This month, we’ll take a deeper look into Audacity, familiarizing you a bit more with the interface and some of the more esoteric but useful features of the application. The full range of features is so rich, I can hope only to introduce you to the most useful.
Last month, we described the open source audio editing application Audacity as the 16-ounce framing hammer of podcast creation and editing. This month, we’ll take a deeper look into Audacity, familiarizing you a bit more with the interface and some of the more esoteric but useful features of the application.
Let’s start by setting up the application to meet your needs. From the menu, select Edit Preferences. You’ll find a window tabbed along the left edge, full of options to tweak Audacity. Let’s walk through the critical items tab by tab.
Audio I/O describes the input and output of the Audacity application, beginning with the sound device from which Audacity will capture audio. In most cases, this will be /dev/dsp in Linux, though your system may vary. You’ll note, too, that the I/O tab includes both recording and playback – yes, input and output.
You’ll also find a drop-down to select the output format: stereo or mono. When opening a new file in Audacity, this selection is used as the default. If you’re recording a podcast that’s primarily voice, mono should be your choice, if your voice will be the only one in the audio. You can select stereo if the podcast will be primarily music recorded directly to Audacity from a media player, and if your sound card supports stereo in. (Mixing and imaging multiple voices is best accomplished in hardware, something we’ll discuss in an upcoming stories in the series.)
Finally, the Audio I/O preferences tab allows you to make a couple of important choices. The first, “Play other tracks while recording new one,” allows you to create multitracked recordings on the fly, or, to mix musical cuts on the fly. If you have a good plan for your podcast episodes, this can save considerable time. Without the planning, though, you’ll face a train wreck in trying to keep up with multiple musical selections when this option is checked.
The last option is very useful. “Software Playthrough” allows you to monitor your recording in real time. This is a great feature for all types of podcasts, but may require a high amount of processing power. It’s especially useful if you’re doing some hardware processing just before sending the signal off to the sound card. The feature allows you to hear what the sound card is hearing, as it’s hearing it. I can tell you firsthand that what you hear through the monitor output of a mixer may not even come close to approximating what you’ll hear in the recorded audio. This feature prevents those surprises.
The next tab in the preferences window is the Quality tab – another important tab for any podcaster. It sets the sampling rate at 44100 by default, the sample format as a high precision 32-bit float by default, and the sample rate conversion. The latter determines the method by which the software will interpolate the gaps between the samples. In other words, if your sampling rate is 44100, you’ll have 44,100 digital samples of the source sound per second. But, that’s only a sample of the original source sound, not a constant stream. Interpolation defines the method by which the software will make an educated guess as to what the other 43,999 gaps per second contain. While it may be hard to believe that these gaps are noticeable, a digital recording without interpolation has a harsher, less realistic sound.
Interpolation in Audacity is broken out into two types: real time and high quality. The Sinc method of interpolation is the chosen method in Audacity, with fast Sinc as the default for real-time interpolation. Fast Sinc makes broader judgments as to the in-between sounds. High quality is the default method for high-quality recording and, as you might guess, makes narrower interpolative choices. You can choose to change the real-time method from fast to high quality. Just be aware that the processor cost is considerably higher. In most cases, the defaults will work perfectly well for podcasters.
The quality tab also contains preferences for dithering. In audio, dithering is the process of artificially introducing some noise into a signal. This is a mathematical necessity related to the sampling format. In the old days of just a few years ago, the representation of the digital sample points was in a 16-bit format, where usage of all 16 bits represented a full-amplitude sample. However, a large chunk of every recording, being less than full amplitude, utilized far less than those 16 bits. And, the lower the amplitude, the more square the wave, the harsher the sound. These audible imperfections were simply the result of a finite, fixed number of available bits.
So dithering is the process of pushing the amplitude of those lower significance bits to a rounder, more realistic state. That’s done by introducing noise at a level less than that least significant bit. That noise pushes the least significant bits up a bit in amplitude. In plain terms, dithering is the process of naturalizing a digital signal by introducing noise.
Dithering is less important when you’re using a 32-bit float to represent the sample points, as is the default in Audacity. The increased dynamic range means that those least significant bits take a more natural form. However, any changes in gain, the ratio of signal output to signal input, will require some dithering. In podcasting, the most likely place for a change in gain is in normalization; adjusting the amplitude of a waveform to its maximum undistorted level. We’ll talk at length about gain and normalization in a later installment in this series. Understand, though, that some dithering will take place in your audio file as a necessary result of changes in gain. And, with the good foresight shown in other areas, the default high-quality ditering method in Audacity, triangle, is perfectly acceptable for most uses.
Let’s move on to another tab in the Preferences settings of Audacity. The File Formats tab allows you to set the quality levels of three file types: lossless, ogg and, if you’ve added the library, mp3. These are the formats to which Audacity will export your working or final files. Within the lossless dropdown (“Uncompressed Export Format”), you’ll find all the widely accepted lossless formats, including wav, aiff, au and flac. Your choice of lossless export format will be a matter of personal preference, as there’s very little difference in the sound of each. In my podcasting work, I’ve always chosen wav as the export format, as it’s a bit more portable than the others.
Better yet, Audacity offers options for the bit size of the exported files. These options include four, eight, 16 and 32-bit options for all the major types. Your choice of bit size will likely be determined by the type of podcast you’re producing. Music podcasts, where full fidelity and dynamics are the goal, are best exported as 32-bit files. Spoken-word podcasts, on the other hand, require less frequency response and a narrower dynamic range to begin with. The tradeoff of fidelity for smaller file size when using a 16-bit export format is pretty insignificant in a spoken-word podcast. That, by the way, is my personal pattern of choice. In my music podcasts, The Roadhouse and the IndieFeed Blues, podcast are exported as final masters in 32-bit wav format. The ClarkCast is primarily a spoken-word podcast, with a musical intro and outro. It’s exported in the 16-bit wav format. Of course, bear the previous discussion in mind if you choose a 16-bit format, especially for a music podcast. You’ll probably want to turn on real-time dithering to help “naturalize” your digital file.
Audacity also offers the ability to set the quality of exported ogg files and the encoding rate for mp3 files. If you have the mp3 libraries on your system but haven’t enabled them in Audacity, you’ll see a “Find Library” button in the preferences interface. Audacity expects the libmp3lame.so file, normally located in the /usr/lib directory on Linux systems. Once Audacity is pointed to the library, you’ll be able to export mp3 files directly. Again, exporting directly to ogg or mp3 will be a matter of preference and, likely, a function of the type of podcast you’re producing. In general, there’s quite a bit more tweaking to be done on music podcasts than on spoken word.
One final item in the file formats preferences tab. At the top of the tab, you’ll see two checkboxes. Both determine how imported uncompressed audio files will be handled. For example, if you’re importing a music wav file into Audacity, these options determine how that file will be handled during the import. The tradeoff in the selection of these options is one of safety vs. speed. The options allow you to either a) make a copy of the file being imported before actually importing it into Audacity or, b) copy directly from the file. The former is the safer option, while the latter is faster. As with so many of the other settings in Audacity, your choice is a matter of personal preference. As I’m usually tied up for eight hours or more with The Roadhouse, I lean toward speed in putting together the show.
The remaining tabs in the Preferences window are pretty standard fare. They don’t require any further understanding of the audio process, as they’re more housekeeping preferences than choices directly related to the quality of the audio. Two items in these tabs do bear some mention, though, for future reference in this series.
The spectrograms tab allows you to adjust the scaling and detail of the spectrograms. These spectrograms are a visual representation of the frequency spread of a file or file section. We’ll use the spectrogram later in determining the underlying noise level of your recording room. It’s also a useful tool for determining how much compression to apply to your audio file. In this option, the higher the number, the more granular the display of the spectrogram. I use the 2048 setting in Audacity for a high level of granularity. That allows me to drill pretty deeply into the characteristics of the file or sample in question.
The other option allows you to set the maximum frequency the spectrogram will analyze and graphically display. This is another option that should be determined by the type of podcast you’re producing. If you’re producing a voice-only podcast, the 8000 Hz default setting should be fine. That’s because 8000 Hz (or 8 KHz) is at the top end of the range of the spoken human voice. You shouldn’t need to analyze much beyond that with a spoken-word podcast. However, if you’re producing a music podcast, you’ll probably want to analyze to the top of the range of human hearing, rather than on the range of the human voice. The range of human hearing is approximately 20Hz to 20KHz, with recorded music following that range pretty closely. That musical range is what you’ll need to analyze, rather than the narrower voice range.
We have one final option to walk through in the preferences window. In the bottom half of the Interface tab, you’ll find settings for the maximum decibel (db) mode display range. Remember that we’re dealing with the amplitude of waveforms in Audacity – the difference between the maximum point of a waveform and a known point, usually seen as 0. The db level is a measurement of the power necessary to drive a waveform. In loose terms, an increase of 3 db in power will result in the smallest noticeable increase in volume. (It’s actually a more precise number than 3, but that will do for our purposes.) In other words, increasing the power from 96Db to 99Db will result in a perceptible increase in volume level. By increasing the mode display range in the Interface tab, you’ll be able to more clearly see these small increments. In effect, you’re setting the scaled height of the waveform in the recording interface. So as you increase the display mode range, your wave form will appear larger when recording.
In preparation for the next installment in the series, let’s take a quick run through the Effect options in the Audacity menu, addressing only the items you’ll be most likely to use. These tools allow you to modify the sound of your file post-recording, or to modify the sound of any files you may import into your podcast. This is also where you’ll find any add-in effects you may later install in Audacity, in the form of Nyquist plugins. In particular, we’ll look at Bass Boost, Compressor, Fade In and Fade Out. The options will only be available in the Audacity interface when you’ve highlighted a section of a file, or an entire file.
Bass boost provides the tools to increase the presence of a selected low-end frequency in your file. This adjustment may be on only a segment of the file, or on the file as a whole. By default, the frequency to be boosted is 200 Hz. This is a pretty wise default setting on the part of the Audacity authors, as a boost at 200 Hz nicely enhances the resonance of both male and female voices. The ability to control the amount of boost is also provided, by sliding the Db slider to the right or left. Remember that a 3 Db increase is the smallest perceptible increase in these levels. And, before committing to the change, you can listen to a quick preview in Audacity using the Preview button.
The Compressor setting will be important to you, especially if you’re producing a music podcast where the voice and music levels may vary considerably. In effect, compression minimizes the differences between the loudest and softest passages of an audio file. I’ll write in detail about the individual settings in the Compressor window in a later story.
Last, but certainly not least, the Fade In and Fade Out effects options also will be important in any type of podcast you produce. Their utility is pretty obvious for music podcasts – the ability to fade out of one music track and into another. But, they’re also useful in voice-only podcasts where a hard edit – a clean break from one passage to another – is too harsh. Fading makes it possible to blend a single voice, or multiple voices, into a seamless spoken-word passage.
We’ve covered quite a bit of ground in this installment of the “Attack of the Pod Penguins” series. And, as promised, you’ve picked up knowledge not only on the Audacity tool, but on some fundamentals of the physics of sound and sound recording, as well. We’ll build on that knowledge in the next installment, as we capture, edit, post-process and export files in Audacity.
