Now that your recording is all cleaned up, it is time to export it, to a compressed lossy format, either Ogg Vorbis or MP3 (MPEG-1 Audio Layer 3, not MPEG-3). (Refer to Chapter 13 and Chapter 14 to learn how to enable MP3 support on Linux and Windows.)
MP3 is the universal portable digital audio format: Everything supports MP3. As we discussed in Chapter 1, there are differences between Ogg Vorbis and MP3. Some listeners claim that Ogg Vorbis delivers higher quality than MP3. At low bitrates, neither one sounds very good to me; rather, they sound flat and without character or depth. At higher bitrates, both sound acceptable, though I think Ogg delivers a more accurate, natural sound. MP3s sound a bit too-bright and crunchy.
Ogg Vorbis supports multichannel surround, while MP3 supports one or two channels. MP3 is encumbered by patents held by several different patent holders in different countries, and the last one expires on 2017. Does this mean you’ll get a visit from attack lawyers if you distribute your work in MP3 format? I don’t know. It seems that a lot of artists distribute their work as MP3s without paying license fees and don’t get in trouble. Ask a lawyer who specializes in these issues.
Even if your listeners all have megamondo broadband, it’s a good practice to maintain a frugal attitude. And they don’t all have fat pipes; estimates of US dial-up users range from 25 to 45 percent. Don’t forget that the Internet is international, so your potential audience includes countries that rely on dial-up and low-speed wireless. Even for broadband users, network congestion and queuing gum up streaming media, so the moral is that streaming audio is still more reliable at lower bitrates.
As always, there are trade-offs. If your podcast host allows it, you can offer your listeners a choice of listening options at different quality levels and both streaming and downloadable files.
Ogg exports are easy in Audacity. You might have read how-tos that say to use a variable bitrate (vbr) rather than a constant bitrate (cbr). Although this is good advice, it is unnecessary because Ogg Vorbis does not support cbr. Select File > Export > Save as type: Ogg Vorbis Files, and then click the Options button (Figure 7-14). Exporting to Ogg Vorbis is simple: Pick a number. 0 is the lowest quality and smallest file size, and 10 is the best quality and largest file size.
A variable bitrate means that simple passages will have a lower bitrate, and more complex passages will have a higher bitrate. This delivers better quality than a fixed bitrate. You won’t be able to calculate file sizes precisely, but the fine folks who invented Ogg Vorbis don’t think much of measuring quality by bitrates anyway. To give a basis for comparison, a quality setting of 3 results in an average bitrate of around 112Kbps, and Ogg fans claim it sounds better than a 128Kbps MP3.
Ogg Vorbis uses lossy channel coupling up to level 5. Redundancies between the left and right channels are combined to save space. This might affect the stereo imaging, so if this bothers you, use a quality setting of 6 or higher because channel coupling is turned off at this level. I use 6. My ears can’t tell the difference between 7 and up. Your ears, of course, may tell you a different story.
For a simple voicecast, you can go all the way down to 0, and it still sounds acceptable.
On both Linux and Windows, you may have to install an MP3 encoder separately. Use the LAME MP3 encoder because it is high quality, free of cost, and open source. Linux users can install it from their usual package repositories, and Windows users should follow the instructions at the Audacity downloads page (http://audacity.sourceforge.net/help/faq?s=install&item=lame-mp3).
Note
Windows users should visit the OSSwin project (http://osswin.sourceforge.net/) to find links to hundreds of excellent open source applications for Windows.
I don’t know if audio geeks love to argue for the sake of arguing or simply have strong passions about everything, because MP3 quality settings are minefields of controversy. Just follow this simple rule, and you’ll be fine: I, your author, am right (except when your own ears tell you differently).
In Audacity you have four categories for setting the MP3 bitrate: Preset, Variable, Average, and Constant.
Preset gives you four different prefab combinations that are recommended by the LAME developers: Medium, Standard, Extreme, and Insane. Standard is fine for music, but you can go lower for a simple voicecast. Medium, Standard, and Extreme all use variable bitrates (vbr). Variable bitrate means the bitrate varies according to the complexity of the sounds, which results in better audio quality. In the olden days, variable bitrates gave MP3 players fits because they were designed for constant bitrates (cbr). But that was a long time ago, and these days variable bitrates shouldn’t present any problems. Variable bitrates are always defined within a range, like the 170Kbps to 210Kbps of the Standard preset.
Insane uses the maximum constant bitrate, 320Kbps. In my experience, the Standard setting (170Kbps to 210Kbps) is about as good as MP3 can deliver, and anything over that doesn’t make a noticeable difference.
Variable has a range of 10 quality settings. Quality setting 5 is 110Kbps to 150Kbps, which is fine for simple voicecasts.
Average is, in my opinion, not worth bothering with. It is a compromise between Variable and Constant bitrates. You set a constant bitrate, but since it’s an average, a little bit of fluctuation is allowed. So the quality is not quite as good as Variable, and it’s a little better than Constant bitrate.
Constant bitrates are supposed to work better for streaming audio over the Internet, and you get a predictable file size. If you want to use constant bitrates, Audacity gives you a full range to choose from.
Set Variable Speed: Standard. Standard gives better quality, and it takes a tiny bit longer to encode.
Figure 7-15 and Figure 7-16 show the MP3 export menus and the various quality settings.
Leave Channel Mode: Stereo, because you have to be desperate for storage space to even think of using Joint Stereo. If you are desperate for storage space, to the point that you’re thinking of encoding your MP3 files at a bitrate of 112Kbps or lower, then give it a try. The term joint stereo is a poor name that does not describe what it does. A popular misconception is that it joins the two stereo channels and destroys the stereo separation. As always with audio engineering, it’s more complex than that. Here is a new audio term for you: human psychoacoustics. Isn’t that a splendid phrase? No, it doesn’t mean psycho like a crazy person with a big knife, sorry. It means engineering audio in a way that takes into account how the human ear perceives sounds, instead of adhering to a strictly mathematical model, which does not correspond to how human hearing really works. It is quite sophisticated, and if you don’t like the way joint stereo sounds, it’s easy enough to do over.
What is joint stereo? In the MPEG-1 Audio Layer 3 standard, it means using any of several techniques to reduce bitrate while preserving an acceptable level of quality by merging redundancies in the left and right channels. These “redundancies” are frequencies that the human perceives as less directional, which are the upper frequencies. It should improve how recordings at low bitrates sound because more audio data are preserved at the expense of stereo imaging.
There are other lossy, compressed audio formats that generate small file sizes, such as AAC, WMA, and RealPlayer. Audacity supports these via the FFmpeg library on Linux and Windows and has some nice export presets. You may also use the external encoder of your choice, which you can access via the Save as type drop-down menu in the Export dialog (Figure 7-17).
Note
See Chapter 13 and Chapter 14 to learn how to install FFmpeg and MP3.
AAC wasn’t very popular until Apple chose it as the default format for its iTunes Store and all of its audio devices. WMA is the Windows Media Format, and RealPlayer is . . . well, RealPlayer. AAC, WMA, and RealPlayer are all closed, proprietary formats that are encumbered with various unpleasant features such as digital rights management (DRM), tracking software, and patent guff.
Once you have decided what format or formats to export your podcast to, select File > Export to make your actual exports. You can make all kinds of exports from the same Audacity project. Note that you can export part of a track by selecting part of it and then selecting File > Export Selection.