"So I've got an automation envelope that ducks the 200Hz band in the Izotope Mastering EQ music like this while I'm speaking"

I think you'll get better results moving the centre frequency to around 1,200Hz, try upping the Q also.

All you need to do is make the voice intelligible and the vital information in speech is contained in a band of frequencies much higher than 200Hz. When I've used that technique I've used the basic Track Eq as it's less CPU hungry. Never bothered ducking it either but YMMV depending on your content. Just take care any ducking isn't noticeable.

Bob.

For all my radio mixes, and most TV mixes, I use a combo of Bob's technique with a slight amount of overall volume reduction in the music track. I think listeners expect background music to soften when a V/O begins, but by slightly reducing freqs centered around 1.2kHz - 2kHz (for a typical male voice, up to 2.5kHz for most females) you can get away with a lot less overall volume reduction on the music track and keep the mix sounding fuller and the voice intelligible. And I always manually set envelopes for both volume and track EQ rather than automating it with a sidechain/compressor/ducker style of process. If you change the EQ and volume envelopes simultaneously, the EQ change can go unnoticed and the volume reduction is less than normally needed. To me it's much more pleasant sounding than automated "ducking."
Larry

Edit: I can only speak for US broadcasters, but another advantage to going this route rather than just ducking the whole music mix is that - especially since DTV - compression/limiting of broadcast media is completely over the top. If you just provide a mix with decent volume reduction on the music track, by the time the broadcasters turn your nice waveform into a flat black line it's going to be pumping and wheezing like you won't believe. Take control of dynamics away from the broadcasters every time you have the chance.

Nick, have you played around with side-chaining? You might get a better result (assuming that's not what you're doing already).

Thanks for the replies everyone.

b.complex - Can you please explain to me what the terms "sub" means in this context? And what sort of settings would you use on the multi-band compressor? I have the Izotope 4-band compressor.

Andy_L - I did have some success with the DB Audioware Sidechain Compressor. I put it on my vocal track and chose the "Send A" preset. Then I put it on my music track and chose the "Receive A, 18dB ducking" preset. It does what it's supposed to, but I still find that it's too "pumpy" or "up and down" for my nature documentary style. I improved it somewhat by increasing the release to the maximum 1000 ms. I don't think it's going to be the tool for my projects, but for a quick, "automated" ducking it's great. If anyone tries this, bear in mind the author's 9-year-old advice about placing the sending track above the receiving track. Probably also a good idea to mute the video and shut everything else on the computer down when doing the audio render to minimise latency.

farss - I'll use the standard Vegas Track Compressor, not the iZotope one. Apart from being less CPU-intensive, zero value on my existing volume envelope (that I created with Excalibur's voiceover function) equates to zero on the EQ plugin, so the envelope is immediately usable when I copy the points from my volume envelope and paste them to my EQ automation envelope, whereas they don't equate to zero on the Izotope EQ so each section of envelope would need to be manually zeroed.

You're also right about 200Hz being way too low. I'm now getting more success with a band centred around 1600Hz and 3.0 oct wide:



atom12 - That's a great idea about using a hybrid of dropping both the vocal frequency and the overall volume, so that's what I'm trying now. I copied and pasted my volume envelope points into my EQ automation envelope. I'm thinking of setting all the EQ ducks to -6dB (maybe more) and then doing the fine control just with the overall volume, otherwise it's a bit much to control over a 2-hour project. Here's what a bit of my timeline looks like with both the EQ and volume envelopes set to approx -6.0dB. My music fade-downs are 2 secs long and the fade-ups are 1 sec long.

Nick, it may not be possible to do this in Vegas, but in other apps like Sonar you can rig the side chain to apply an eq to another track, so that, for example, a -6 db cut is gradually applied at 1.2Khz when the triggering signal crosses the threshold. This should "clarify" your VO narration while leaving the music sounding basically unchanged--at least in theory. :)

Apologies again if you're already tried out this approach...

Andy, It might be possible to set that up in Vegas, but I don't know how to do it other than fudging it by recording an automation envelope with the Blue Cat Digital Peak Meter, inverting it, and pasting the points into another automation envelope that controls an EQ or multi-band dynamics plugin. I don't know how to do it "live" and would be very interested if anyone knows how.

A problem with the sidechain approach is that, as far as I know, it can't anticipate, so you can't have the EQ or volume ducking start gradually in advance of the voiceover, and so the result is less smooth.

I use combination of both also. I find the frequency duck envelope by itself is distracting, and yes the non-attenuated low end could hit broadcast compressors too hard..
Unfortunately VP does not have 'normal' side-chain capabilities, and for some reason, the side-chain/ducking comps just don't sound good... in my opinion anyway.

This is totally obvious but just in case you overlooked it, you can also just statically EQ the music track and your vocal track. The idea would be to look for the formants in your voice, boost them a bit, and cut the same in the music/ambient track--not as a sidechained effect or automated in any way, but for the duration of the entire video.

Then there's no pumping whatsoever. This is really just a simplified version of what goes on in multitrack top 40 mixing, where instruments are edited down to just the bare minimum needed to identify them so that there's room enough for everything in the mix.

Probably--probably--no one would ever notice the missing frequencies in the music, provided there are enough other sonic elements like voice and background ambiance to round out the mix. Anyhow...that's about all that's left in my idea bag. Good luck!

I had thought of that Andy, but I'm kind of nervous of spoiling the music. I don't trust my ears enough!

Hi,

Could anbody explain term "side-chain compressor" - can use normal compressor (because I have learnt it from my guitar processor), but can't understand "side-chain" meaning. Nothing ensue from plain translation ... is it connected with mid-side processing ?

Best regards,