Minimising vocals on background track

D

deltoidjohn

Veteran
Hi All,

I've seen methodology for solving the below problem here and on other forums before, but can't find anything with the search function here or on Google - I suspect as a result of me not using the correct keywords.

I've got a background song with vocals which I am trying to mix underneath my main dialogue audio (it's an interview on a lifestyle tv show, recorded with a pair of lav's). I'd like to target just the vocals on the main track and lower the level, without completely lowering the volume of the whole track - anybody able to help me out? Do I use the EQ effect? Or Multiband Compressor? What frequency should I be targeting for the vocals? I know I can't completely remove the vocals, I'm just trying to balance them a bit better against the dialogue over the top and make it a little less distracting.

Usually I'd prefer to run a song with no vocals, or to get both a vocal & instrumental only version and mix them as required, but in this case I've only got the vocal version.

Thanks in advance for any help!
 
I don't think that's possible with PP. Maybe Audition, but I'm skeptical the results will be any good. The ideal here is to have the music and vocals on separate tracks, separate files.
 
Last edited:
That sounds tricky....mainly because I presume there's a fair amount of frequency-range overlap between the singers on the background track and the speaker dialog you want to preserve. However, I've sometimes been able to improve tracks like this.

Like Jim S suggests, I'd try perhaps Audition (since you'll have access to it if you're on PPro CC) and specifically the Sound Remover effect:
https://helpx.adobe.com/audition/how-to/audition-sound-remover-effect-cc.html
I've had mixed success with this effect, but one hint is to use several passes to remove the audio you want removed.

I've had more success with the similar, but in my experience better, tools in Izotope RX. However, RX isn't free:
https://www.izotope.com/en/products/repair-and-edit/rx.html

And then some EQ, compression, etc to reduce the bad and boost the good. But really, it depends on the particular track and problem. And experience really helps here. Can you perhaps pop the track over to an experienced post-production audio gal or guy? They'll have the tools, know which knobs to turn, and how much to turn them.

But if you want to or must DIY, check out various online tutorials. A good place to start: Izotope's RX Cookbook and Audio Repair guides. The concepts will help you use whatever tools you have available. Good free stuff:
https://www.izotope.com/en/support/support-resources/guides.html

Good luck!
 
deltoidjohn said:
I've got a background song with vocals which I am trying to mix underneath my main dialogue audio (it's an interview on a lifestyle tv show, recorded with a pair of lav's). I'd like to target just the vocals on the main track and lower the level, without completely lowering the volume of the whole track...
Click to expand...

That's sorta like asking to take some of the salt out of the cake you just baked.

The problem is that your vocals undoubtedly use the same frequencies as the music you don't want to disturb. How do you tell the two apart? This is similar to the de-reverb problem, except that a couple of companies have actually made some strides with de-reverb filters. But vocals are much more of a challenge.

If you had the stems, you could just lower the vocal track(s) in the mix, no problem. But that requires going back to the mixing stage, and I'm pretty sure you can't do that based on how you worded your post.

That said, some of the royalty free sites sell their music with and without vocals. That is, full up or just the "music bed" that underlies the vocals. If that's possible you might ask if they'll mix you a special version with lowered levels on the vocals. Long shot, but you don't give enough information to be more specific.
 
I've managed to get a multi-track version form the record label so it's all sorted now.

I'm surprised to hear so many people saying it can't really be done though. I've watched tutorials showing how it's done, I've read the theory on how/why it works, I've heard the effect in practice in plenty of films, I just can't find any of those tutorials again now. It's not like I'm trying to remove the "salt" to create a karaoke track, just flatten the whole cake so I can balance a pie (dialogue) on top of it!


For the record though, I always leave the salt out of any recipe when I'm baking a cake.
 
Ah geez; I totally misread your problem. But ya, the separate-and-invert method (as in at least one of the above links) in just about any decent audio editor (including the free and good Audacity) works well...or you can use a plug-in to automate the process...or use some EQ... But your method of getting a multitrack version is best. :)

Sorry I didn't help. Glad you're sorted.
 
You must log in or register to reply here.
Back
Top