Topic: Blend to mono selectively - channel removal??

[chk]

I've got a stealth stereo recording I made and the person a couple seats away (one of my daughters, ha) was a bit spirited in her sing-along'ing.

I'd like to selectively blend the recording to the other channel in certain sections of the show. Still want to retain as much of the true stereo image as I can but it's not super-critical (it's a stack recording of Zach Bryan from 3/31/24 and came out pretty good).

I am experienced with doing mixes/matrixes in Audacity. But in this case wondering if there's a straightforward way to do this? I know I can break the tracks out independently and fade in/fade out with the good channel / bad channel, but that seems like it would be pretty labor intensive.

Would love to be able to select certain sections in Audacity and make it two channel mono with the good channel, is something like that possible?

Not sure if there's an easier way to do this and am open to suggestions. I've tried RX and other software fixes, but this is too random and I've decided doing the channel combining is the best solution since it sounds pretty good that way.

Thanks guys.

-chk

[if_then_else]

You could also cross-fade between the channels. (Respectively fade the "problematic" one out and in again.)

[AbbyTaper]

If all you want to do is make certain sections dual mono, just copy the channel you want and paste it over the same section in the other channel.  It will be noticeable where the transitions are, obviously.

[kuba e]

I don't use Audacity. But If_then_else says that it is possible to do crossfades in Audacity, then it should be easy. I don't know if I understood correctly. Do you have a second stereo source or just want to replace the right and left channels with each other which picked up less voices? The procedure is similar in both cases. The following is for replacing the left/right channel. Copy the desired left/right channel to a new stereo track. Adjust stereo width (e.g. free plugin ozone imager from izotope) on the new stereo track so that it is as close to original recording as possible. And then do a slower fade-in/fade-out for the original recording and the copy at the desired places.

Keep the original recording as well. Sometimes I talked to my friends while recording. I always tried to remove it. But over time, it's fun for me to listen to our conversations.

[robgronotte]

Everyone who has kids, please teach them concert etiquette while they're young. If you're singing loud enough that others can hear you, you're singing too loud. And if you're talking loud enough that anyone other than the person you're talking to can hear you, you're talking too loud.

[kuba e]

It depends on the type of concert. Somewhere audience singing is welcome. I don't mind funny comments either. I personally like it a lot when the audience is alive, cheerful, and it is captured on the recording. Of course there are limits, what is already annoying. Everyone perceives it differently, so tolerance on both sides is needed.

[Gutbucket]

Izotope RX or Ultimate Vocal Remover?

If not.. or possibly in combination..

What are the good stereo qualities of the stack tape you would like to retain?  Does it convey actual stereo image placement of sound sources?  Or is the stereo aspect of it mostly an more open non-monophonic portrayal of audience. reverberance, and room sound?  If more the second than the first (which is typical of a stack tape), and the clean channel sounds good on its own in isolation except for being less involving and somewhat dimensionless in a stereo sense, you might use just that channel and apply some flavor of pseudo stereo processing to it.

Advantages of that approach are that the work will mostly involve dialing in the most appropriate pseudo-stereo sound, followed by applying that to the entire recording, rather than doing lots of cross-fading (less post work), and it avoid transitions back and forth that could be noticeable and distracting.

There are a number of pseudo-stereo techniques you might use, which has been covered in other TS threads.  Most of those will use just the "good" channel and discard the other one.  I like using Mid/Side processing for stuff like this, partly because it can retain both original channels- if you can EQ the channel that has the extraneous singing in it such that the voice is significantly attenuated (deeply scoop out the mids, leaving the lows and highs - don't worry that it sounds incorrect on its own) you can use that as the Side channel and the clean channel as the Mid channel.  The channel with the attenuated signing then becomes the "stereo effect channel".  After Mid/Side processing its content will be heard out to either side, with the clean channel dominating the center of the playback image.  Stereo involvement will manifest mostly in the low and high frequency regions, while the midrange will be focused and centered.  Any attenuated extraneous singing that may still be audible will tend to be diffuse and out to the sides.

[chk]

Thanks for keeping it on topic Gutbucket and for your thoughts.
It’s not terribly loud singing, only in a few parts here and there. But it’s a little distracting and she would love for it not to be as noticable on tape (lesson learned). So really just a matter of making the content more enjoyable.

The image will lose some depth in mono and i have worked with M-S processing and some of the stereo-izing imaging tools.
But that is certainly the easiest route and best bang for the buck to just do that. It’s not a show or tape where it makes too much difference in a true stereo image.

Optimally, i’d still love to find a way to selectively blend the single channel to L/R if anyone has any more ideas…all ears.

[nulldogmas]

Ultimate Vocal Remover is definitely worth trying, though if she's singing along with the vocals, UVR might bundle both into one track, which doesn't help.

What you really want here, if you're trying to add in a bit more good channel to the bad one but not go full mono, is some kind of Mix Paste option, which I don't think Audacity has. But you could maybe fake one by splitting the channels into separate tracks for just the problematic sections, then playing with the gain and balance sliders until you get the optimal mix? (I.e, if the bad channel is the right one, try sending the right-channel track to be panned far right at a reduced gain, while the better channel is panned just left of center at a higher gain.)

[Gutbucket]

Ultimate Vocal Remover is definitely worth trying, though if she's singing along with the vocals, UVR might bundle both into one track, which doesn't help.

In that case, try applying UVR to just the channel that contains the "extra" vocals. In that way the other channel with the primary vocals in it won't be effected.  If in the targeted channel it unavoidably causes a reduction of primary/wanted vocals as bycatch/collateral-damage, the aformentioned Mid/Side technique that uses the channel with the clean vocal as Mid will keep the good vocals centered and any attenuation of that in the other channel will only effect the apparent stereo width of the vocal.

Alternately you might use UVR or RX to stem out the good/desired vocal from the clean track and then pan that to center or however is necessary.

[nulldogmas]

Alternately you might use UVR or RX to stem out the good/desired vocal from the clean track and then pan that to center or however is necessary.

Oh, duh, this is a *much* better idea. Essentially, make the vocal track mono, and leave the rest stereo.

And UVR is free, so it can all be done at no cost.

[chk]

Thanks guys will play around with UVR and see how that works. Never used it before.  Like these ideas. Will report back… Thanks

[robgronotte]

I have been able to remove some near audience singing along with Izotope RX, but it's very tedious.
Basically I find the place where I can hear audience singing, zoom in on it, alternate between viewing left and right channels until I find the alternate vocal line which is louder (brighter) in one channel than the other, then trace and delete that vocal line.

[chk]

Yeah luckily it’s not throughout the recording, but RX would be way too tedious.
Playing around with UVR. Amazing tool. It’s gonna do the trick I think. Thanks

[chk]

UVR is an amazing tool.
Ended up going the easy route - pulled 4 tracks. Stereo L/R for both instrumental and vocal. Used the 2 instrumental tracks as-is, the better vocal track mixed to dual channel, little bit of ozone stereo-izer, and then some processing and clean-up.
Thanks for the suggestions guys. UVR is incredible considering it’s free. Has to impact iZotope’s business but that’s another topic…

By far the best model on UVR I tested was MDX-Net; MDX23C-InstVoc HQ
Very little artifacts or noise. Granted i didnt test a ton but did a little digging around prior to testing a few.

Hope this helps someone. Can also be really helpful bringing up vocals on poorly balanced recordings, etc, etc.

It won’t be long before you’ll be able to pull multitracks from stereo recordings.

AI is silly.