlaps”. Overlaps don’t make sense. Only one event/part is played anyway.

At the other, if you have some gaps (silent places), you can extend the event, to close these gaps.

Both of them you can use while comping, but in fact, you shouldn’t run in to it, if you use the Comp tool properly. Or you can run in to it, when you Quantize an Audio.

Hmm, I don’t think I am using the Comp tool properly. I mean, I know how it works to drag / cut the pieces I want, but I end up with a ton of tiny little bits that become audible when I play back my perfect take and I have to zoom in to find these tiny little pieces.

Also I don’t understand why having gaps would be an issue? All it would create is pure silence - is that not a good thing? Or is the goal to have one single event/part that starts at 0:00 origin and lasts the exact duration of the song?

This workflow is new to me so if you could explain a bit more or provide a link to a good resource or tutorial that would be awesome. I find most tutorial videos I’ve watched just show the cutting action in the lanes but nothing beyond like what the completed take should look like or if we should apply any detect silence after comping…

see screenshot for what is driving me crazy - this doesn’t seem to happen in the tutorials i watch but happens plenty in my own workflow.
shit like this is driving me crazy.JPG

Hi - sorry if I’m off on this one, but isn’t there a command “Clean up lanes” or something similar that takes care of that stuff?

Hi,

The “Clean up lanes” function just delete completely empty lanes. It doesn’t take care about the gaps.

I’d turn on the AutoFade and AutoCrossFade settings (right click on a track panel on the left), then you don’t have to worry so much about crossfading manually. This will introduce a (default I think) automatic 50ms fade or crossfade for each part which helps to get rid of any annoying clicks. However, I’ve found that manual crossfading is still sometimes required between comped takes.

As for the comping, its usual to have to go through fine tuning the edges of each event and moving breaths around to ensure it flows correctly. So it can be a time consuming manual process. I often do it with headphones and the track soloed so I can hear the details.

My process is to highlight the parts I like and mute the ones I don’t. Then I adjust the boundaries of the parts to get each edge smooth (e.g. fade-up over the breath, fade down quickly at the end, if auto-fades isn’t doing the job) and to make any transitions smooth (usually auto crossafade works well here). Sometimes I use different breaths from other places if the comp doesn’t work very well, really whatever gets the best result. Then I use the Track Versions to create a copy, which I bounce down. Then I tune this version (Melodyne or VariAudio) usually in sections (otherwise it can impinge on the gui performance). I’ll then get on with the other instruments, but when I’m finally happy with the vocal in the mix I’ll do another track version copy and bounce down the tuning and the whole lot into one single wav file (for future compatibility).

Mike.

Nice use of track versions, much less cluttered and fewer clicks than copying/making real copy/bouncing for each stage. Thanks for the hint!

Thanks for your input Mike… I looked into setting up those automatic fades/crossfades settings which were not enabled by default. However, if these are enabled, when exactly are they being applied “automatically”? I’m guessing this means we don’t need to use the X key anymore so I’m just curious about when that auto fading/crossfading operation would be triggered.

Also, have you come across the issue I depicted in my screenshot where there are a bunch of minuscule leftovers as a result of using the Comp tool?

Hi,

The auto-fades and xfades are calculated in real-time as the song is played. The fades are applied at the start or end of an event when they don’t overlap any other parts. The crossfade is used whenever two audio events overlap. I find that this does the job most of the time, but if there’s an odd sound or a blip with the auto-crossfade then I first move the boundary to see if that helps and if not then I use the X key and adjust completely manually. If you apply a fade or a cross-fade then this over-rides the auto version.

I have noticed that there are very small events left sometimes but I’ve never really known why and it hasn’t happened very much at all. But actually I don’t use the Comp tool much, I use the usual scissors because I always find that I can’t cut all lanes at the same place because each take is just different.

Mike.

Thanks, when you say applied real-time as the song is played… do you mean that its not actually “rendered”, that it’s just like a filter or effect that is being passed thru during playback but not actually part of the file? If I export the song, will these be included?

PS. Yeah, maybe I’ll go back to using the scissors too…

Indeed, like a plugin say, the auto fades/xfades are not imprinted to the files unless/until you bounce or export to create new files.

Mike.

I think the comping tool would work a LOT better if there were an option to comp our takes a a clean unused lane that does not automatically let the unused take “peak” our from under the used takes. Have all the other takes “stacked” like that is were the confusion comes in for me.

Also, I have key commands set up for “nudge region by +1 frame” that when applied to audio in a comp stack causes unuaul behavior - rather then nudging the region, it “shuffles” the selected region below another take. Any advice on this?