While I’m sure you save your projects with the uncompressed versions of all its constituent waveforms, when it comes time to send out a file to a client or upload a piece to your site, you’re most likely going to encode as MP3. In Audition, this is very easy, and most of us are comfortable doing it. But, it’s worth explaining exactly what’s going on “under the hood”, and what some of the more obscure options do.

First, a word on MP3 in general. This is a fantastic technology that’s not only made it easy to take high quality audio with you, but it’s revolutionized many industries as well. Whereas it used to be a hassle to deliver commercials and programs to radio stations (think FedEx and CD’s in a best case scenario), we now just e-mail each other MP3’s. As a radio production guy, I also incorporate MP3’s into my production, so I’ve probably got about 16 open right now. But please. PLEASE! Never save your projects with those MP3 files as they are. Hard drive space is cheap, so please always make sure you’re saving your sessions and projects with the full, uncompressed audio so that you’ve got a perfect copy to work from every time.

Never discard your original files. Never save any part of a final project as MP3.

Please only use MP3 to exchange audio casually. E-mailing a spot to a client? Yes. Posting samples of your work for people to stream/download online? Yes. Submitting a master to a record label? No. Saving loops for later use? No.

I have posted (right) a capture of the MP3 options dialog. If yours looks smaller when saving, click the “Advanced” button to show all the options.

(Note to Audition team: the “Help” button and “F1″ don’t do anything in this dialog. Actually, that’s not true. Sometimes it just brings up Windows Explorer. This may be a bug.)

In my example, I’ll be saving at 224Kbps, stereo, a 44.1 kHz sample rate, and a constant bitrate. I have no custom options selected. Pretty easy to understand. Now, let’s break it all down.

CBR or VBR?

Constant bit rate or variable bit rate? Generally speaking, at comparable quality, VBR will yield lower file sizes. It works by altering the encoding bit rate depending on the audio coming into the encoder. If, perceptually, there isn’t much audio content, the bit rate will momentarily be low, thus saving file space. However, because each file is different, it’s tough to predict the exact resulting file size. (You can usually “ballpark” it once you’ve done it a few times, though.)

When encoding in VBR, you are given a choice of ranges of bit rates, rather than one specific number. The lower number will indicate the lowest possible bit rate, and the higher number the opposite. For example, if you’re encoding VBR using the “Highest Quality” (150-240 Kbps) preset, you know that the bit rate will never go below 150 and will never exceed 240. For average audio, then, its average file size is comparable to 192Kbps audio encoded as CBR, but with the advantage of perhaps using higher bit rates when needed.

In contrast, constant bit rate will yield a predictable file size by using the same bit rate through the entire file. You will also be able to see the exact ratio of compression, too, when encoding in this method. In the example in the screen capture, you see the ratio of “6.3:1″. In English…this file, versus its uncompressed counterpart, will be 6.3 times smaller. If your source audio is 100 megabytes uncompressed, the resulting MP3 will be just under 16 megabytes.

Personally, I never bother with VBR, simply as a matter of habit. Any player out there today can play either, but I’m not short on hard drive space, and I’m very much married to my presets. Variable bit rate files, however, are increasingly popular. Play around and see what you like.

MP3 or mp3PRO?

mp3PRO is an extension of the MP3 codec that allows for “CD quality” at around 64Kbps. Unless you have a specific use for this, and you know the audience will be able to play back mp3PRO files, I would stick with MP3. It’s a wonderful technology (especially for podcasting, Internet radio, etc), but player support is often sub-par, presumably because of licensing problems. Your options in mp3PRO are roughly analogous to the standard MP3. Just remember that, on average, you can halve the bit rates you’d normally use in MP3 to achieve the same effect in mp3PRO.

Bitrate

That’s the dropdown dialog box you see below the options we just discussed. In my case, I’ve selected 224Kbps. (You can also specify this manually below.) The higher the number, the better quality the audio. Unless you’ve got ears of gold, anything above 256Kbps will be overkill. (Indeed, the original specification for MPEG 1 Layer 3 audio stops just above that at 320.)  In fact, most people won’t be able to detect the difference between 128Kbps and 192Kbps.  You’ve probably found a good system that works for you.

But, if you’re new to this world and aren’t sure what to select, here is a quick guide based on the material (we’re assuming you’re saving in stereo unless otherwise noted):

Podcasting:
- 32Kbps - A great option “B” to provide for the bandwidth impaired.
- 64Kbps - Great for longer programs. If it’s mostly chatter, however, consider converting to mono (see below). TWiT (This Week In Tech), one of the most popular podcasts/netcasts online, delivers the MP3 encoded at 64Kbps in mono, for example.
- 128Kbps - Great short programs, or if you’re featuring music or other quality-demanding content.

Music:
- 64Kbps - Good for posting a low quality preview for fast streaming.
- 128Kbps - Acceptable Quality
- 192Kbps - Good Quality
- 256Kbps - Great Quality

Production:
- 128Kbps - Good for showcasing, airchecks, and demo reels.
- 224Kbps - Good for sharing workparts and finished material.
- 320Kbps - A great way to archive old material and workparts. You burn the original session files to a backup, then keep the MP3 files around so you always have them at least available to listen to.

Just Voice:
- 128Kbps - Great with “convert to mono” option selected, see below. (Most “voice guys” I deal with in the radio industry send out their work in this manner.)

Convert To Mono

If all you’re sending out is a voice over, or you need a “lo fi” version of an MP3 quick, this option is for you. It saves you the step of having to do the conversion on the audio itself…the MP3 encoder will take care of it transparently. Be aware, however, that the highest bit rate available in mono is 128Kbps. (This options is not available for VBR files.)

Maximum Bandwidth

If you read back to my post about digital audio basics, you’ll recall that the highest frequency a file can represent is denoted by halving the sample rate. Accordingly, Audition will set this value to be zen with your sample rate. Trust me when I tell you, leave this alone in 99.99% of cases.

CBR Bitrate

If you’re encoding CBR, this is a chance to specify an exact bit rate. The standard is to use increments of 16, but it’s technically possible to encode anywhere from 20Kbps to 320Kbps. So, yes, you could encode at 312Kbps if you so desired. You won’t have this option, obviously, if you switch to VBR encoding mode.

Sample Rate

Regardless of the sample rate of your source material, you can encode the MP3 with a different sample rate. I would strongly discourage delegating the job of re-sampling to the encoder. This is something I would do inside Audition (use “F11″ or your shortcut key) prior to generating the MP3 file. There are two reasons to avoid this. First, not all bit rate and sample rate combinations are valid. You could, in theory, encode a file nobody could play. Second, I’m not sure if the audio would be correctly low pass filtered, which could be a quality control problem.

This is just me, but I stick to sending out MP3’s at 44.1 kHz unless there is a very specific reason to do otherwise. It’s universal, and it works. (But yes, still save your projects with the uncompressed files at the original sample rates.)

Codec

You can read more about this in the Audition help file. The only thing I have to say about this is…leave it at “Current - Best Quality” unless you’re an audiophile who’s also a glutton for punishment.

Allow “X”

(Editor’s note: I usually leave these UN-checked. These are expert or legacy options that have specific uses that you won’t usually need to worry about.)

…Mid-Side Joint Stereo: If you’ve created a specific stereo image, this affords you an opportunity to preserve it during the encode. However, technically speaking, your file won’t truly be stereo. Rather, it’s stored as a combination/difference between the channels.

…Intensity Joint Stereo: This will obliterate the stereo image by saving some frequencies in mono and others in stereo. It will reduce your file size a bit, but at a great cost to the quality of your resulting file. This is very much a legacy option. Almost a faux mono. Never use this on stereo material like music.

…Narrowing of Stereo Image: With this setting, the encoder will try to increase the “big picture” quality of the file by quashing the stereo image where it’s deemed perceptually pointless.

Set “X” Bit

These are meta-data bits in the file that denote specific properties. Think of it as additional tagging for your files. Almost a “for your reference” bit.

Padding

This is one option specific to the decoder that will be used. Padding in MP3 files is so boring to explain that it’s likely to put you into a coma. Suffice to say, unless you have specific knowledge that you need to encode with “no padding” or “always padding”, just leave it at “ISO Padding”.

If you’re wondering why, in some cases, there is a tiny amount of silence at the beginning and end of your MP3 file, this is where it’s coming from…but it’s the decoder, not Audition. This is a technical limitation in the way MP3 files are structured. It’s best to accept and ignore it and know to never to save anything that you would need to “loop perfectly” later on as an MP3. (There are codecs out there nowadays to do “gapless” encoding, but guess how frustrating they are to use…)

Set all decoding to 32-bit

If you’re in my school of thought of “always work in 32-bit”, then check this box any time you save an MP3. It’s an Audition-specific command that will cause the file to be up-converted to 32-bit automatically when you open it. Saves you the trouble of having to mash the F11 key. Your iPod or Winamp will just ignore this instruction when playing back audio, don’t worry.

Encode Stereo as Dual Channel

If you’re a quality snob, hit the check box. This will give you a true stereo MP3 file, since the left and right “streams” will be saved and encoded distinct from one another. In reality, though, this option is actually only useful if what’s going on in the left channel has nothing to do with the right channel. This would be a good option, for example, if you’re providing a click track on the left channel and a guide track on the right. In that case, you’d want each to be completely separate from one another. However, in everyday use, this is usually unnecessary.

Write CRC Checksums

For the unfamiliar, CRC and technologies like it, allow for a checksum to be embedded into the file. The checksum, sometimes called a “hash” or “fingerprint”, allows the integrity of the file to be known before it’s played back. I usually leave this checked by default, since it’s a good sanity check. It will only cost you a few bytes in storage, something you’ll never notice.