How to automatically balance L-R levels

[Joelistics]

For batch processing I'd use Soundgrinder Pro. Works really well and supports VST. Don't know of any plugin that does this.. Never had to... Mono not an option I suppose?

[AlexDavis]

It's an option! Just not a good one. Appreciate your input though, I'll look into the batch processing tool (although I love Audition's).

Anybody else know of anything that can do this!?

[sat159p1]

Maybe a normalizer which operates in unlinked mode?
Maybe the comp/limiter in unlinked mode, too?
Maybe the M/S processor with an option to lower the sides volume?

Any of these will probably result quite unnatural sound, but you propably know it already.
The good batch processor for that is in DSP Quattro.

[wado1942]

You'll be hard pressed to find an automatic option that will do a good job.

[AlexDavis]

Yeah, that's what I'm thinking. Thanks guys for the suggestions! If there was a way to split to mono - and have the L channel look to the R channel as a reference for volume levels... THAT would be the bomb. I'm sure there could be a script written for that, but certainly not something I'm capable of doing.

Alex

[huejahfink]

Originally Posted by wado1942

You'll be hard pressed to find an automatic option that will do a good job.

Yes... even if you find/write something that does automatically match the left and right via weighted RMS or something similar, it's highly likely that a pure maths approach is not going to give you the results you want, and that you will find other issues with the sound such as background information now shifting around uncomfortably etc.

I'd love to be proven wrong, but I from my perspective I'd say that your options are probably limited to

a) do it all by ear
b) make it all mono in a batch processor
c) select one setting to batch the files which narrows the files a bit, compromising between mono and full stereo.
d) learn to live with it.

From what you have been saying I would suggest either c) or d), with b) as a last resort.

[Laurend]

I don't know when a mono conversion can be preferable to a slightly different stereo image from file to file

[AlexDavis]

Hue- I ALMOST completely agree with you. Sadly. The only thing that is keeping me searching right now is foolish hope - and if I prove you wrong I'll let you know!

Laurend - quite simple - A mono set of files is preferable to slightly different stereo files in a sample library... One where absolute consistency is crucial for an effective instrument!

-Alex

[Andrew Kinsey]

I guess you could also try using a M/S processor to narrow the stereo signal enough to find a happy medium.

[Laurend]

OK,
SoundForge offers a batch capability, and it features RMS normalisation which can be a solution if you process a stereo file as two mono files assuming the phase isn't involved in your stereo issue.

[wado1942]

I don't know what kind of mic configurations you're using either, but intensity is only part of stereo placement. If you're using a stereo pair of mics, there's a greater difference in tone and also a difference in timing. It only takes a .1ms difference in arrival time between two mics to shift the stereo image pretty noticeably.

[dcollins]

Originally Posted by AlexDavis

I'm wondering if there are any plugins that will make sure that the volume levels between the two channels are relatively equal. I would need some solution that can be batch processed, because I don't have an extra month to spend pouring through these files one at a time... again.

The answer, as it always is when "automatic" and "audio" are used together, is:

Nope.


DC

[AlexDavis]

Thanks guys for the thoughtful responses!

Andrew - narrowing the stereo image is def something we are considering, thank you for giving us that idea!

Laurend - Audition has the capability too (RMS Normalization), and although the realist in me is fairly sure it won't work, I'm going to give it a go. THANK YOU

Wado - Excellent point, it's amazing how sensitive the ear is to these tiny, tiny differences. I'm sure it's not just about volume. I'm going to experiment with changing the timing (by small, small amounts) as a means of matching the samples.

dcollins - I tend to agree with you, HOWEVER, I've been surprised by what can be automated, as long as one remains vigilant to check the results. We have been using Melodyne to tune many of our samples, and it's even possible to create a batch processing workflow using QuicKeys... But like I said, even in an ideal scenario, 1-2 percent of the samples will be bogus. So careful inspection is necessary.

After all this, it seems easier just to scrap one of the mics and go mono. But I'm not quite ready to give up!

[JTransition]

Sounds like you need an intern.
Good luck

[Thomas W. Bethe]

Wavelab can do batch processing and has a very nice processor for panning correction.

I too agree that automatic and audio do not really belong in the same sentence but I understand what you are about to undertake. Been there-done that more than once.

Suggest hiring someone just for this project???? I am sure there are a lot of people looking for work with good backgrounds in audio.

[AlexDavis]

Thomas - can you show me some info on that panning correction? I don't own Wavelab but I'd love to learn more about that!!!

Yes, an intern... God that would be nice

-Alex

[Thomas W. Bethe]

Originally Posted by AlexDavis

Thomas - can you show me some info on that panning correction? I don't own Wavelab but I'd love to learn more about that!!!

Yes, an intern... God that would be nice

-Alex

From the WL7 Manual

4.6.12 Pan Normalizer
This dialog allows you to ensure that both channels of a stereo file have the same level or
loudness. Since loudnesss is more important than peak volume in the perception of sound,
this is a powerful tool to get the best possible stereo balance.
This process uses two passes to first analyze the audio and then render any level changes
required.
You must have a stereo selection in a stereo file to apply this process.
You can access this dialog from the Audio File Workspace via Process > Pan Normalizer....
This function is also available as a batch processing plug-in. You can access it from the Batch
Processor Workspace via Plug-ins window > Multipass Plug-ins > Pan Normalizer.
For an explanation of each parameter and interface feature, click on , or the 'What's this?'
question mark icon.

\Full manual here/ ftp://ftp.steinberg.net/Download/Wav...veLab-en-7.pdf

[AlexDavis]

THANKS FOR THE TIP! I might consider buying wavelab just to see if that works!!

[Foldedpath]

The free Audacity program has an option for applying Normalize separately for each side of a stereo track:

Normalize - Audacity Manual

This is just peak normalize, and if the Wavelab function does RMS that might be more useful. Regardless, I've used that setting in Audacity a few times for balancing stereo tracks, since there isn't an equivalent function in Samplitude that I've been able to find. Audacity can also be set up for batch jobs, but I've never tried it.

[mixkitchen]

Pro Tools splits interleaved files on import. Import to two mono tracks, select all, normalize, export audio as files from the menu in the region list.

[kiopo]

How big a deal is it that the image is slightly different, for the project? I wouldn't want to risk making something sound over-processed for the sake of a bit of natural movement. If it's really vital to have it in one place, mono and a touch of realistic reverb would be my bet. That, or narrowing the stereo field.

Peak normalizing will not help you in this situation. I'd be surprised if RMS did either, the difference in tone will be very noticeable.

[Alécio Costa]

I will sugest taking one saturday-sunday off to stay at the studio doing the thing the best as you can...

[Cheebs Goat]

What is the nature of this project? Is it a solo instrument with a stereo room sound? Why are there 2000 recordings? Is it for a sample library instead of a song? Is that why center placement is so vital?

The answer sort of depends on what you're trying to do, but so far the best idea sounds like it came from the dude above me: sum it to mono and fake the stereo with a verb plugin.

[whiteaxxxe]

this is not the most professional hint ... but ... (I am not sure, if I should really say this, but I do):

AAMS does in cases when there is no need to be perfect a good job. it is able to batch the whole lot of your files, including eqing, compressing (in a suitable way) and panning. dont know how it does it, but it does.

if you need a solution fast, I would give AAMS a try.

(sorry, I know ... automatic mastering ... awful. but sometimes ... )

[Laurend]

Originally Posted by dcollins

The answer, as it always is when "automatic" and "audio" are used together, is: Nope.

I can't agree with that statement. The real question is: How many months (years?) are you ready to spend developing an effective automated process to save a week of manual job? Just imagine you have to test different processing solutions, to select one, to tweak it, then apply it to all samples and check if the results are correct. If it isn't OK, then restart the development of the processing from the beginning. I'm sure the manual option will give the fastest results.

[dcollins]

Originally Posted by Laurend

I can't agree with that statement. The real question is: How many months (years?) are you ready to spend developing an effective automated process to save a week of manual job? Just image you have to test different processing solutions, to select one, to tweak it, then apply it to all samples and check if the results are correct. If it isn't OK, then restart the development of the processing from the beginning. I'm sure the manual option will give the fastest results.

Does your machine have good taste? Sensitive aesthetics and a familiarity with a wide variety of music and client needs?

"I know I've made some very poor decisions recently, but I can give you my complete assurance that my work will be back to normal. I've still got the greatest enthusiasm and confidence in the mission. And I want to help you."


DC

[Laurend]

Machines don't have any easthetic sens nor any musical culture. They just do the exact job they have been designed for.

[dcollins]

Originally Posted by Laurend

Machines don't have any easthetic sens nor any musical culture. They just do the exact job they have been designed for.

Next you're going to tell me the AE-35 antenna was fine all the time.


DC

[antithesist]

"I'm sorry, Dave. I'm afraid I can't do that."