Vocals too hot....What to do in post?

craftech wrote on 4/12/2004, 10:18 AM
I taped a musical a few months ago and had to use a body mike feed from a kid who didn't know what he was doing at the sound panel. After running my own stage mikes I ended up with an awesome sounding recording of the orchestra and the chorus, but also fed into my mixer was the body mike feed from the aforementioned kid and the final result were vocals mixed in which are way too hot. My mixer had to be set and left while I shot the video and I guessed at the body mike feed level because the kid wouldn't let me do a sound check off his panel.
Since this is the first time I have run into this problem I am not sure which audio correction tools to use within Vegas to fix this. The vocalists are at an ear piercing level while the rest sounds great. I don't want to lose any of the overtones and harmonics of the rest if possible.
I have the standard tools in Vegas and also Sonic Foundry Noise Reduction to work with. It will be AC3 audio in the end for a DVD.
What I am looking for is the general procedure. Which tools? What frequencies should I work on? Would "normalize" make it better or worse? etc. Specifics would be greatly appreciated if any of you have the time to write them.

Regards,

John

Comments

Chienworks wrote on 4/12/2004, 11:05 AM
Normalize won't have any effect since it operates on the entire audio uniformly. This is very similar to the "remove the vocals from the song" question, though not quite as blatantly impossible. ;)

Probably about the best you can achieve is to use EQ to reduce the standard vocal frequencies and hope that there is enough of the background sounds below and above this range left to balance. If the offending vocal is mono and pretty much dead center in the mix then you can use the vocal removal techniques to produce a (very crappy) version with his vocal mostly cancelled out. You could then mix this in with the original to get something somewhat inbetween (although with lots of potential phase cancellation problems). Good luck!
MJhig wrote on 4/12/2004, 11:41 AM
I agree with Kelly that you are most likely not going to "fix" this.

You don't say you have Sound Forge but just in case you forgot to mention you do, here's what I would try;

Load the file in SF, observe the Spectrum analyzer while playing and get an idea of the what frequencies the majority of the vocals occupy.

Set up a band in Multi-band Dynamic to correspond with that band of frequencies and compress heavily, also each band has it's own gain so lower the gain in that band and raise the other bands.

Simple example to give you the idea;

Vocals occupy 1 kHz to 4 kHz

Set up band 1 = Low shelf to 1 kHz
Band 2 (vocals) = Band > 2 kHz Center > 2.5 Oct.
Band 3 = High shelf 4 kHz

Again don't expect miracles and you can substitute another multi-band compressor and Spectrum Analyzer if you don't have SF.

MJ
larry-peter wrote on 4/12/2004, 11:47 AM
Here's something you might try that saved my butt once. It's a tweaked version of the "Voice removal" trick, but doesn't sound quite as bad and can save a little bit of stereo imaging. As usual, the vocals have to be center and mono for it to work at all. Use the track "duplicate" function to make a synced copy of the mix. Apply a "band-pass filter" plug-in, with as sharp a roll-off as you can get from 400 Hz to 2500Hz. (You might be able to get close with a multi-band EQ, but you'll do more damage to the phase of the signal and results won't be as good) Then invert the phase of the filtered track. As you adjust the level of the filtered track you should be able to attenuate the primary voice frequencies without affecting (much) the high and low freqs of the music. Play with the low and high cutoff settings on the filter. The midrange is gonna have wicked cancellation problems, but you might get something salvagable out of it. Dittos on the good luck!
craftech wrote on 4/12/2004, 2:10 PM
Thanks for the responses.
I don't have Sound Forge, but I'll give some of the above a try. I think I have enough really good audio to play with the equalizer in the vocal frequencies.
1Khz to 4Khz is what you said. Any rules of thumb here? The female vocals are really the most noticeable in terms of ear piercing levels.

I hate the thought of destroying my own really excellent stage and orchestra miking because I was stuck with a body mike feed from an arrogant and ignorant high school kid.

I take it the Noise Reduction plugin won't help. For what it cost, I haven't used that tool once. But I guess I haven't used the audio tools in Vegas that much either, except to raise and lower the audio envelope in spots when certain specific parts needed it.

Regards,

John
MJhig wrote on 4/12/2004, 3:12 PM
Can you post a minute or two of this track in MP3 format on the web (Soundclick if nothing else) with all faders in Vegas at 0 dB and plug-ins in the track bypassed?

I can give you some suggestions if so even with just Track Compressor possibly.

MJ

Arnar wrote on 4/12/2004, 4:00 PM
Best way to do this is to put up a chain of plugins like this...
( a very well kept mastering trick)

Ms encoder ( mid side -Waves makes one)
Eq
Ms Decoder.

this will split the signal into the Mono component and the stereo component(mid = Left /Stereo = Right), and you can eq the mono component seperately by splitting the eq so it processes each side (unchain the eq so it processes left-right seperately)
or you could simply just lower the middle?

you could even insert a compressor to compress just the mid info.

Dont tell any mastering engineers that i told you this...



LarryP wrote on 4/12/2004, 6:10 PM
Bob Katz goes into MS techniques in his book "Mastering Audio". Fine book IMHO.

Larry
craftech wrote on 4/12/2004, 8:12 PM
Can you post a minute or two of this track in MP3 format on the web (Soundclick if nothing else) with all faders in Vegas at 0 dB and plug-ins in the track bypassed?

I can give you some suggestions if so even with just Track Compressor possibly.

MJ
-----------

I signed up for Soundclick and will post an MP3 sample as soon as I figure out how. So far I can't find anything on the site which tells me how to upload an MP3.
I am printing everyone's suggestions and will try them all. Thanks for the overwhelming response to my dilemma.

Regards,

John

PS: Just found my answer regarding Soundclick, but I guess I can't upload this material to their site:
"Important: copyrighted samples, loops from other artists, or remixes are not allowed. This affects all charts, not just this new genre. " Thanks for the offer to evaluate an MP3 sample of the audio track though.
SonyTSW wrote on 4/12/2004, 9:59 PM
craftech,

>> I don't have Sound Forge

There is a limited time fully functioning version of Sound Forge 7.0 you can download from here:

Download Sony Trials and Demos
craftech wrote on 4/13/2004, 6:29 AM
Thanks Tracy. Really appreciate it.

Regards,

John
MJhig wrote on 4/13/2004, 6:42 AM
You could post it on Chienworks Vegas users site I believe.

MJ
craftech wrote on 4/13/2004, 12:59 PM
You could post it on Chienworks Vegas users site I believe.

MJ
---------------------------
Thanks MJ,

I uploaded a sample from three songs here. It should appear as "Mpeg3Sample" under my name (Craftech) when Kelly approves it.
http://www.vegasusers.com/audshare/


John

6 hours: Still waiting for it to be posted
craftech wrote on 4/14/2004, 6:05 AM
Thanks Kelly,

Regards,

John
MJhig wrote on 4/14/2004, 12:37 PM
Ok John,

Post your email address so I can send you the *.veg file. As I said, don't expect miracles with this type of "fix" especially with only the standard Vegas tools. I was able to get much better results with a multi-band compressor but so be it.

This girl is all about POWER even if her channel wasn't too hot. Her harmonics cover the whole frequency spectrum making it more difficult to isolate her.

She's a good singer and under the right conditions could be made to sound very good but this is a stereo file and I can't listen to her anymore today, yow ;-)

MJ
craftech wrote on 4/14/2004, 1:08 PM
Thanks for the help MJ.


Regards,
John
MJhig wrote on 4/14/2004, 1:48 PM
OK, it's been sent. Let me know if for some reason you don't get it. You might want to edit out your email addy before the spam-bots find it.

MJ
craftech wrote on 4/14/2004, 3:38 PM
Got it..............

Thanks again MJ

John
craftech wrote on 4/15/2004, 6:36 PM
MJ,
I saved the effects chains you sent me as an Audio Left Preset chain and an Audio Right Preset chain, but am unable to figure out how to apply one to one side and the other to the other side of the audio track in my video.

John
MJhig wrote on 4/16/2004, 6:42 AM
The chain is the same for both but you need to split the stereo track to 2 mono tracks.

Assuming now you have one stereo audio track with one of the plug-in chains, Right - click the track's header > Duplicate Track.

Right-click the event in the original track (or Select events to end to select all events in the track) > Channels > Left only, do the same for the duplicate track except select Right only.

Now pan the original track 100% Left and pan the duplicate 100% Right.

Right-click on the pan fader for both tracks and select Balance -6 dB Center.

The reason for this panning is it takes her voice out of the more obvious center reducing her presence a little more.

MJ