Subject:Vocal EQ ?'s

Posted by: larryo

Date:11/27/2001 6:58:32 PM

Using Sound Forge 5.0 as an audio editor on Vegas, I've run into some frustration trying to get a good vocal EQ. On the front end, I use a Shure Beta 58, a CAD Equitek, a SM 57. This runs into a tube line processor, then to a Eurodeck little mixer. This then runs to a Delta break out box where I sample in stereo 24 bit/96. You'd think I'd get best results using the CAD, but I'm pretty pissed off with this units overall sound. The best "natural" sound for my funky mid-range warble is the Beta 58. Isolating the track after recording, it sounds pretty good. I open in Sound Forge, mute the puffs,moans and garbage in between sentences, I correct DC offset, I normalize. Clean, even, good sounding as a stand alone track. I open track back in Vegas and it's dull. I add effect, EQ, compression, I redo, I undo, I quit. Using spectrum analysis, I find my peak frequencies typically are in the 300-400h range. My question is to the experienced:

Is there a general EQ setting most of you use that feels good in most mixes? I know a vocal can get crowded by other instrumentation/panning/etc. I'm talking about an in-your-face clean but not over compressed crisp presence? (minus the purchase of a Neumann mic). Any suggestions regarding using effects/EQ/Wave hammer & spectrum analysis will be greatly appreciated.

LarryO

Subject:RE: Vocal EQ ?'s

Reply by: VU-1

Date:11/27/2001 11:57:15 PM

Congratulations! You have fallen into the same trap as countless other mixing engineers.
Talk to any of the BIG BOYS (Al Schmidt, Mick Guzouski, Lord Algea, and so on) and the first thing they will tell you is 'DO NOT solo your tracks in order to get them sounding right for the mix. You'll get them sounding great by themselves but as soon as you pop them back into the mix, they will fall apart on you.'

You touched on this in your post. Nine times out of ten, the correct sound of the vocal in the mix sounds way off when soloed. As you said, there are alot of things fighting each other for attention in a mix. You have to find the right "spot" for each one in the freq. spectrum so that they all work together.

The attributes you mentioned: "in-your-face, clean, crisp & present" all suggest a dominance in the upper-mids region. The presence range is about 2k-5kHz. However, down around 2k-3k, the sound is a bit "hard". Try using between 5k-6k and maybe even up to 7k. This will add some sizzle to your vocal and help it to cut thru the mix. You can beef it up by boosting down around 230-285Hz. Use a medium Q up top and a narrower one for the bottom. You said your peaks were 300-400Hz. I think that's a little high for the guts of the vocal. There are many other instruments that can fill that space quite well (guitars, for one). If the vocal starts to sound thin, add in a little bit of something in the middle - somewhere between about 800Hz and 1.3kHz. I always apply EQ AFTER compression (watch your levels!). USE YOUR EARS!!

You said you didn't want it over-compressed. Actually, if you compress it too much, it will lose its punch. Could be that you are using too fast of an attack time - that chops down all the transients that it needs to have that punch. Go for a medium attack time and a short-to-medium release time. Can you give a little compression (1-3dB) on the way in? Admittedly, its much easier to set a compressor using a VU meter (than a digital peak meter), but try to get about 2-4dB gain reduction on the vocal while mixing. Also, sometimes "in-your-face" is achieved by simply adding more volume.

Another thing you may need to pay more attention to is the way you are (or are not) working the mic while singing/recording. Alot of times, a good close-up vocal sound is just exactly that - close up (to the mic). I recommend that you DO NOT pipe any effects (even reverb) back to your headphones while recording. While it may make you sound like you're preforming better with them in, you will not get a clear picture of what you are really doing. A good analogy is playing a guitar with & w/o distortion. It always sounds like you can play better & faster when the distortion is on. Truth is...its just smoothing the sound out and covering up alot of mistakes!

I'm curious as to why you are using Sound Forge to "mute the puffs,moans and garbage in between sentences". You can do the same thing - much faster & easier - in Vegas. (True, DC Offset must be done in SF - do it first, then save.) Use the output volume on the last plug-in in the vocal chain to reach desired max track level w/track fader set to 0dB. Then you can adjust the track fader to mix (or use volume envelope).

I hope all of this helps. Good luck!

Jeff Lowes
On-Track Recording

Subject:RE: Vocal EQ ?'s

Reply by: larryo

Date:11/28/2001 8:04:30 AM

Thanks, Jeff (OTR), for the response. My use of Sound Forge to mute unwanted transient noise on vocals is simply because I've been a Sound Forge user before a Vegas user(started with SF 4.0xp with CD Architect). I've been playing with the envelopes and such in Vegas and see alot of the functions I have been performing in SF can be done in Vegas. I gave some misinformation regarding my dominant frequency on my vocal takes. Actually they're typically in that 230-285HZ area. Regarding compression on the way in - I don't currently have the gear. My tube processor adds a bit of a compressed feel to it, so this does help somewhat. Let me pose another question if I can? I've gotten into the habit of normalizing tracks in SF after running many "statistics", finding the hot spots, fully zooming in on them, highlighting the wave (from valley-peak-valley), dropping volume of these areas (typically 33%), re-normalizing, and starting all over again until I'm confortable with the overall level. Kind of a "selective compression". This has worked well for me in SF in the past for 2-track mixes, but in 5.0 using 24bit/96, I notice some strange little pops that left in these regions. This has occurred on some vocal takes that in experimentation I converted to mono. So my question is, is there a quicker and more practical way to get these results for effective normalization of an entire track without being slave to a few nano-second hot spots? Thanks again for all the info.

Larry

Subject:RE: Vocal EQ ?'s

Reply by: VU-1

Date:11/28/2001 9:50:50 AM

I can think of 3:

1) Normalize the file using RMS normalization instead of Peak normalization. Set the "If Clipping Occurs" box to "Apply Dynamic Compression". This does a pretty good job unless you really bump up the RMS level.

2) Use the Track Compressor, Wave Hammer (Compressor part), or the Graphic Dynamics across the entire file. Set the Attack Time to VERY fast (.1 - 1.0 ms) & the release time to a few ms (5 - 15). Adjust the Threshold & Ratio to catch and drop the peaks (use the wave view to figure these settings out).

3) If its certain freqs. that are bothering you, use Multi-band Compression. Do the same as above, but sensitive to frequency ranges.

Incidentally, Multi-band compression can be used to clear up a mix & pull different elements out of a mix.

JL
OTR

Jlu!