As the tools for music production become increasingly more sophisticated and affordable, the home studio continues to evolve as a facility where we can produce income-generating projects of all types. We now live in a world where record projects, commercials, and music scoring or sound-effects design for TV and film are commonly produced off-site-a practice that was inconceivable ten years ago.
Not to be left behind, the radio-broadcast community has jumped on the bandwagon. Many of the same tools that musicians and recording engineers use for music production are also used in the construction of radio material. The difference with producing material expressly for radio broadcast (such as in-house station IDs, commercial spots, and station promos) is not the equipment but the attitude. These spots are usually relatively short, tightly edited, loaded with snippets of music, and often heavy on sound effects, either to create an aural environment (say, making the dialog sound as though it were recorded on a downtown street corner) or to generate a sonically aggressive promo.
Most important, you rarely have more than a minute to get the message across, so the announcement takes precedence over all else. Ron Shapiro, imaging/creative services director at KCMG Mega 100 FM in Los Angeles, puts it this way: "Brevity is best. If a promo goes beyond 60 seconds, you're likely to lose the listener. Most of the producers I know generally attempt to keep these promos between 30 and 40 seconds. For the most part, people perceive the promo as just another commercial and will tune out if it's too long."
SEE SPOT RUNBefore we proceed, let's define some terms. Although we often use the word spot to indicate any short piece of material, a true spot is a paid advertisement (a commercial), and it is always sold as a definite, time-limited unit. Most commercials in radio last 60 seconds, though sometimes you'll hear a 30-second or 10-second spot.
On the other hand, a promo is a piece of random length whose sole purpose is to promote the station or hype an event that the station is sponsoring. An ID is a fairly short piece that identifies the station in one way or another. For this article, we'll focus primarily on promos and IDs, and we'll use the term spot loosely.
TIME OUTI couldn't help but wonder whether the 15-, 30-, and 60-second promos so common to TV exist in the radio world. As it turns out, such promos don't exist in a station that is run "live." What exactly does "live" mean?
An automated station that takes a satellite feed and other network cues will probably need to use 15-, 30-, and 60-second promos; without them, the station is likely to have dead air (unwanted silence) following the commercials until it rejoins the network. (It's also important to remember that if a commercial goes longer than the 60 seconds that the network has allotted, the network feed will more than likely break in during the end of the advertisement.)
As a result, time tends to be far more crucial in an automated environment than at a live station like Mega 100, which uses live DJs all the time. In a live environment, the DJ isn't forced to hit a particular cue at exactly the top of the hour.
Shapiro emphasizes, "For the most part, radio is much less rigid than TV. In radio, we have a little more leeway because songs are of differing lengths and each DJ speaks differently. In a live situation, a 30- to 40-second length has proven itself to be a good time frame for promos."
When it comes to producing radio promos and ID spots, the announcement, or voice-over (VO), dominates every other production concern. If the listener can't understand what's being said, the entire spot has missed the point. That's not to say the music and sound effects don't matter, but these elements play far more of a supporting role than they do in most other areas of audio production.
Because radio is a 24-hour, seven-day-a-week medium, the volume of material that's needed to "fill the pipeline" is enormous. Hence, for the home-studio operator who has the right production chops, there can be huge rewards in radio work. With that said, let's find out what's involved.
MAKING TRACKSBefore we examine the production process, you first need to understand what makes up a radio spot. For those of us who grew up as musicians, perhaps the most apparent difference between music and radio production is the number of tracks that are found in a typical project.
For the musician or composer, if you have only eight audio tracks, you probably use a lot of MIDI tracks. For an audio-only project, most musicians prefer to have 16 to 24 audio tracks or more, and it's common to find 96-track production rooms. For radio spots, though, things are very different.
The radio promo usually consists of between eight and ten audio tracks at the multitrack stage of production, and rarely more than a dozen. Why is that? According to Shapiro, "A typical radio promo has two tracks for voices. This frequently takes the form of a primary VO and a secondary voice that responds, or otherwise interacts, with the principal voice. Music and effects usually consist of stereo pairs, each separated for better control over the mix. My multitrack environment frequently consists of primary and secondary VOs on tracks 1 and 2. Music bed number 1 goes on tracks 3 and 4, while music bed number 2 resides on tracks 5 and 6. Ambient crowd noise or sound effects will follow on tracks 7 and 8 and perhaps tracks 9 and 10." Figure 1 shows this track arrangement.
There are no hard-and-fast rules when it comes to radio production. If you find that you need a certain type of EQ or effect on a particular segment of a track, you will most likely find it easier to accomplish your goal if you isolate that section of material. With some simple cut-and-paste editing, you can move a segment of a particular track onto a dedicated track. In doing so, you immediately have the freedom to process that segment without the fear of affecting the remaining material.
LIBRARY RESOURCESIn the overwhelming majority of instances, radio producers rely on libraries to provide music and sound effects. Why not create your own? Simply put, the answer is lack of time. In order to feed that 24-hour, seven-day pipeline, it's infinitely more practical to use libraries. In the time that it takes you to compose, record, and mix a 30-second music bed, another radio producer will have finished several spots. According to Don Elliot, house voice for KFI AM and KOST FM in Los Angeles, "Radio is a very fast-paced business. The demands of a 24-hour, seven-day system force us to deliver program material in sizable quantities. Music and sound-effects libraries are the only realistic way to meet the demand."
One of the best-known music providers is a company called Brown Bag. It is considered the Rolls Royce of the radio-promo business, and its music is available as CD Red Book audio, so you usually play the music bed into your recording system. (Most multitrack editors don't let you rip audio from CD in the digital domain, and radio engineers rarely use 2-track editing programs, which often have this feature.) Other companies make their libraries available as WAV files, allowing you to import the music data directly into your digital audio workstation. Music and sound-effects libraries are also available on hard disk.
Because the music and sound-effects libraries are professionally mixed and mastered elsewhere, they generally occupy one or two stereo pairs of tracks. If composing is your passion, perhaps you should look into producing music libraries. But if bringing all the elements together and creating something that leaps out of the speaker and drives the point home in 30 seconds would be a welcome challenge, producing radio promos might be for you.
TOOLS OF THE TRADEDue to its random-access editing capability, the computer has become the principal force in radio production. Although you could do the job on a linear system (such as an MDM), a computer-based system provides far more visual feedback. Which multitrack editing software you use really doesn't matter; the important thing is to have the level of editing capability that is found only in a software editor.
You'll also need a good sound card or other computer audio interface. The important issue here is to select a system that effectively communicates with the other audio hardware you own. (EM has covered this subject repeatedly over the years. For a look at PC sound cards, check out "Playing the Slots" in the March 1998 issue. "Digital Pipelines," in the April 1999 issue, offers an overview of digital connectivity, including sound cards. Finally, EM's Personal Studio Buyer's Guide features a huge listing of digital audio workstations, including audio interfaces and sound cards.)
You will certainly want the ability to mix digitally to DAT, CD-R, or MiniDisc, so either S/PDIF or AES/EBU digital output capability will be a must. In many cases, a simple 2-channel digital I/O system will serve you well. Because source music and sound effects from libraries are already professionally mixed and mastered, you can play the program material into your system and slide it around to fit your needs later.
Sometimes you will need to sample a sound for use as an effect. In the radio world, the two most popular tools for this purpose are portable DAT and MiniDisc (MD) recorders. MD has experienced less than a wonderful reception in most sectors of the audio community, mostly because it employs ATRAC, a data-compression algorithm that sacrifices a calculated amount of the audio signal in order to record smaller files. However, the format has done quite well in radio broadcast, where the signal will be compressed several times anyway. (I'll discuss this in detail later.)
When Elliot was asked why he prefers to sample with MD as opposed to tape, he put it this way: "With tape, locating a specific point is far more cumbersome than it is with MD. When I want to check my work to ensure that I've captured the sound I'm after, the rewinding process is time-consuming. With random-access capability, MD provides a faster and easier way of getting things done."
THE VENERABLE VOIt's not uncommon for an announcer to serve as the "voice" of a radio network, giving that chain of radio stations a unique, identifiable sound. To circumvent the physical limitation of having to reside near any particular broadcast facility, many of the better-known announcers have their own home studios equipped with ISDN lines that enable them to "phone in" their part to the production facility. Since we all have to start somewhere, however, we'll continue to focus on getting the job done locally, with local talent.
It's important to recognize that no particular microphone is ideal for every recording job-or for every voice. Choosing a microphone that is best suited for recording voice-overs is largely a matter of personal preference. Of course, every effort should be made to find a microphone that is a good fit for the voice-over talent.
Due mainly to their low cost and durability, dynamic mics are more common than condensers. Most radio stations tend to lean toward the Electro-Voice RE20 and RE27, Sennheiser MD 421, and Shure SM7. However, you'll also find mics like the Neumann U 87 and U 89 and the Sennheiser MD 416 in more elaborate production facilities. That doesn't mean you can't use a mic from, say, AKG, Audio-Technica, or Rode; radio producers who read music-technology magazines sometimes choose these mics as well.
As is the case with every recording application, positioning a microphone requires considerable attention to detail. Close-proximity miking, in particular, requires careful positioning in order to minimize problems caused by sibilants and plosives. Sibilance-a hissing s sound-commonly occurs when recording dialog. Plosives are popping sounds caused by words that include the letter p and, to a lesser extent, t and d.
The more you perfect the art of close-proximity dialog recording, the less time you will spend editing the recorded announcement-and as you already know, time is in short supply when it comes to radio production. As is the case with so many aspects of recording, the old saying "Garbage in, garbage out" rings true here. So learning good mic technique is critical.
One of the most common techniques for close-miking dialog is to take a side-angle approach. By positioning the microphone from the side but very close to the speaker's mouth, you can significantly reduce the occurrence of plosive sounds. With this technique, the voice-over talent speaks across the front of the microphone as opposed to speaking directly into it. "If you're experiencing too much plosive and sibilant noise, go for the side-angle approach," recommends Shapiro. "With our station's E-V RE27 N/D, I've found the side approach works best."
If the microphone is positioned at the side, a pop filter might not be necessary. However, if you're miking the talent directly from the front, you'll need to place a pop filter between the speaker and the microphone. Pop filters are available from a number of sources, including Popless Voice Screens, Middle Atlantic Products (Popper Stopper), and Windtech Microphone & Windscreens. You also can build your own; see "DIY: Build the EM Pop Filter" in the October 1998 issue.
NAKED DIALOGAlthough the finished promo will most likely have music accompaniment, the VO is generally recorded without music cues because the music can become distracting to the talent as they read. It's not uncommon, however, to play the music for the talent beforehand and have them rehearse with the music to get a feel for how the dialog needs to fit.
But as with any creative endeavor, there are no rules. Some VO talent prefer to hear the music in their headphones, so it all boils down to a matter of personal taste. Shapiro notes, "I find that it's really not that crucial to hear the music bed when you're recording a promo. Obviously, if the talent needs to match the tempo of the music, then we'll use it."
How much dialog is used in relation to the rest of the promo spot? Good question. I doubt anyone has ever conducted any research specific to this point, but the general consensus appears to be that the dialog will constitute a good 80 percent or more of the actual promo.
Whatever you do, don't waste time getting the announcement started. More often than not, roughly half a second of music or sound effects will "establish" the promo prior to the actual dialog. Sometimes the dialog even begins right on the music or slightly ahead of it. Remember, the purpose of the promo is to inform, not to entertain-so forget about that four-bar intro.
TROUBLESHOOTINGAs mentioned earlier, proper microphone placement and good recording techniques can minimize your editing load, and this is an extremely important point. However, you'll still need to polish even the best of takes. Virtually all professional recording programs offer tools for dealing with plosives, sibilants, and other common artifacts. Learn to use them. Many of these algorithms do a wonderful job of cleaning up dialog. Also, numerous plug-ins for this purpose are available from third-party software developers.
Use de-essers to reduce the occurrence of sibilant sounds, and use a similar combination of compression and EQ-which can be very effective-for eliminating plosives. These tools commonly fall under the category of dynamics processors and frequently have several parameters for fine-tuning the process. Figures 2 and 3 illustrate Sound Forge's capabilities in this department. These tools are located in the program's Multi-Band Dynamics area.
In many cases, you'll need to do some more tweaking in one or two stubborn segments, including editing right down to the individual sample. Many programs have provisions for pulling the offending audio segment's waveform down or even redrawing the audio data. And let's not forget what EQ can do. With parametric equalization, you should be able to notch the offending sound into submission.
An important consideration to bear in mind is not to overdo it. Ask yourself, "Just how important is this?" I'm not advocating sloppy work, but try not to lose sight of the fact that the dialog will probably be compressed a good three times or more before the listener even hears it. That ever-so-brief sound that annoys the heck out of you might well be masked by the time music, sound effects, and compression all work their magic. If you overdo things, the dialog can easily become unintelligible-and that can't be allowed to happen.
Experience is a key factor in knowing when to let go. "Sometimes you know what's not likely to be heard once it hits the air, so you simply don't dwell on it," says Elliot. "I know that sounds lazy, and I don't mean to come across that way. The reality of the situation, however, is that there is so much material to produce in order to fill the schedule that if the offending clip is so minute we know it will never be heard, we just let it pass. We always try to deliver the highest-quality product, but there's a limit when so much material needs to be produced."
A CUT ABOVEThe promo's copy is almost always timed prior to the recording session so the engineer knows that it will fit. However, in order to get the message across in a designated time frame, the dialog occasionally needs to move considerably faster than any announcer could realistically speak. Time compression is a truly wonderful aspect of working with software editors. In the days of tape, speeding up the audio playback invariably meant transforming your announcer into a chipmunk. Nowadays, using time compression won't alter the pitch of the program material unless you want it to.
Most programs can handle at least 10 percent time compression without noticeable artifacts. This is usually sufficient because, as noted, most recordings are "timed out" prior to the session. Hence, the original performance and the desired change are usually fairly close-typically not more than about 5 percent apart.
Another common task when working with dialog is to edit out all breath marks. That's right, folks-radio announcers don't breathe! The subtle "aaah" that invariably gets recorded prior to any spoken phrase simply gets deleted. When it comes to removing all those blank spaces in the dialog track, many programs (including Innovative Quality Software's Software Audio Workshop, or SAW; see Fig. 4) enable you to set a threshold level at which any data beneath the specified value will be deleted. This sort of function can be a tremendous time-saver.
Normally, the dialog that follows the deleted breath mark remains where it originally was in time. But if you need to speed things up, here's another tip: rather than deleting the breath mark and leaving the dialog in its original location, you can slip the audio forward and butt one spoken phrase up against the preceding phrase. This alone can shave a considerable amount of time off of any audio segment. With relatively little effort, this technique, combined with time compression, can give the effect of someone speaking a mile a minute.
Another related practice for dialog editing is to position a critical portion of the announcement precisely with a sound effect or other occurrence in the promo. This technique is frequently referred to as "back-timing to music." Perhaps the phrase "You too could win a million dollars!" needs to coincide with a cymbal crash, horn blast, or similar effect. By "slipping" the track into position, or editing out a portion of a pause or breath mark, you can accomplish this easily.
Yet another common technique is to simply duplicate the dialog track and paste it to a new, open track. By offsetting the two tracks ever so slightly, by about 10 or 15 milliseconds, you can easily create a much bigger-sounding voice than you had originally. Even better, pan the tracks hard left and right in the mix for that really big sound. Try this yourself, and you'll immediately realize how many times you've heard this technique used.
SQUASHEDCompression and normalization can be effective tools when you're massaging an announcement into shape. However, you must always pay careful attention not to distort the audio. Here again, good recording technique is more than half the battle. A few peaks here and there are natural occurrences, but if the original audio file's waveforms are all over the place and have a lot of widely varying peaks, you may need to normalize the track or segment. Similarly, if your audio levels range from very low to very high, you should also consider normalizing and compressing the track. So what exactly is normalization, as opposed to compression, and how do they interact? Let's take a look.
When you normalize a file, you alter its overall volume so that the highest-level peak in the file reaches a defined limit. The entire signal is adjusted in the same proportion, which does not change the relative amplitudes of the frequencies within the signal. If the highest peak is at the selected limit, the rest of the signal will also be at its highest possible level without clipping (distortion) and without affecting the overall sound quality.
Compression, in contrast, is used to fit a large signal into a small dynamic space by squashing the signal's dynamic range so that the peaks are not excessive. But this does not affect the whole signal in a uniform way, as with normalization, so the sound is altered. In radio, where the dynamic range of the audio signal is usually larger than the receiving equipment can handle, compression can effectively fit the audio neatly within the limits of the equipment's capabilities. (For more on compressors, see "Square One: Dynamic Duos, Part 1" in the December 1994 issue of EM.)
So how do compression and normalization work in practice? If the original audio file has wide level fluctuations, ideally you would want to rerecord the track. But as this is not always possible, you might consider normalizing the file to achieve a better overall "average" level. Once the level of the VO is more consistent, you can then apply a gentle compression (say, a 3:1 ratio) to tighten up the sound. In the same way that compressing a bass track tends to give more punch to the part, a compressed VO will typically yield a greater degree of articulation, making it easier to understand the words-provided you don't overdo it. In radio production, you typically apply compression a few times, with this stage of production being the first, so don't get carried away.
Normalization can be extremely effective if, for example, you have primary and secondary VO parts in which one voice is considerably louder than the other. By normalizing the two tracks, you can effectively put the two voices on a more equal footing.
Do you always need to normalize the VO? No; if the original recording is clear and doesn't exhibit many peak level fluctuations, don't mess with it. Do you need to use compression? Except with very experienced announcers, the human voice is all over the place in terms of dynamics, so compression is usually a good idea.
PLAYING MUSICEven though most radio facilities use library music, chances are you'll still have to perform some editing on the music bed. For many musicians, perhaps the most annoying aspect of listening to radio or television spots is the proliferation of stone cut, or hard, edits that have little if any regard for musical phrasing. I never cease to be amazed at how music gets chopped to fit the time frame of the spot. Audio editors are now beginning to pay closer attention to musical phrasing and, whenever possible, are attempting to soften the blow.
Here's a good tip to make your edits more palatable. If you need to segue from one music style to another, try to make the transition smoother by using your audio program's crossfade function. In Figure 1, you'll notice there is a crossfade on tracks 5 and 6. In this case, the music makes a quick, but musically acceptable, transition from a section that contained vocal tracks to one that is purely instrumental. Taking advantage of your editor's crossfade capability can work wonders in this type of situation.
THE MIXDOWNIn the process of creating a mix that translates well to a variety of portable audio systems, home systems, and car systems, it's important to check your mix at low volume levels and at a normal studio level. "My biggest concern," Shapiro says, "is how the mix will sound on a small radio, such as a mono radio alarm clock, as opposed to a big home system. I always test my mixes by listening at a very low level to ensure the voice is intelligible over all the music and other background material. If the material passes that test, I know it will sound fine once it hits the air. I'm most concerned with getting the message across."
Because most music and sound effects come from libraries that are already professionally mixed and mastered, the bulk of your focus will continue to be on the VO. When you're equalizing the VO, you will generally find that the majority of your EQ tweaking will occur in the midrange, as this is where the human voice resides.
In order to give some added emphasis to the dialog, you should experiment with the midrange frequencies. As is usually the case, moderation is the key to success. Don't forget that if you boost the high frequencies too much, you're likely to increase the sibilance in the track. If you overdo the low frequencies, you're likely to "muddy" the track and perhaps make plosive sounds more noticeable. Generally speaking, applying a slight boost to the midrange, around 3 or 4 kHz, will bring the announcement straight to the front of your mix.
When you're adding signal processing to the music and sound-effects tracks, your ear must make the final decisions. With explosions and other sound effects, a good, long reverb tail can make all the difference as the sound tapers off into oblivion. With music beds and effects, subtlety tends to be the key factor.
Once the announcement, music, and sound effects reach the point where you feel they are just as you want them, it's time to mix. To expedite your mixdowns, you may find it helpful to place certain types of elements on the tracks in a consistent manner. Get into the habit of placing dialog on tracks 1 and 2 and stereo music beds on tracks 3/4 and 5/6. Sound effects can go on tracks 7 and 8 and, if necessary, on 9 and 10. As Shapiro notes, "Developing a consistent work approach has the benefit of enabling you to work faster with less error. In doing so, you may not even find it necessary to get out your Sharpie to label so relatively few tracks-you just know what goes where."
DUCK, DUCK, GOOSEWhen listening to a promo spot, you've probably noticed how the music bed tends to drop down beneath the VO once the announcer begins speaking. This process of lowering the music level is known as ducking, and it's one of the most common techniques employed in radio production. These days, the ducking process is generally handled by the editing program's automation system and level changes are generally visible onscreen (see Fig. 5). By taking advantage of automation, you can achieve perfectly adjusted levels on a consistent basis. In addition, by writing these level changes into automation, it's easier to edit and fine-tune the mix later.
Typically, you ride the faders that govern the music bed, lowering and raising them as the spot requires. Generally, a change of a few decibels at most is all that's required, though your ears and the person who's paying you will be the final judge. The change in level should be enough to get out of the way of the voice, but not enough to make the music sound like it has moved across the street. If your computer program supports such editing tasks, you may prefer to draw the level changes. Some engineers find that doing so provides them with a greater degree of control.
When you're ducking the music for a promo, knowing whether to make a rapid or a gradual change in level is something that comes with experience. For the most part, your ears will tell you what works and what doesn't. "Sometimes a sudden level change does the job, while other times it sounds too abrupt," says Shapiro. "I generally prefer to make these changes gradually, especially when it comes to raising the music bed toward the end of the announcer's dialog. The beauty of using a computer for this task is that you can see where the VO begins and ends, so you can drop your levels and gradually bring them back up to coincide with the end of the announcement."
READY TO PRINTGenerally, the second application of compression will occur during the mix to stereo. For this, TC Electronic's Finalizer, inserted in-line between the workstation and the mastering deck, has proven to be a popular tool in radio production. Shapiro notes, "The Finalizer is a truly wonderful tool for radio-broadcast work. It contains an excellent assortment of presets that, without any editing, are ideal for compressing the mix on its way to the DAT recorder."
As a general rule of thumb, the compression ratio at this stage of production is in the 4:1 range. As before, this is fairly gentle compression, the goal being to even out the entire stereo mix, perhaps making it sound louder. It can't be overemphasized that while compression is common in radio production, you must use moderation, because the station's broadcast technicians will compress the audio signals yet again on the way to the transmitter. Too much compression, and you're likely to end up with something that is unintelligible and without any real dynamics.
While DAT and CD-R are certainly common mixdown media, it should be noted that many random-access systems have provisions that enable you to mix internally to a pair of open tracks. The delivery medium of the final product is not really that important, as long as it is common. Stations will transfer your material to whatever medium they want.
However, it is important to recognize that in many smaller U.S. markets, as well as in Latin America and Asia, you may need to deliver the finished product on MiniDisc. Due to its near-CD sound quality and instant start capability, MiniDisc has had its fair share of success in the broadcast market, partially replacing the conventional NAB (National Association of Broadcasters) cartridge that was used for years. Whatever you do, never supply a station with a cartridge. It is a standard in the industry, but alignment is rarely the same from station to station.
At Mega 100 and many other large radio stations, the finished spot will typically be archived to an audio server where all files are numbered and categorized. This is generally accomplished by way of a real-time transfer via a DAT deck or CD player digital output. Once on the server, the promo most often resides as a broadcast WAV file. From there, the audio can easily be recalled at any time for broadcast.
AIR ITAt this point, you should have a fairly good understanding of what you will probably encounter when you land that first radio gig. However, I have a few closing thoughts.
For starters, I can't stress enough just how fast-paced this business is. Since time tends to exist in far too little supply, do yourself a big favor and learn how to use your equipment before you commit to a project. If you miss your deadline or in some other way fail to deliver, you will most likely never hear from that employer again. Your first shot will probably be your last if you jump the gun before you are prepared.
Second, listen and learn. Emulation is in many ways the best method of learning a new skill. Listen to radio: know what's being done and, more important, what isn't done. If your work doesn't fit with thestation's style, you'll end up wasting more than just your own time and effort.
Many vocational schools for the creative arts as well as colleges and universities have programs to help you get started. If you prefer, try Don Elliot's interactive CD-ROM, which teaches the ins and outs of radio production (see Fig. 6).
Radio production is a fascinating field that provides a wealth of challenges and the opportunity to stay steadily employed. So don't just sit there-go for it!
A percussionist, Roger Maycock spent years hitting things before realizing he should be holding sticks. He now sits at a computer attempting to find the Off button.