Hi,

I spend hours trying to match the central channel in order to sound the tracks like the singer is standing in the fixed position from the mic. EQ'ing the central channel so that the voice in the tracks is at the same levels.

Is there any way I can automate this process and save my hours of hardwork?

Any help will be greatly appreciated.

Thanks.

 

What do you mean? I think you need to be a bit more specific and detailed.
Lead vocals are 99.9% of the times mono recordings and centered (and located dry in the Mid channel). The spatial effects (reverb, delays, etc) are usually a mix of Mid and Sides in stereo.

 

Last edited: Dec 31, 2021

Not talking about reverb or delays. I'm talking about the levels of the vocals in the stereo track. How do I automatically adjust the levels of the vocals in a track to match that of another track?

There are bunch of EQ matching plugins that promise to do this but leaves lot to to be desired and manual adjustment. Is there any specific plugin/method I could apply to get the work done faster in more efficient way?

Hope this explains it.

 

You ride/automate the volume of the vocal. Then compress (maybe even in several serial steps/compressors). That way you get a more consistent loudness. Watch out for sibilance (which you can de-ess) and breaths, which you also can volume automate down, if needed. Some clip-gain sibilance and breaths.
There's also parallel compression for vocals, to make it more compact and consistent.

You then mix the track so that the vocal sits well (relatively and subjectivly). Then you level-match the other songs on the EP/album so that the vocal has somewhat consistent loudness throughout. The vocal is usually the "red thread" that ties multiple songs together. I do this as a mastering engineer.

 

Last edited: Dec 31, 2021

Thank you for taking interest in my thread Baxter. I'm working with already mastered and finalized tracks. Making anthologies. Any easier way to accomplish this?

 

Usually you can match the perceived position (front to back) of the vocals by looking at how loud the area around 500-1000Hz is in comparisson to the rest of the instuments (by using an analyzer on the mix/master bus), then processing accordingly (with EQ, compression, on a bus or individual channels, whatever you feel will work best). Too much level on that area will make the vocals feel closer, too little and you'll bury them, regardless of how loud the vocal channel is on it's own.

How much is too much really depends on what you're doing, but it's a nice little trick to compare stuff.

 

iZotope Ozone Master Rebalance lets you adjust the volume of the vocal in a finalized track. You'll still have to automate the volume manually.
Either way, this will mess up the mix. Make sure your client knows there's no way around this without the stems, at the very least.

 

Not really. You have the vocal as loudness reference and turn down the volume of songs that have "too loud vocals", if that is really a concern. Otherwise just normalize the overall percieved loudness (LUFSi) of each mastered song (turn loud ones down to reach a common loudness value).

 

Last edited: Dec 31, 2021

Thank you for the response. That sounds like a neat trick. I use a similar trick for overall loudness consistency across the mix. How do I do such analysis. I'm using iZotope RX8.

 

RX also has this feature but this doesn't make the levels of instruments in the entire mix to sound consistent. Based on my tests, it altered the makeup of the track differently based on the complexity of the source material.

I need something like Loudness match that also works for the vocals. Thank you for the response though.

 

Will this also normalize the loudness of the vocals?

And what happens if I do this twice? does running it twice introduce any artifacts?

 

I'm confused about what you are asking. The vocal is fixed with the rest of the music. You loudness normalize (turn the volume down) so you get similar loudness as the other tracks. Just like Spotify/Youtube/Tidal/etc does. I don't know how else to explain it.

 

Last edited: Jan 1, 2022

In standalone mode? Don't know about RX8, but RX9 comes with an analyzer called "spectrum", it's probably on RX8 too. The only issue is that this analyzer in particular doesn't have a slope funcition (that i know of) so visually it could be kind of hard to tell what's going on, but it's there. I think you can actually use external plugins, in which case you can get a free spectrum analyzer like voxengo's SPAN.

It's pure especulation though, i never use RX in standalone mode.

 

Hi, @ceo54 . No, there isn't.

There is not a way to logically define your query other than rephrasing it to mean you are looking for a way to automate leveling out "perceived" loudness of lead vocals, across an playlist of recorded works in their final mixes. In normalizing, quite as @Baxter is already pointing out, the mean (as in: midway between extremes) of perceived loudness across a given playlist is itself the normal. This can not be differentiated for lead vocal - or any other component of the mix - other than processing it in isolation from each individual mix, and regardless of their integral normal. This leaves you with contradiction in logical terms, rendering any answer to your specific question "subjective, if true", or: false. This forbids automation. To objectify your purpose in AI would require feeding it the preferred (= subjective) settings for an immense number of example mixes, in a reference playback system, to a given set of ears (yours) - which cannot be achieved other than... manually. Subjectively ⇒ by you alone, for your case.

_

 

I understand, don't worry about it. And thanks for the help.

 

Thanks, will look into it.

 

Thank you for the detailed explanation. What you said makes sense. I'm new to music production and have very limited use for it. I will keep my options open and see if, with time, I can discover any initiative and effective way to accomplish this.

 

keeping the power or peak of a track somwhat on the same level all the time - normally any compressor with a "gate" and an "autogain" function can do that.

but i am afraid this does not make sense at all.