Adjusting audio timing using technology - is the results acceptable?
 

Hey all!

Still a newbie here - pre-apology inserted here!

I have tried many different workflows (too many to recall all now) to try to utilize Elastic Audio to help me fine align the timing of some instrument tracks. The tracks are not horribly out of time or anything like that. But, I have to admit when all the instrument are fine timed, the overall mix does sound better - especially when target key ornaments where specific notes come together to synergize the moment.

My results is sometime different when using Elastic Audio (which makes it difficult to predict the outcome). When using on all acoustic instruments, in my case, I generally:

1) select consolidated audio clip in the track
2) turn on polyphonic
3) quantize (if possible) to a grid time or manually align one by one (which sucks when the analysis markers don’t auto locate the start of transients as expected (big misses on acoustic guitar)
4) fine align if quantize misses any or move the wrong direction on some transients
5) listen

I generally accept some “wobble” in the the audio playback here as the track is still in a real-time mode for polyphonic Elastic Audio (EA). I just listen for timing issues and adjust as necessary to get right

Then (as I have learned from some suggestions here), change the EA mode to X-Form to basically process the audio to a file or Render it so as to change from real-time to a file as normal. It’s at this point I believe I should expect to hear no “wobble” in the audio but, more times than not, I still do (not always on the same instrument or timbre of audio material- which is the mystery). Then if it does sound good, I commit the changes and turn off EA entirely. Sometimes I get “wobble” at this point also.

Sometime I don’t even go from polyphonic to X-Form at all .. just go directly to commit from polyphonic and commit timing changes then and the results can be better than going to X-Form first. But, not always, as at times the “wobble” from the real-time polyphonic alignment process doesn’t go away at all when committing.

Sometimes I swear I’m going crazy as one day I turn everything on and listen to the track and no “wobble”, next day I am working on another track or edit totally unrelated to the track that sounded great the day before and the “wobble” is there. I begin to wonder if the system resources / pro-tools environment settings are jerking with me. But, for mixing I have the max samples (2048) and a 2GB hard drive buffer cache setup as I’ve normally used for a while.

Having said all this (and I know that was a lot), has anyone else experienced the same or perhaps found workflows the work in a predictable manner and deliver quality results? If so, I am all ears... I have thrown a lot a time in flipping things on/off back and forth with audio quality that comes and goes.

I appreciate all feedback, comments or recommendations up to an including not using it, using something else (ex Logic FlexTime), etc.

As a footnote, I have also briefly tried melodyne’s time alignment feature but, that seems to take more time than EA manual alignment but, I have not spent a lot of time there to say I know 100% how to use it yet.

Thanks in advance!








Sent from my iPhone using Tapatalk

Re: Adjusting audio timing using technology - is the results acceptable?
 

I just quit using elastic audio a long time ago. The audio quality sucks.
So I just edit in the traditional way, but using Control + "+" or "-" to nudge the audio within the region forward or back according to the Nudge value. (command for PC)
This lets me edit way, WAY faster and get better sound results.

Just try it, separate the region you want to move (press B), then Control + "+" or "-", then apply crossfades to boundaries.

Re: Adjusting audio timing using technology - is the results acceptable?
 

So to be clear - cut/break, nudge and cross fade is the “Traditional way”? Not to doubt you just wanted to make sure this what I understand and maybe anyone else who reads this somewhere down the road.

Now... so did start out doing that and seem to the butcher method as depending on how good you make the cut, move and fade little artifacts in the audio tend to show up (when track is solo’d but, not when in full playback of mix ... is that normal using the traditional way? I am not saying leave pops or clicks type stuff just noticed there are some dips or maybe quick changes in the music. I guess that just comes with how much time you spend on cleanup after the cut and nudge?

Appreciate the info. Sounds like your results with EA has been like mine! Ever used Beat Detective to do this cut, nudge and fade process in some automated fashion? Did it help or just add more time in the end?



Sent from my iPhone using Tapatalk

Re: Adjusting audio timing using technology - is the results acceptable?
 

I'm sorry, by traditional I meant the old "move things around and then adjust crossfades". Not by time stretching.

But the method I described is not what most of the people is used to.

And indeed, you should be careful with crossfades.

Re: Adjusting audio timing using technology - is the results acceptable?
 

For me, X-Form is the only way to adjust timing(**). Both Monophonic and Polyphonic cause way too much garbled audio. X-Form is very clean, but since it renders every little change, the trick to using it is to separate audio into clips that fit in the EDIT window(this way, the render is only working on smaller clips, so it happens in seconds instead of minutes).

** I should also mention that Melodyne can be used to move stuff around in time. Just COMMIT it once you like the result:o

Re: Adjusting audio timing using technology - is the results acceptable?
 

So I am clear on your workflow there...
1) you breakup any fully consolidated / single / long audio clip into multiple chunks/parts
2) you turn on X-Form (you do not turn on any Polyphonic or Monophonic or any other algorithm method that X-Form first or at any other time?)
3) because you did #1 you can more quickly adjust the transients as desired without it taking forever processing a fully consolidated audio clip. I have experienced this many times and it makes you want to Zzzzzz - that's a great tip!
4) when completely done with time adjustments in #3 turn off X-Form & Commit
5) cleanup any breaks/ gaps / add fades and consolidate. (I am assuming here)



In my brief use of melodyne time align... it seems a bit hard to get the grid and auto "snap" to time slider to get what I expected. I will spend more time on this ... as I felt it should have worked better than the results I got.


Thanks for feedback...

Re: Adjusting audio timing using technology - is the results acceptable?
 

When you use EA for what it was designed it works just fine most of the time. Maybe the OP is using EA on multi mic sources, that doesn't work, for drums, etc the best tool is Beat Detective

Re: Adjusting audio timing using technology - is the results acceptable?
 

That's correct^^^ And, amagras is 100% correct that EA is not good on multiple mics such as a recorded drum kit or a guitar with 2(or more) mics. For drums, Beat Detective or manually cut/move/crossfade is best. If I need to EA an acoustic guitar that was recorded with 2 mics, I will pick the best sounding single mic and mute the other track:o

Last advice: its not a bad idea to DUPLICATE PLAYLIST first. That way, no matter how messed up your edits get, the original(untouched) audio is still available for another try:D

Re: Adjusting audio timing using technology - is the results acceptable?
 

That's the best practice, that way if EA messed up a passage you just have to copy it and paste from the original playlist, that's standard practice for autotuned vocals

Re: Adjusting audio timing using technology - is the results acceptable?
 

Albee that’s good info and I have to agree from all my trials and errors and hours spent sorta trying it all -that method definitely delivers the best quality results (granted the splices don’t bring about anything new to deal with)! I have not tried or have the reason to yet to deal with the multi sourced recording and EA... YET! Thanks for the tip ahead of time - this makes sense and will come in handy I’m sure.

Thankfully I have been duplicating the playlist as I go through the different phases of you will. However early on I didn’t and had to look back through all the playlists and find my original tracking days and pull back those takes for each track. More of a pain but, found them, lesson learned!

Thanks for all tips and suggestions... very helpful indeed!


Sent from my iPhone using Tapatalk