Elastic Audio caveats. artifacts.

[sweetness]

I've got a cool million vocal takes. the only playlist i like has miserable timing. It's fairly breathy intimate slow song vocals. I want to mess with the timing in elastic audio. but I am pretty weary of getting artifacts on an up-front vocal. I usually save it for vocaligning harmonies. or tightening up rhythm parts.

The question is: who uses it on the central lead vocal? and what pit falls do i need to look out for? I remember all the digital glitching that came with stretching tempos in Recycle ten years ago, and I don't want that to end up on an attempted sexy time vocal.

sincerely,
confused in clinton park

[ronwasserman]

I've found Melodyne to work better than EA in certain situations like this. You could go get the demo plugin and give it a shot.

[midnightrambler]

+1 for Melodyne, the user interface is way easier to use than EA too.

[sweetness]

Are you plug in or stand-alone?

any tips for both pitching, and playing with timing?

i thought the interface is rather cumbersome for timing. easier with AE when it's laid out with the grid showing and tracks above and below to match to.

Melodyne bothers me, i just do pitch, but i hear massive artifacts on the breaths and odd vocal noises.

don't want to have that on the main vocal at all.

[midnightrambler]

I do a lot of tuning and timing with vocals (I prefer stand alone because you can set up keyboard shortcuts).

I can only tell you what I do:

- Treat tuning and timing as 2 distinct passes, ie try not to do both at once. I prefer tuning first

- Cycle 2 or 4 bars at a time

- Don't tune breaths

- Don't tune 's' or 't's, they tend to sound awful tuned. Chop them out manually

- Try not to tune the beginnings of notes

- Take your time

- Don't over-tune, a singer being slightly sharp (for instance) can sound exciting, depending on the context

- Don't tune small words, passing notes, etc, unless you really have to

- Try not to zoom in too much. Tuning and timing every single note will end up sounding horrible

- Remember that changing the timing of one note will affect what's come before or after it. You may need to manually chop if you want to leave the previous note where it is, for instance

- Seeing where the blobs land on the grid can be a good indicator of timing but it's not foolproof. Definitely need to trust ears rather than eyes. (haha! Stating the obvious I know)

- Sometimes blobs need to be chopped twice and each bit subtly tuned differently, rather than relying on Melodyne to tune it 100%.

- Always compare the tuned and timed vocal to the original (I have a key command set up to play the original untouched vocal)

- If timing is difficult to get, mute all blobs and sing along where you think they should be. Unmute a blob at a time, and get it to where you feel it should, and continue until all blobs are playing

- Use the "smart" tool for efficiency - works a bit like ProTools in that the tool automatically changes depending on where the cursor is in relation to the blob

Um, that's it off the top of my head. Hope some of this helps

[sweetness]

Excellent reply, midnightrambler..

Regarding breaths and T's: even when I leave them alone, there will be a subtle (and sometimes not-so-subtle!) change. Melodyne is doing some processing to smoothly transition between it's pitch changes and the original unprocessed breaths.

This is a huge red flag for me, what artifacts are there that I am missing??

Try not to tune the beginnings of notes: Do you chop the slur at the beginning and leave that alone, and then just tune the held portion of the note?

I would like to hear more about how you are timing. And also more about your keyboard shortcuts. Timing is incredibly difficult for me. I end up using the take with the best timing, and usually awful pitch, just because i can't get timing quite right with melodyne, snapping is bad, and manually nudging is never enough, Elastic audio with a good vocal reference track above it gives me much more precise results that I can zoom in on....

[The Missile Silo]

I found while using EA especially in monophonic mode that it was creating weird anomalies, so I would switch between poly and rhythmic until I found it improved. EA is certainly not a perfect science. For my last record I was 50/50 with the Melodyne Plug-In, but now that I have the Editor version (which I have yet to fully test) I'll still be between the two.

When I am using PT's EA feature, I just make sure that I solo the heck out of each track that I use it on, and as much as possible. As mentioned, I also do audition between the various other EA modes (Mono, Poly, Rhythmic), until I find the one with the best result. For some reason, with certain audio, some EA modes just produce different results (better or worse).

This has just been my experience with it, hope some of it might help you.

Ron

[therecordinghouse]

have you guys tried rendering your final EA tracks with X form? it sounds much better than the base algorithms.

[sweetness]

I am probably behind the curve, I don't use EA X-form. Do you just edit in monophonic and then dupe the playlist before switching over to X-form?
Thanks!

I'm wondering if the others who expressed reservations about EA have used x-form...

[midnightrambler]

You should switch the algorithm to X-Form just before committing audio for best quality. Apart from drums, where Rhythmic can sound better. (Though this is also contentious.... there's been loads of debate about this on here already)

EA is great for timing. I just prefer to do both tuning and timing of vocals in Melodyne. YMMV, as the yanks say.