Once again, lots of heavy-hitters weighing in this month. Thanks to each of you for taking the time from a busy day to join in our discussion!
Q It Up: What processing do you use when recording voice tracks for your more common tasks? What pre-amp are you using? Do you have an outboard processor in line for preliminary processing before bringing the voice track up in your DAW for final processing? Which external processor are you using, and what are settings you like to use for most VO sessions? Once you get a voice track into your DAW, what plugs do you use most often to perform any final processing on voice tracks? When EQ-ing voice tracks, what are you trying to achieve? When do you know it’s EQ’d “right”? What is your goal with compression on voice tracks, and how do you achieve those goals? Please add any other thoughts you may have regarding the subject of processing the voice.
Vaughan Jones [vaughanandjoy[at]aapt.net.au], Prime Radio Network, Maroochydore, Queensland, Australia: What pre-amp are you using? Yamaha O2R in-built.
Do you have an outboard processor in line for preliminary processing before bringing the voice track up in your DAW for final processing? No. The only thing processing my v/o on the way in as a pop filter, and sometimes I even take that away if I trust the voice.
Which external processor are you using, and what are settings you like to use for most VO sessions? Again I use the O2R for external compression, but I have a number of presets that I can swap between, depending on the degree of outboard processing I require, if any. Some have both Compression and EQ and some have just one or neither in varying degrees of severity.
Once you get a voice track into your DAW, what plugs do you use most often to perform any final processing on voice tracks? L1 Limiter, Waves Ultramaximiser, Req, Filterbank, Dverb, C4.
When EQ-ing voice tracks, what are you trying to achieve? Cut through and/or variation
When do you know it’s EQ’d “right”? When it cuts through the bed clearly, without sitting too proudly on top. Also, in many cases, I am looking for a variation to give the message some momentum, when dialogue is heavy.
What is your goal with compression on voice tracks, and how do you achieve those goals? I like to make my tracks as fat as possible, whilst preserving as much dynamic range as I can. It’s a really visual thing for me, I know what it should look like; I know what looks too thin and empty; and I know when it’s looks too much like a G.I.’s flat top haircut.
Please add any other thoughts you may have regarding the subject of processing the voice. I used to feel like I was really processor heavy and I was conscious of it. Now I feel like I process less than many... but along the way I have learnt that there is no right or wrong answer... I don’t worry what people think of my processing any more, and I don’t judge others on theirs. Also, a lot of my decisions depend on what’s in the broadcast chain. What works for me at one station, won’t at another. My message to new starters is: “don’t be afraid to get experimental - it’s a wide envelope and you’ll know when you’ve gone too far.”
Andrew Frame [andrew[at]bafsoundworks.com], BAFSoundWorks, Lehigh Acres, Flordia: Are you insane man? Those are trade secrets! Would you ask Kentucky Fried Chicken for the 11 herbs and spices? Would you ask Coca-Cola for the ingredient ratios? Would you... wait, yes we all would. Okay.
We keep it all very simple. When we record voice tracks, there is zero processing. An AKG-C300B feeds a JoeMeek ThreeQ. That goes to an A/D box that USB2’s into the workstation. There are no outboard pieces of any kind. The JoeMeek has an analog optical compressor, but we only use the pre-amp and phantom power stages of the unit.
The reason is, we want a pure signal to begin with in case we have to go back later and do any pickups - which, in this business is a given. I receive voice tracks from about four dozen talent all over the globe, and everyone has a different mic-preamp-preprocessing setup, so when I order a voiceover, I stipulate “no processing”. Often, those tracks are combined to make an ensemble commercial, so it’s important that we receive audio that is clean and unprocessed.
There are a myriad of different EQ curves that come in, and detuning them is hard enough. But if someone had compressed their audio in advance, it would only add to the problem. And, some people work in broadcast stations where their compressors are set to mirror the air studio (for voicetracking), meaning the production room compressors are set to work hard. This kind of setting is completely unacceptable for the work we do. Fortunately, all my radio people know where the “processor bypass” button is located.
Once the v/o is in the DAW, we’ll edit to de-breath and clean out all the boogers and bad takes, tighten spacing and insure the read will fit the time on the script. (Talent ships us raw audio in most cases.) At this stage it’s mostly selective limiting and normalizing to give a clean, but high-amplitude signal. A 32-bit WAV of the file is then saved as a working file, and the original talent audio is backed up as a safety/master. Then in the multi-track, we’ll use real-time combinations of limiting, compression and expansion, EQ or any other tricks on the working file to give it our signature sound.
If we’re sweetening a voiceover that is going to television post-production (which we do more than anything else), we then export the audio as a 48kHz, 16-bit WAV or AIF file, and ship it to our customer. We don’t use MP3, MP4, AAC, or any other kind of lossy files.
If we’re producing the audio for radio or television, we’ll then start laying in the rest of the audio elements to build the production.
I have found the native plugins on Cool Edit 2.1 to be very, very good if you know how to use them. I have a Waves Gold bundle installed as well. The not-so-secret is to use them lightly. EQ is rarely ever more than 3db up or down (mostly down, we believe in subtractive equalization). Mostly, EQ is 1 or 2db at a select frequency, with the parametric’s “Q” set modestly, around 0.5.
Compression is no more than 2:1, and that doesn’t kick in until the signal is at -6dB. It slopes down until -40, then I apply 9:1 expansion to create a fast and clean noise gate. We work to keep as much dynamic range in the audio as possible. The radio and TV station processing chain is going to crush it with compression, so we work to give them a loud, clean final mix. We use compression as a way to soft-limit and bring out the fullness of the mix.
When it comes to EQ and dynamics, we go for a tight low end with no booming, intelligibility in the midrange while avoiding the “AM radio” sound, and a crisp top end, with no sibilance issues. The compressed voiceover should have all the same qualities. Light compression gives you a pleasant smoothness to the entire bandwidth of the audio.
Ryan Drean [ryandrean[at]gmail.com], www.RyanOnTheRadio.com, www.ryandrean.com: When I record my own VO, I have found, through recommendations from friends, the Waves AudioTrack Plugin to be best. I run it live as I record (with a bus) so the processing happens on the fly. Thus I don’t have to ever waste time “bouncing down” in Pro Tools. I use either a Sennheiser 416 or a Neumann TLM 103 depending on which studio I am in. The AudioTrack plug gives good gating, some nice subtle EQ and a bit of compression, but I really use it for the gate. I run the Sennheiser through an Aphex 230 which again I mainly use for very subtle Comp/EQ/Gate.
Once I am in the production world, I am using mainly 3 plugins on all the voice tracks. Start with the C1 Gate then the REQ 6 Band then the L1 Ultramaximizer+. This chain is pretty simple but gives you a lot of power and dynamic range when changing the VO. You can make it REALLY crunchy and hot or just give subtle coloration/roll-off, etc.
FYI - my template starts with 6 VO tracks. Four of them are bussed to this chain I described on an Aux 1. Two others start pushed right and left with an extra Filter for when I want the higher filtered stereo spread effect to be quickly applied. I produce many different voices and sadly, there is no miracle setting I can just use with everyone. I am always making adjustments on the L1 and the REQ. One thing I like, which always seems to make the VO swim better in the mixes, is laying the L3 Multi on the master track. Try starting with the “Hi Res CD Master” preset and go from there.
One more thing... I start with the bus mentioned above sent to Aux 1 with a very filtered EQ. But then I can change the bus quickly on one or more VO tracks, send them to Aux 2 and quickly give them a more full sound. i.e. I have voice guy going through crunchy Aux 1 but then add some listener audio that I want to sound more natural so I send those tracks to Aux 2 which is already set up which a more natural EQ setting. Just saves time when you can set everything up in your template, ONCE. Hope I helped!
Ric Gonzalez [Ric.Gonzalez[at]CoxRadio.com], Cox Radio, San Antonio, Texas: We use AKG 414B’s in all six prod studios and voice booths. These microphones run through Focusrite VoiceMaster Pro processors. Additional processing depends on the individual producers and what they prefer in adobe 3.0. I don’t usually add more processing other than normalizing. The AKG414B’s are incredible microphones and paired with a good processor you don’t really need to add much more. Exceptions would be special fx like phasing, filtering, etc. But Adobe 3.0 comes with a folder full of goodies for this.
Frank Scales [fscales[at]kloveair1.com], K-Love & Air 1 Radio Networks, Rocklin, California: I use an Avalon 737 Pre (usually flat EQ, 3:1 compression, Neumann TLM103) into Pro Tools. I usually post with Waves L1, L2 or L3 (try not to overdo the threshold) and EQ 7 with a slight boost on the high end. After a sub mix (if I’m sending out dry tracks), I’ll normalize to 0.
Johnny George [vo[at]johnnygeorge.com], Johnny George Communications Inc.: In my DigiStudio, I work on Pro Tools with very little processing. I’m using the common Digirack Compressor/Limiter & the Expander/Gate plug-ins only.
Gate setting: Range –80.db / Attack 15.1 us / Hold 63.2 ms / Ratio 3.0:1 / Release 29.3 ms / Threshold –41.2 db.
Clean Limiter: Knee 8.0 db / Attack 300.0 us / Gain 3.0 db / Ratio 20.0:1 / Release 120.0 ms / Threshold –5.0 db.
Outside of my Bock Audio 195 microphone, I unitize a Presonus TUBEPre set to about 40% Drive and 30% Gain.
That’s it. No extra toys or EQ. I took them all out of my chain long ago when I realized as a VO Actor, most all my clients who were doing post-production, wanted clean audio without a bunch of tricks. My mic goes directly into my pre-amp and then directly into my Pro Tools input set at 50%. The settings above are relatively flat in comparison to what I was doing previously when I produced all the production for a finished project.
On audition material only, I may compress the VO once to beef it up a bit & normalize it at about 70% since I’m competing with other talents who are using radio production studios to send in their tracks. Many of them have inline broadcast “beef” already set by their engineers. I’ve had several agents tell me when I first started sending in “clean” tracks, they sounded great and clean, but were wimpy in volume and girth. Once I compressed them, they felt they were more competitive.
Heikki Wichmann [Heikki.Wichmann[at]nrj.fi], NRJ, Helsinki, Finland: When I’m recording I use comp/limit/de-esser on my mic-preamp. I use SPL Channel One as my preliminary processor when I’m recording to DAW. I use Comp/Limit, low-cut and de-esser and some EQ so it means the mic-sound is being processed almost ready in SPL unit.
In my DAW I use EQ and Limiter, Waves plug-ins on voice-tracks. I try to do as natural sound as possible. Well, very limited dynamics anyway. I use EQ to clean “pops” and make the sound as clear as possible. Sometimes I use EQ as an effect to give some words little more punch.
Jeff Berlin [jberlin[at]jberlin.com], www.jberlin.com: I use the plug in “Channelstrip” from Metric Halo to process voice tracks in Pro Tools. For dense multilayered production, I like to filter then heavily compress voices so they easily “swim” in the mix – letting the music and effects rock the foreground, while voices remain easily understood from the back.
To know when the mix is “right” requires monitors you can trust – for me those are Yamaha NS-5’s. Before I got those, I used to turn around and drive back to the station when I heard a promo in the car that I thought I mixed properly – only to be unable to understand the voice, or have the voice drown out music. Once I got the NS-5’s – what I heard in the studio was what I heard in the car, at home, in stores. I always want the VO to be loud enough to be intelligible, but no louder – a delicate trick given the variable listening environments radio reaches. Processing is a powerful tool for accomplishing that.
When I provide voice tracks for radio stations, I usually submit 5 or more different settings – giving the producer a range of choices from distorted & filtered, to big and beefy, and to relatively clean. I create these settings using a mix of outboard gear and Pro Tools plug-ins, and roll on them in real time as I record. No further processing is really needed on the end producer’s part, but that’s always their discretion.
I’ll get out of the way now, ‘but I could fill up the entire page on this subject :)
Wawro, Wally [wwawro[at]wfaa.com]: WFAA-TV Creative Services, Dallas, Texas: At least 95 percent of the voice work I handle comes in via mp3 or ISDN. The announce booth has become a lonely place! I run PT-HD2 and all microphones and ISDN audio are routed through the Digidesign PRE. From there I have a set of Summit TLA-50 tube levelers for some light compression before the audio hits the converters for PT. Both of my regular VO guys, Doc Morgan and Jim Pratt, have clean, good sounding mic chains, so I really have to do very little processing once I get the tracks into the system. About the only EQ I use is the standard Digi 7-band plug-in. I’ll dip the audio around 5 db at 200 Hz to get rid of the “chesty-ness.” Depending on music, effects and strength of the read, I’ll add a db or two at around 2.5k for presence. I will De-Ess if necessary.
I’ll bring extra compression into play for female voices; my ears say they need it. A female voice may also get some additional low frequency treatment via the high pass filter on the Digi EQ.
MP3 files vary since those primarily come from voices the sales producers hire for local spots. That may need a lot of tweaking! The same goes for field audio from the news photogs, that can vary all over the place.
I rough mix everything and once I’m happy with that I put the finishing touches to the mix via the Izotope Ozone 4 plug-in. I’ve tweaked one of their presets to give me a sound that’s large and punchy but retains some dynamic range. The result is a “signature” sound that competes nicely with the over processed stuff found on a lot of national and regional agency production.
Oh, if anyone one cares, I have a Heil PR-40 in the booth. I’m also evaluating an Audio Technica 4050 which is part of my personal mic collection. I monitor through the venerable Electro-Voice Sentry 100’s, which is the station standard.
Mitch Todd [Mitch.Todd[at]siriusxm.com], Sirius/XM Satellite Radio, New York, NY: I prefer to print to tape or a DAW as cleanly as possible, but with a little “safety limiting” going in. Usually NO EQ adjustments on the low end, a little midrange tweaking, “opening up” the upper frequencies and no more than a few db of compression going in.
Though you didn’t ask, my two mics of choice for virtually all broadcast VOs are Neumann’s TLM-103 (for the large diaphragm condenser at a very reasonable cost) and the Sennheiser MKH-70 (for a bit more $)! The TLM-103 is very versatile as most people can sound pretty good on it, and it’s “presence band” emphasis is very nice. The shotgun is great for that mid-range “cut” and minimizing background noise.
What’s up front is as personal a choice as the mic itself. I prefer “channel strips”, all-in-one tube preamps with EQ & processing. On a budget, I think the DBX 376 Tube Channel Strip is a great value (at around $500). Modify it with a higher quality tube, and you get a LOT of pretty clean “bang for your buck”. Of course you can go up from there to the Manley Vox Box or Avalon 737SP channel strip being the smoothest and cleanest my crispy old radio ears have heard.
But ideally, it should be tailored to the talent if you’re buying a set-up just for one individual’s use. With so many options out there now from Blue, Royer ribbons, etc… the options are endless. This old analog fogey has also been pretty impressed with a few of the USB mics popping into the scene. As for the box’s settings, generally a 2.5 to 3 to 1 ratio only hitting about 3 to 4 db of compression is usually the safest way to proceed… you can always add more later. EQ to taste. In post, it’s all relative to the context of what you’re producing. That’s why I don’t like to use a formulaic template once your VO is tracked… the subtle EQ and compression settings (or not so subtle) plus” band-aid” plug-ins like de-essing, must be done and adjusted accordingly during the production itself, not in a “bubble”.
But I do recommend one thing when auditioning mics and boxes (or recording and mixing audio in general): Close your eyes more often and always let your EARS be the judge! Some of the ugliest boxes and plug-ins sound great… and vice versa. Don’t let your eyes lead your ears astray!