Looking for a good workaround for freezing groups

[diegoferrera]

Hello!

I have been trying to reduce a lot of the CPU workload in my sessions, so naturally I have been freezing/flattening a lot of my tracks. I do use many third-party plugins (they're what I'm used to and I do love them), and naturally they can eat up a lot of CPU. I do also use group processing, especially for multi-mic'd sources like drum overheads, guitar cabs, etc.

I've been trying to figure out the best way to go about freezing/flattening/printing a group when I'm ready to commit to them within my session. I've tried resampling the group into a new audio track, but when I phase-check the print with the original group they don't perfectly cancel out. I've even tried just importing the grouped tracks into a new session without any of the processing at all and it still doesn't phase-cancel, there's a little of top end I'm hearing after the phase flip in the print. I've been told to try to just stay away from group processing in general or third-party plugins for this reason, but I would love to at least try and find a way to make it work. Would love to know any of your thoughts!

[maximejerryfraisse]

I've even tried just importing the grouped tracks into a new session without any of the processing at all and it still doesn't phase-cancel, there's a little of top end I'm hearing after the phase flip in the print.

Do you mean no processing at all, even on the individual tracks? Are you using warping or pitch transposition? Are you recording in 32-bit? What’s the audio sample rate of your project and your mixed audio files?

Also, I’m curious why you’re looking for a perfect match between the print and the initial group track. Some processing algorithms use randomisation, and in such cases, if the two files don’t match perfectly, it doesn’t necessarily mean a loss of audio quality.

Personally, I’d recommend using Export Audio to freeze a group. It’s faster than real-time recording.

[cotdagoo]

I've tried resampling the group into a new audio track, but when I phase-check the print with the original group they don't perfectly cancel out. I've even tried just importing the grouped tracks into a new session without any of the processing at all and it still doesn't phase-cancel, there's a little of top end I'm hearing after the phase flip in the print.

From what I understand Resampling uses the master output for it's audio - it's possible that you have some plugins or filters on the master channel that's causing the resampled track to not cancel out perfectly. I know some of my master plugins do a bit of top end polish or boost certain things.

Group Freezing is currently being beta tested with an options.txt entry to enable it in the latest Live 12 betas on Centercode - http://ableton.centrecode.com

[diegoferrera]

I've even tried just importing the grouped tracks into a new session without any of the processing at all and it still doesn't phase-cancel, there's a little of top end I'm hearing after the phase flip in the print.

Do you mean no processing at all, even on the individual tracks? Are you using warping or pitch transposition? Are you recording in 32-bit? What’s the audio sample rate of your project and your mixed audio files?

Also, I’m curious why you’re looking for a perfect match between the print and the initial group track. Some processing algorithms use randomisation, and in such cases, if the two files don’t match perfectly, it doesn’t necessarily mean a loss of audio quality.

Personally, I’d recommend using Export Audio to freeze a group. It’s faster than real-time recording.

Correct, no processing at all on the individual tracks or master. To test, I just imported my overhead mic group from my session into a blank session with nothing on it. I tried to resample the two mono overhead mic tracks in a group into a single stereo track, flipped the phase of the print and am still getting a lot of top end from the drums coming through. No warp happening at all, it's disengaged in the original clips as well as the print. I record in 32-bit 24khz.


And I'm looking for a perfect match solely from a "maintaining the integrity of the original recorded audio" perspective. These are well recorded tracks and I was just curious of if I was resampling groups would it alter the sound of the files in any way. I was unaware of the randomization you're talking about, now I'm definitely curious!

[diegoferrera]

I've tried resampling the group into a new audio track, but when I phase-check the print with the original group they don't perfectly cancel out. I've even tried just importing the grouped tracks into a new session without any of the processing at all and it still doesn't phase-cancel, there's a little of top end I'm hearing after the phase flip in the print.

From what I understand Resampling uses the master output for it's audio - it's possible that you have some plugins or filters on the master channel that's causing the resampled track to not cancel out perfectly. I know some of my master plugins do a bit of top end polish or boost certain things.

Group Freezing is currently being beta tested with an options.txt entry to enable it in the latest Live 12 betas on Centercode - http://ableton.centrecode.com

I tested with a totally blank session with no plug-ins at all, still encountered the same problem. I will definitely have to check out the beta testing, thank you so much!

[maximejerryfraisse]

I record in 32-bit 24khz.

I assume you meant 44.1 kHz? Can you confirm that your original recordings are also 44.1 kHz?

[diegoferrera]

I actually meant 32-bit 48khz, my mistake

[maximejerryfraisse]

And just to avoid any confusion, can you confirm the Bit Depth is set to 32 in the Record menu?

I actually meant 32-bit 48khz, my mistake

Do the original files are also in 48kHz?

[diegoferrera]

And just to avoid any confusion, can you confirm the Bit Depth is set to 32 in the Record menu?

I actually meant 32-bit 48khz, my mistake

Does the original files are also in 48kHz?

Yes they do

[maximejerryfraisse]

I just did some tests. Perfect cancellation when Bit Depth is set to 32 in the Settings > Record menu. Obviously, when Bit Depth is set to 24, a lot of residual audio remains…

[diegoferrera]

I just did some tests. Perfect cancellation when Bit Depth is set to 32 in the Settings > Record menu. Obviously, when Bit Depth is set to 24, a lot of top end remains…

Interesting, okay I'm going to go back in and see what I missed