Podcasting Audio Production 101

The following is a sort of quick and dirty guide to audio post-production for podcasts.

I’m trying to provide a fair view of the playing field, but can’t possibly cover everything. To begin with, there is no comprehensive list that is “everything there is to know” about
audio production. The process simply doesn’t work that way. Not only is audio production an ongoing process, it is a process that should adapt itself to the specific material that you’re dealing with. What you must
learn to do as a producer, more than anything else, is to listen. Everything else is just figuring out what to do when you hear something you don’t like, or how to create the sound that you have in your head.

Thus, there is no one “right” way to produce an album or a podcast. The project itself defines its needs, and the more flexible you are, and the more expansive your “toolbox,” the better. At the same time, each producer inevitably develops their own signature sound which somehow winds up a part of the final material, even if they were made with the same exact tools as another.

For most podcasting tasks I use Adobe Audition (which is essentially the same as Syntrillium’s Cool Edit Pro 2.) I’m not saying that this is the best software to use – chances are if I was working on a Mac I’d be using
Digital Performer. However, it is more than sufficient for the task at hand, especially when coupled with a robust array of Direct-X plug-ins. For this I will refer to plug-ins in the Waves Platinum Bundle. I have and use many others myself, but many of the FX in this bundle are invaluable.

Editorial:

The first rule of thumb is simple: garbage in, garbage out. You can tweak things out in post- turning a flat recording into shrieking static and a nauseating array of trailing 6 part harmonizers and multi-tap arrays set to the first couple numbers in the Fibonacci sequence simply because you can.  However, it is distressingly difficult to make an awful recording sound clear and pristine. Most seasoned producers can’t even do it. Don’t cause yourself this kind of pain. Get as clear of an initial recording as possible.

That means turning off background noise. (Shooting your neighbors, if necessary.) For podcasting, it also means investing in a mic and an audio interface that allows for a 3 prong (balanced) connection. You can get a variety of USB 2.0 or firewire audio recording units. Most snobby producers will tell you the D/A conversion on these units is not great. Well, they are right, but keep in mind, you are encoding all your work into mp3 format. So getting a $100 Shure SM-58 and a $300 M-Audio box would be a reasonable investment, but you really don’t need that beautiful, $3000 Rhodes, or a tube mic pre-amp that sounds like the kitchen of a Diner when it’s heating up.

When editing people’s speaking voice, you want to pay careful attention to rhythm. In other words, there are times when you want to delete “ums” and “ahs,” there are times when you don’t. (Also note: you’ll freak yourself out with how quickly you come to recognize stutters in a speech pattern just by how the waveforms look.)

I cut my teeth on this process editing doctors talking about Pharmaceutical drugs. We had to edit their rambling diatribes to a script that they oftentimes went off. We had to be fascistic with the deletion of ums and ahs. It was hellish. I don’t think I’ve fully recovered.

You don’t need to do that to yourself. But you would be amazed how much time you can save if you are a little conscientious about cleaning up stutters and unnecessary pauses.

Also be aware of the uncomfortable or unnecessary silence, as opposed to the pregnant pause. The former you need to delete with impunity. The latter, if removed, will leave a sterile, pernacious wasteland.

Alright, I’m exaggerating. But the distinction is important.

There is a whole lot more that could be said about the recording process when you’re recording musical instruments- but thankfully, the bar is still pretty low with podcasting. Know what the fuck you’re doing, and you will be a genius. Relatively speaking.

Some notes about Audition:

Though I will leave it to you to look up the basic instructions for operating the software, there are a couple things you should bear in mind.

First, make all your recordings 44.1 kHz 16 bit. You want a higher bit-rate for your studio album. For a podcast, those settings will be just fine.

Save all your local audio files as .Wav or .Aif. If you have to work with mp3s for your background music, or if the teleconferencing software you use records mp3s, it isn’t the end of the world, but that double encoding is part of the reason why podcasts tend to sound like garbage. On the other hand, since most teleconferences are recorded over phone and saved at a horrifically low bit-rate, double encoding is the least of the issue.

You can create automated changes in volume, panning, % of FX, and so on in an individual track by enabling envelopes. (You do this in Audition by going under View > Show Pan Envelopes, View > Show Volume Envelopes, and Enable Envelope Editing.)

There are a lot of tricks you can employ make these low quality recordings as clear as possible, and filter out the noise. I’ll get to some of that soon.

FX:

Once you have your basic recordings together, drop them into the Multitrack mode. (You do this by hitting Control-M when in the single waveform edit view, or selecting Edit -> Insert in Multitrack.)

Be aware of which track you are grouping things in. For instance, you may want to put sequential “narrator” vocal tracks in one track, and your interview recording in another, and full volume music in another, and
background music in another.

The reason for this is that you will next be adding non-destructive, real-time FX to each track, which will apply to the entire track. You want consistency between certain elements, and this is part of how you do it.

To apply FX like this in Audition, simply click the blue FX button associated with each track. (Note: Acid actually works more or less the same way.)

This calls up a dialogue box. You will be picking your FX from the Direct-X dropdown. Trust me, most of the presets that come with Audition aren’t worth your time.

I will cover just a few of the primary FX you should be dealing with:

EQ:

Use the Waves Q10 paragraphic EQ (or another paragraphic EQ, if you aren’t using the Waves bundle). Make sure you don’t select the Mono version unless if it is a Mono track. (I assume they make this distinction so you can use up less CPU if you don’t need stereo processing. But who knows, I’ve never bothered to look into it.)

You’ll need to play around with the Paragraphic EQ to get a handle on how it works. Do a little Google research. What I want to talk about is subtractive EQing.

Subtractive EQing is an approach which can be used in two ways.

First, you grab one of those points on the graph (Q1-10). Now push it up into the + (you are adding gain to that frequency) and reduce its range somewhat, (this will effectively turn a gradual wave into more of a spike.) Now play your wave, and slowly drag your point from the high frequencies to the low. When you hear a tone that is just God Awful, drag the point into the negative an amount that sounds right to you. You are effectively reducing or removing that offensive tone. Now you can go in and find the harmonics of that tone if you like.

The second form of subtractive EQing, which you won’t deal with as often in podcasting, is when you take the frequencies of certain tracks that are taking up “unnecessary real estate.” Remember, you only have 2 speakers or headphones (usually), and you have all these tracks playing simultaneously. Most of the incredibly low frequencies on a high hit or crash cymbal, for instance, are just going to be garbage. You can reduce these and all you’re freeing up is empty space for the things that should be in that “zone.”

(I could get into visual metaphors at this point that I’ve come up with fellow producers while really high, but I’ll spare you. Suffice it to say that submarines, sonar, and giant squid are often involved.)

The only time you’ll have to deal with this in most podcasts is when dealing with background music. If you scoop a bit from your music out of the frequency range of most speaking voices, (let’s say 100-1000 Hz, it’ll vary), you’ll free up some “speaking room” for yourself without having to cut the volume on your music track quite so much. Let your ears be your guide.

There is a Zen art to EQing. You will have to play with sound for many, many hours before you start to get it. All I can say here is this: beware ear fatigue. Especially when it comes to EQing. If you have been working eight hours straight with loud monitors, it is altogether possible that you will come back tomorrow and discover that you created a monster. You also don’t need to always subtractively EQ- for instance you can bring out a certain timbre in a voice or the pop of a snare drum by accentuating certain frequencies. But it’s better to try to take away before you try to add.

Finally- If there is a consistent hum or tone in your recordings, you can also use the Waves X-Hum to reduce 60 Hz hum from electronics, the inversion effect of poorly grounded recording
gear, and the like. This plugin is more self-explanatory than EQing.

Noise-Reduction:

Even though you should try to create the best initial recordings possible, they aren’t always terrific. That’s what plug-ins like Wave’s X-Noise are for. You can find this under the Direct-X plug-in menu (supposing you have Wave’s Platinum installed, if not, there are many other decent pro or quasi-pro noise reduction plug-ins.) This is one of those plug-ins that works best by using your ear and playing with it for a while. Generally,. set your Resolution to “high,” and when in doubt, click the tab that toggles between “Audio” and “Difference.” The “Difference” plays the noise that’s being filtered. Oftentimes it is easier to distinguish when you’re filtering out unnecessary noise and when you are cutting into your actual coherent sound.

A word of warning: if you overdo your noise reduction, the wave file will sound worse than it did beforehand.

Effectively Boosting Your Vocals:

One of the most important elements of the vocals in a podcast is that they are clear. The Waves R-Vox plug-in is great for this, and it is also incredibly easy to use. Apply it to your channel, set the Noisegate to be above the noise-floor but below the volume when you are speaking softest, and set the compression higher to increase your volume.

You may find that you want to add a little more room color to the vocals as well. That’s what the reverb plug-ins are there for. (Also for doing A Capella versions of “Bela Legosi’s Dead.” I don’t suggest you share these with others, however.)

A Note About The Signal Chain:

As you add FX, it actually matters what is being applied first, second, third, and so on. You will learn what works and what doesn’t in what situations, there are simply too many permutations to possibly go through in a single article. One word to the wise: apply your reverb after any sort of intense Compression or Limiting, such as the R-Vox plugin, as compression essentially increases the volume of the noise-floor above a set level, and reduces the volume of the peaks above a certain level to create more consistency- but if you’re not careful some of the reverb will be included in that and you’ll get some fairly odd results. (Not that I haven’t done that intentionally
once or twice…)

Mixing:

Like everything else that we’re covering in this all-too-brief article, mixing is a complex art that can take a lifetime to master. Your ears should always be your guide, above and beyond your experience.

Beginner rules of thumb:

You want to mix your files together so that most of the time, the volume level is consistantly coming close to 0 (max) on the meter, but not clipping (in Audition, and in most apps and with most hardware, if it clips a red box or circle will become lit above the meters.)

If your mix does clip, but only periodically, don’t worry about that. It’ll be fixed in your mastering process through the use of a Master Bus.

Placement.

You can also create a much “roomier” mix by being aware of stereo placement. I know this is podcasting, but not everything has to be dead center. (And unless if you want to create a schizophrenic, nightmarish experience, like I did with some of my more “experimental” albums and projects, don’t overdo the stereo panning, especially the real time shit.)

Keep any background music at a low enough threshold that it doesn’t interfere with voice-overs. Of course, there are times when you may want to intentionally make the voices more a part of the music. Ignore this advice in those cases.

Mastering:

Once you’ve set the volume levels, its time to maximize your sound.

(I feel like I need to say that while flexing my muscles and pouring a mixture of creatine and human blood down my open gullet.)

Be that as it may, you “maximize your sound” most easily by setting up an effects Bus.

Go to view->show mixers window from the Multi-track mode, or hit Alt-2. On the mixer, make sure the buttons to the side of the mixer entitled “Out” and “Bus” are both depressed.

Now click on the first track, where it most likely says Out 1. (It may something Out 2, or Out 3, depending on how your sound card is configured in relation to the software. You can get those options under the Options menu.
I’m not dealing with that here.)

Now click “New Bus.”

This is where you will be selecting the FX you will be applying to all of your tracks. For each of your tracks, on the mixer, where it says “Out 1″ click that and select the Bus
from the dialogue box, rather than your sound card.

For the FX you use, I suggest you select another Q10, to EQ the total audio output, and also the Waves L2. There are many others that can be useful in certain situations, but those are the easy necessities.

For the Waves L2, select the preset of “High Res CD Master,” and then tweak the levels to taste. Make sure there’s still a little breathing room, but that the peaks are hitting the top.

You will notice that once all of your tracks are set to go through the Bus that, if your mix was clipping before, it isn’t clipping now. That’s from the limiter in the L2. Now if you crank up the volume levels on all the tracks, it will never clip. The problem is, of course, everything will get squashed and sound awful. So you’ll want to set your levels again so that you find a decent balance between loudness and, well. Squashed awfulness.

When you are happy with the mix, in Multitrack mode select “edit > mix down to file > all tracks.” This will create a single file out of your session.

Save this as an mp3, and then make a backup as a full resolution .wav file. Many people podcast at pretty low bitrates. You don’t need to be many people. I suggest somewhere between 128 and 256 kbps depending
on how large a role music plays in the show. If the show is mostly spoken word, 128 is just fine.

Now open this mp3 in iTunes, and right click on it in the playlist. Edit your ID3 tags, and add all the information you like to help categorize and define the file- a brand image, the production credits, the website, and so on.

Viola, another episode for your podcast. Well, those are the basics. Have fun experimenting.