Procedure for Maintaining Phase-Coherence - Multi-Mic (Drum Kit) - Elastic Audio

[skylar_battles]

1. Enable elastic audio on the track group
2. Determine which track you will use to derive analysis from. Unlike Beat Detective, you must only use 1 track for this.
3. Now suspend groups with Command Shift G
4. Select all tracks in the group excluding the track you’ve determined to use as the main track to edit off of
5. Delete all analysis markers on those tracks
6. Command Shift G again to enable track groups
7. Then change to warp view to start working

From here you have 2 options:
1. Start editing the warp markers on the guide track manually. The other tracks will follow and no analysis markers will conflict
2. Quantize normal from event operations and all tracks will warp and follow the track with the analysis markers

In either case after you’ve completed steps 1-8, you can then edit off off other tracks besides the main track designated at the beginning of the process, and all tracks will follow with no conflicting warp markers.

[Wireline]

Depending on how steady the tempo MUST be, I print a click track to use as my guide.

[skylar_battles]

A guide in what sense? In the way I've explained I don't think printing a click track could be part of the process.

[Wireline]

Step 2 of your description in OP

[skylar_battles]

Analysis must be derived from a component of the performance so a click track cannot be used to derive analysis for this workflow.

[Wireline]

No such thing as must…it’s just what I do when tightening tracks at customer request/they are so bad they need drastic help.

Your way works, as do dozens of others…

[skylar_battles]

I'm just talking specifically about not causing issues with phase relationships when stretching. For example, if you just quantize the drums without clearing analysis markers from all tracks except one, there will inevitably be transients in the performance where analysis markers are located ever so slightly away from each other on the different drum tracks. Because of this when you quantize, the drum tracks will not move perfectly in relationship to one another like when using Beat Detective so the drums will start to phase. I'm sure your approach makes sense as well, I just honestly think we may be talking about different parts of the process or using the term "guide" differently. So please do to explain if you don't mind doing so as after extensive experimentation, I thought clearing all other analysis markers for all tracks in the group except one was then only way to maintain phase relationships mult-mic elastic audio :)

[audiogamble]

Hey. I've tried this. It does quantize everything, but I can definitely hear comb filtering in the multitrack group. I'm using elastique pro, which is supposedly phase coherent across channels. Any tips?

[skylar_battles]

If all the analysis markers are deleted on all tracks besides the one you've determined to use before quantizing, the audio should be treated correctly for phase accuracy. Did you Command Control Click Elastiqué Pro to set it to rendered processing instead of real-time? It should sound better but not in any way that changes phase accuracy as far as i know. X-form or Polyphonic might be better also. Whichever algorithm you choose just don't forget to set to rendered processing before turning off elastic audio and printing your work. Some users have had issues with ElastiquéPro on short passages as is documented by Zplane I believe as well.

[audiogamble]

Yes. Did exactly the procedure you said. This was for an electric guitar performance. I recorded a DI, and 3 mics.

I removed all analysis markers and then only re-added the markers for the DI track. I then, with groups active, quantized the elastic audio events. Just as you said, this used the transient analysis from the DI track to also quantize and warp the other tracks in the edit group.

It did set nice accurate warp markers and quantize them together in a group, but the resulting sound was definitely sounding comb filtered on all algorithms, real time or rendered. A quick a/b between a playlist I kept intact without warping showed that the problem was introduced after the quantizing.

Did you do anything differently from what I'm describing?

I was hoping it wouldn't be the case, since they advertise this algorithm as being phase coherent between tracks.