Cheat Sheet: DSP Functions

Cheat Sheet delivers concise, explicit information on how to do specific recording/audio-related tasks. This installment describes the types of DSP you’ll find in typical waveform editors and DAWs.

Allows reducing or increasing the file’s overall level. There may also be options that determine what to do if you set a level that would cause clipping, such as adding limiting or allowing soft clipping (saturation).

Takes audio that’s in the clipboard and mixes it in at the selection point in a file. Usually you can adjust the level of the signal being mixed in.

Lets you specify a maximum level that a file’s highest peak will attain. Typically this is 100% of maximum level, but most normalize functions let you normalize to any arbitrary level. For example, normalizing cuts intended for CD to –0.01dB of full scale will make sure that no signal hits 0, which would likely be interpreted by the CD duplication facility as being distortion.

This is similar to peak normalization, but lets you specify a particular maximum average level. This is sometimes applied to multiple cuts so they all have the same average perceived level prior to creating a CD. However, you still need to use your ears as the final judge of whether an operation like this accomplishes what you want.

This may be part of the Peak Normalize function, or a separate function. It finds the highest level in the file and inserts some type of marker, or places the cursor at this point.

Displays the current duration, and allows you to enter a new duration referenced to time and possibly tempo. You may have a choice of algorithms to improve the quality of the stretching, e.g., optimized for voice, optimized for beats, etc. Try all of them when stretching to determine which sounds best.

There are usually two ways to change pitch, one that doesn’t preserve duration (i.e., transposing up shortens the duration, and transposing down lengthens it) and one that does. The latter is more likely to add artifacts to the sound, particularly with relatively significant pitch changes. As with changing duration, you may have a choice of algorithms to optimize the final sound quality.

Flips the signal so that the file starts playback at what was the end, and ends playback at what used to be the beginning.

Flips the signal’s polarity. In other words, the positive peaks become negative peaks, and the negative peaks become positive peaks.

This changes the sample rate without changing the pitch or duration. For example, if you’ve mixed to a 96kHz file, you’d convert the file’s sample rate to 44.1kHz before trying to create a CD. Note that not all sample rate conversion algorithms are created equal, and some algorithms sound better to some people.

This fades the signal in or out over a selected region. You will probably have a choice of curves, or a way to draw a specific curve shape for the fade.

This lets you convert a stereo signal to mono, or mono to stereo. When converting to mono, you’ll usually able to specify whether you want an equal mix of left and right changes, or a different balance. Converting mono to stereo generally places the mono signal at equal levels in the left and right channels, but again, you may be able to change their relative levels.

Places the right channel in the left channel, and vice-versa.

Invoke this to make sure that the zero-crossing point of a waveform is actually at 0 instead of some other value.

You can think of this as being similar to a noise gate. If you specify a particular threshold, audio that falls below this threshold may be converted to silence (no signal at all), reduced by a certain amount, or deleted, depending on how the editor implements this function.

Inserts an arbitrary amount of silence at the insertion point in a file. The length may be based on a time you enter (e.g., 1 second), or on a region you defined. If you define a region, that region will be converted to silence, and the audio to the right of the selected region will be “pushed” further to the right to make room for the silence.

Allows superimposing an amplitude envelope on the file. Usually this is a line where you can add “nodes,” then drag the nodes around to create the desired envelope shape.

Sections of the file that occur before or after a selected region will be discarded, leaving only the selected region.

Sets a loop start and loop end point within the file. Upon reaching the loop end point during playback, playback continues from the loop start point, plays through to the loop end point, then returns to the loop start point, ad infinitum.

This is like looping, but upon reaching the loop end point during playback, playback reverses and plays backward until reaching the loop start point. Playback then reverses again toward the loop end point; the looping continues back and forth indefinitely.

A portion of the file immediately after the loop end is mixed in just after the loop start and faded out, or a portion of the file prior to the loop start is mixed/faded in just prior to the loop end, or both, depending on how crossfading is implemented. This creates a more seamless loop as there’s no abrupt transition between the start and end points of the loop. You will generally have the option to determine the duration of the pre- and post-loop material that’s mixed in.

This multiplies the spectrum of one piece of audio with another—typically what’s in the clipboard with the file having the focus. The resulting sound therefore has elements of both sounds, but doesn’t sound like either one.