Q It Up: What approach do you take to mixing a project? What elements to you deal with first? Do you lay out voice tracks and set VO levels, then adjust levels of the other elements around the VO? Do you use limiting or compression on individual tracks -- VO, music or SFX -- to minimize time spent tweaking levels? Do you use EQ to adjust the presence of elements in the mix? Is dynamic range your goal, or do you compress the stereo master to maximize “loudness” and minimize the tweaking? Describe your approach and the techniques that get you to your ultimate goal.
Don Elliot, Levine/Schwab Broadcasting, Hollywood, CA: Pretty simple rule of thumb for promos is to use as much compression on the vocal without audibly hearing the "effect" and let the music retain as much of the dynamics as possible.
Try to create an acoustic hole in the track… that is, by equalizing a little room in the speech range, the track will sit there really fat and full in the mix without wiping out intelligibility of the vocal. Feel free to go back and tweak any sudden impact of the instrumental track that could wipe out words because of the dynamics of the track that you left intact.
I try to make the vocal track have the same EQ as the music track in order to give an "apparent sameness of source". Since everybody knows that it's a guy sitting in a little room talking over a track in the first place like a vocal karaoke, if you will… it most certainly doesn't have to sound that way. It's up to us to manipulate the theater of the mind to also include some audio tricks like these in the smoke and mirrors part of our wizardry and audio sorcery.
One way to do this is to add a little Aphex Aural Exciter to the vocal track, if you sweep it judiciously through the voice track range until you feel it being the center of the spotlight. It's important to do some fast punching in and out for an "A-B" comparison. Basically, if you can consciously hear the effect, it's too much.
If you were doing this for a station that you actually work at, you will develop an ear reference to Air, so that you are aiding and not becoming a victim of the station air chain's processing. You can adjust your own production sound to the resulting air sound to develop that coveted "ear reference". If it's going out of house to another station, you don't have that advantage. By the way, if all of your efforts go down the drain from too much of the station processing, it's probably time to revisit that weakest link in the chain, because at that point of realization, it's probably destroying more product than just your promo. Nothing's worse than struggling to succeed in spite of yourself. In other words, you have to have the engineering department on the same page with you. This is sometimes difficult because transmitter guys probably don't have your ear, since they think more in terms of "RF". Figure out a way to communicate with them that doesn't make them insecure with them thinking you want to have a piece of their turf. They have to understand that you're going to make them look good and vice versa. Then you have won the battle.
The hardest thing I ever had to accomplish in the radio business is to convince coworkers that Radio is not a zero-sum game. Perhaps they always thought that it was. So as the saying goes, "be the change you want to see" on your staff.
And that's how to fix politix in the mix!
Mix responsibly and DO try this at home!
Ty Ford, www.tyford.com: It depends on the project, but if the words are important – as in ad copy or lyrics to a song – I'll sort of get them where I want them and pull up the rest of the tracks up around them. Then I'll set my master gain reduction dynamic range plugins in the Stereo Two Mix (master fader).
I may use EQ, comp/lim on each track to get tracks to fit.
If the VO starts to get lost, I may use EQ, or comp/lim (or a little of both) to bring it out. I may use subtractive EQ as well as additive EQ.
Reverb usually pushes the source away from the ear. If have used reverb or other smeary effect, I may reduce them to let the VO dry up and poke though more.
I normally use two master reverbs, a fast and a slow with sends and returns.
When everything’s sounds right, I'll go back to the master comp/limiters and make any final adjustments.
Jay Helmus, Audio Engineer, www.jayhelmus.com: I'm sure everyone will have a vastly different approach to this...
Personally for me, I don't use shells often, but I wouldn't say never. I typically make stuff from scratch. Not because it's better, just because that's the way I've always done it. I find it gives me more creative freedom, but that's just me. Here is my own personal process for a typical radio promo or imaging piece:
1. Import VO tracks and edit together a full pass I like. Sometimes I'll keep alternate takes for later, but I try to avoid this if possible. I find committing to takes early helps the flow.
2. Process VO to get my "generic natural" setting. This is where I'll EQ, compress, limit, perhaps add reverb etc... I'm not looking for a hugely affected or filtered VO here (yet). I just want a nice present (and natural) generic setting. Some distortion is fine. Not too loud. I need some headroom.
3. Import SFX and music beds. Usually I import more than I need so it's all in the session once I get going. I don't like stopping the flow to look for zippy zappies or sfx if I can avoid it.
4. This is the big stage where I start to "construct" the promo. I usually put in the music and sfx and shape it around the VO. This usually means bringing stuff in and turning it down so it doesn't overpower the VO etc... adjust timing, move things around, adjust volume, add zips and zaps, EQ stuff if necessary etc. I will also use this stage to "effect" the VO where needed to create texture and movement. I'll create high-pass tracks, distorted tracks, delayed and panning tracks etc. Sometimes I create these "effected" tracks from scratch, or from the original generic setting. Sometimes I use tracks with settings that I've saved throughout the years. All depends. But this the stage where most of the 'mixing' and sound design happens.
5. Personally, I try to get the mix right before anything on the master -- or at least, as close as I can get it to sounding great.
6. Mastering: I'll add subtle compression, limiting, sometimes EQ, saturation, widening, and general "polishing" effects on the master buss for the loud and punchy final sound. Sometimes (often) this will affect the balance of the mix slightly, which is fine. I just usually go back one final time and quickly adjust any music or sfx tracks that have become too loud. Doesn't take long.
Then I send it to my PD where he rips it apart and tells me how awful it is!! :) hahaha, kidding!!
Andrew Frame, BAFSoundWorks: Sacrifices seem to help. Depending on the job, we may use a 'possum, or
chicken, or the neighbor's house.
Outside of that, everything gets adjustment before going into the multitrack. I have so much work that recycles, it's a time saver down the line. Whether it's last year's re-use of a sale spot with a new price-point, or last week's copy revision, having all the parts pre-processed means the repeated mixes will sound consistent.
Voice work is cleaned up. De-breathed, normalized, all the junk cleaned out, and timing adjusted to get the desired feel. Dramatically bad EQ on the raw read is cleaned up - typically a massive boost around 100Hz has to be dealt with. Music and effects levels get evened out. No compression at this point. Just cleaned up.
Then it's all imported into multitrack. There, any track-by-track companding and overall levels are tuned, and any timing corrections made. Each element gets a discrete track. Mono voice, stereo effect voice, sound effects, bangs and lasers, music - all separate.
Then the final stereo mix gets a wee EQ tweak (like 1dB top and bottom), an overall light companding (1.5:1 max).
The goal is to find the balance between dynamic range and a quality sound. Radio and TV stations apply crushing amounts of compression, so I specifically do not do that in the mix I send. I let the station's processing do the heavy work. I strive to deliver a clean, loud product that will sound good with or without post-processing (like on a streaming-only station like Pandora).
After all that, we usually deep-fry and eat the chicken. Can't let a good sacrifice go to waste.
Al Peterson (Talk Radio), Radio America Network, Arlington VA: Fortunately for me I'm not in the "We Play More Music" battle with the guys crosstown, so I don't sweat high-energy mixes. Doing long-form spoken-word Talk Radio projects, my approach is more towards intelligibility and clarity rather than sizzle.
Purity of voice comes first. Hosts are live in our studios; or call in via Skype or SourceConnect; or check in with a Comrex Access. Assuming good gear and correct mic technique at the sending end, I'll gently compress first then noodle the EQ to put in a little bottom and fix the artificial highs added at their end (inexpensive condenser mic brands like MXL internally over-EQ the highs to sound bright). Too much compression destroys dynamic range -- you want to hear the host get close up to share an intimate detail, then lean back and wail when he/she needs to. If their technique goes off the rails, I'll stop them and redo the portion they booted.
Telephone guests can require more attention than the telephone hybrid provides. Some long-distance calls come in with a metallic tonality or a steady whine around 8 kHz, so they get recorded on a separate track, away from the host voice, and I'll apply a noise gate and a brickwall EQ that slopes out at 4.5 kHz. I also keep a DTMF filter close by to erase the touch-tone digits a guest will invariably press during a phoner. Another advantage to separate tracks for host and phone guest is that if they step on each other, I can time-slip the two tracks so the conversation sounds better and not intrusive.
I also take a look at the voice recording with the Spectral view. If there is a stripe up around 15.7 kHz, it means there is an old-style CRT television set on somewhere and the mic is picking it up. It won't be heard on AM or FM radio, but it shouldn't be there anyway so I usually notch it out with a para EQ.
When the theme & bumper music goes in, complementary EQ is used to dial in a broad notch at voice frequencies, so I can ride the music louder and we can still hear the host. I prefer that over sidechain ducking which works, but can be heard as 'breathing' to a disturbing degree on FM talk stations.
On mixdown, everything gets a hint of peak limiting and is normalized to -12 dBFS. While that sounds low to many readers, mastering a voice recording all the way up to -0.1 dBFS adds nothing and will actually cause it to be rejected by some of the pro houses and satellite companies I work with. Processing a talk radio waveform to be so flat & level you can use it to hang a picture is counterproductive.
If there is anywhere I'll allow more-than-normal processing, it's on the podcast mixouts of our shows. While stations want a drier, uncrushed sound to add their own sonic signature to a program, punch and clarity are important when your audience is using earbuds, so additional squish gives a file a "radio sound" that many listeners expect.
Denzil Lacey, Zava Media: I have a number of templates set up for different projects – Branded Intros, Sweepers, Songs, Promos all have different settings and functions. It's a very busy job, so having a solid start makes all the difference. I want the station voice to have consistency, so having that laid out and templated saves a lot of time - it only changes if the voice is recorded in a different location.
I personally feel that Dynamic Range is hugely important - especially when beatmixing and working with pre-mastered audio - do you really want to compress that again in your DAW? It can make nasty sounds. In all my projects the audio will not exceed -3db gain reduction on the master channel. The Optimod or On-Air processing will always do a much harsher job on leveling everything out - so starting low is always key.
Ethan White, Commercial Producer, Vista Ideas Group - Vista Radio: I usually work with the VO first, but with very immersive spots that have lots of SFX and changes, I'll put the VO aside and make sure the I have the right sounds in my session so that I'm not going back and forth too much.
I'll adjust VO levels while editing, and lay out each VO part accordingly, then after that I'll deal with other work elements chronologically. I use EQs on beds regularly to make room for the 1khz-4khz range on the voice, but it’s not always necessary. I try to keep SFX as they are because a lot of them are already processed a little. I’ll adjust levels, but not compress them unless I’m using heavy compression as an effect. Dynamics are usually not the goal sadly, but when I get the chance I take advantage of producing with as much dynamic range as I can. These are usually the more scenario or story-like spots. I do use a multiband compressor on my master bus, but have started using limiters less. After bouncing a spot, I'll normalise to -1db.
My EQ settings: I process each VO with a parametric EQ, 6-8db's of compression, and a De-esser. I roll off everything below 80-100hz at 12db/Octave, a pretty scooped EQ between 200hz and 700hz, and pinching out any harsh 1khz-2khz frequencies with a tight notch. With female VO's I might bump up everything from about 10khz up just a little to balance out all the cutting.
Hope that sufficiently answers all the questions! Love this magazine!
Scott Shafer, Regional Production Director, Waco/Temple-Killeen//Bryan-College Station, iHeartMedia: What approach do you take to mixing a project? I want a good clean overall sound. I don't think you should have to strain to hear anything in the mix.
What elements do you deal with first? My first thing is to lay down the VO. I clean it up and then I have my go-to scripts to process the VO. It normally consists of a little EQ-Compression-Reverb-Normalizing.
Do you lay out voice tracks and set VO levels, then adjust levels of the other elements around the VO? Yes, since that's the main focus. And thank goodness for fader automation; it's nice when levels are really up & down.
Do you use limiting or compression on individual tracks -- VO, music or SFX -- to minimize time spent tweaking levels? It depends on the track. If it needs something, I'll tweak it. Let's say you are producing a club spot, and you want the music pumping but don't want to drown the VO. You can use a side-chain compressor to automatically control the music level while the VO is active for example. Most of the time it's pretty good on its own, but I have compressed tracks, and it was too much, it lost its dynamics. In this case the spot sounded better with its dynamic range full & open instead of narrow and crunched down. I like to A-B the tracks to compare how they sound.
Do you use EQ to adjust the presence of elements in the mix? I will, but if the levels are good to start with, I leave it alone. If it ain't broke...you know the rest. There is something to be said however about subtractive EQ. You don't always have to boost something to make it pop in a mix.
Is dynamic range your goal, or do you compress the stereo master to maximize “loudness” and minimize the tweaking? It might sound crazy, but loudness is not always a good thing. I'll get pre-produced spots sent to the station that are grainy and distorted. I will occasionally put a limiter on the master track to boost the overall loudness.
Describe your approach and the techniques that get you to your ultimate goal. I would say if I was just mixing music, dynamic range would be more of a focal point. But since I'm dealing with Music, SFX, VO, etc., I want the message to be clear, and the additional elements to color or enhance that message as much as possible. That said, sometimes a dry VO is a very powerful thing. With the production that comes across my desk, I read over the copy and see what approach to take. It really depends on the spot. And then I go to work, or assign it to someone I think would be a better fit. Afterwards, if I'm happy with it, great. If not, I'll go to a studio with different monitors to listen or get another set of ears. And sometimes just walking away and coming back a few minutes later will make the mix sound different because your ears get used to hearing certain sounds and frequencies. It's like shopping for cologne, sometimes you have to smell coffee beans to neutralize your sense of smell. In this case, neutralize your ears!
Michael Shishido, 94.7 KUMU Programming, Dir. Creative Services, KUMU / KDDB / KPOI / KQMQ, Honolulu, HI: I'm a bit of a minimalist when it comes to radio production. I try not to overthink things. I always run voice work through at least two presets in Adobe Audition. In "Tube Modeled Compressor" there's a setting called "Vocal Attacker." I use that to even out the levels of the voice work. Then I apply hard limiting to -3db, in order to maximize the loudness of the voice. That's it for 90% of the spots I produce. I almost never alter music. If I do, it's to make a passage louder. If it's a concert-type spot, I might run that final through one of the mastering presets in Audition to brighten everything. I've found that most production beyond that is lost on clients, listeners, and even some radio people listening in-house. Keeping things simple also eases the work flow and makes for an efficient process.
DJ Mike, Chris-Mar Studios: The first thing I do when mixing any project is to have a vision (or sound) on how you are hearing the mix. To start, if it's a promo/spot, I look for the right music bed. I then look for any sfx or sound bits that I may need. I import them into my DAW. I write out the script for the promo/spot and vo. Once the vo is completed, I import into the DAW. I go over each level for music bed, sfx or sound bits. I look at the vo track(s) make sure the levels are correct. Okay now comes the simple part. On the tracks mixer I always use track 1 for vo and 2, 3 & 4 for music beds etc. I adjust levels as need to. I listen with my eyes closed to hear the promo/spot. Once everything sounds good I mix everything to a final mix. Everyone has their own way of mixing. This is a way I came up with to make my mixing sound the best.
Kyle Whitford: Voice first. Music, then SFX.
Mixing requires a touch of un-sanity. Try vastly different versions. Various combinations of compression and EQ.
Make a rough version, take it out to the car and play in on the car speakers standing outside on the sidewalk to see if the voice cuts through enough but the music still rides up nicely. Are the SFX right?
Play it down a phone line and listen for that mix.
Make models of other spots you know sound good and compare yours back to back.
Play it on a boom box as your play softball with your kids, so you can hear it like real people do.
Get a timer and have it come on as you wake up in the morning so you can be surprised by the sound like in actual day to day living. Are you too close? Wait a few days, do it again.
Play it on the big studio speakers where nobody will ever be listening just to hear God's version.
Listen through a Studer reel speaker for a treble version and make adjustments.
Email it to yourself and listen on your phone, an iPad, and a laptop.
But most important - Go out in real life and hear your spot coming from a taxi cab radio, or a TV in a bar, that's the real test. Then go back and mix it again.
The mix is never finished.
I just try to think what they do. Not much compression at all. Maybe a bit of EQ, and I try to sweeten the room sound due to old plaster and interesting acoustics. Mine will never cut a record, it’s mostly for demo, cold tracks to be used elsewhere, narration, and similar. I don't have or want a drum booth. So it's about clean more than shaking the surround sound system.
I usually just normalize, add some top, and roll of some bass. I have listened lately to a few how to sound like a voiceover with Adobe Audition, and man the world is in sad shape.
There are plenty of examples on YouTube of what not to do. Crank up compression, bass, and sound like a 1950’s announcer. I thought we were done with that. Maybe it's just me that is. My thinking being that if I win jobs by sounding artificial, living that life isn't something I want to maintain.
I really believe the dark side if you will is going for a pumped up artificial tone and bag of tricks on the channel is not what I want to set my best foot forward with.
I just try to do my best work and the engineering should reflect that.
■
Thanks to all who responded. Your input is valuable and appreciated. If you have a question you’d like to see posed to the RAP Q It Up panel, email it to