Streaming Media Bible - Streaming Magazine Column

Streaming Magazine Column - August 2002


	Table of Contents

	Chapter Excerpt

	Updates

	Order the Book

	Resources

	Tips & Tricks

	Streaming Magazine Column

	Feedback

	Author

	Home

Audio Production - Fun with Signal Processing.

One of my favorite things about audio production is that audio is very malleable. Using techniques such as equalization (EQ) and compression, you can not only add extra sparkle to your programming, but also correct small imperfections that might otherwise compromise quality. These techniques are collectively known as signal processing.

Essentially all signal processing falls into two categories: Equalization, where the level of particular frequencies is adjusted, and Dynamics Processing, where the level of the entire signal is adjusted. These techniques turn dull, lifeless audio into professional-sounding programming. Most importantly, it gives your content the extra polish that makes it stand out above the competition.

Equalization (EQ)

Equalization changes how audio sounds by boosting certain frequencies and/or attenuating others. Turning up frequencies is known as additive EQ, for instance "warming up" a signal by adding bass. Turning down frequencies is known as corrective EQ. Corrective EQ can remove unwanted noise, such as hiss or rumble.

EQ can be done using hardware or software, using graphic or parametric equalization. Essentially, all you do is identify the frequency you want to adjust, grab the appropriate knob or fader, and boost or attenuate to your heart's desire.

Table 1 provides a rough guide to the frequency spectrum to give you an idea of where to start. To make sure you're working on the right frequency, try adding too much - this "pops" the frequency out of the mix and makes it easier for you to identify. Once you find the right frequency, add or subtract a moderate amount. Use your ears to decide how much sounds better - err on the side of caution.

Table 1: Useful EQ ranges

EQ Range	Contents
20-60 Hz	Extreme low bass. Most multimedia speakers and, in fact, most home speakers cannot reproduce this. Boosting here can make your audio sound muddy and lead to distortion. It is good practice to roll off these frequencies.
60-250 Hz	The audible low-end. Files with the right amount of low end sound warm, files without enough sound thin. Files with too much in this range sound boomy.
250 Hz-2 kHz	The low-midrange. Files with too much in the low-mids are hard to listen to and sound telephone-like. Sometimes cutting in the 250-350 Hz range can clarify your mix.
2 kHz-4 kHz	The high-midrange. Most speech information resides here, so you don't want to push this range in your background music. In fact, cutting here in the music and boosting around 3 kHz in your narration makes it more intelligible.
4 kHz-6 kHz	The presence range. These frequencies provide the clarity in both voice and musical instruments. In particular, boosting 5 kHz can make your music or voiceover (not both!) seem closer to the listener.
6 kHz-20 kHz	The very high frequencies. Boosting here adds "air" but can also cause sibilance problems.

Dynamics Processing

Dynamics processing involves adjusting the overall level of the signal. The most common form of dynamics processing is known as compression. Compression can be thought of as an automatic volume control that operates according to rules you set. Since last month's column was subtitled, "Level is everything," compression is the most powerful tool in your audio arsenal.

Compressors "even out" signal level by turning down the loudest sections of a signal. This can be particularly helpful when you have widely varying levels, for instance a panel with a number of different speakers. Since the compressor automatically turns down the loudest speakers, you can set the overall level higher, which in turn gives you a more professional sounding program.

Compressors have threshold, ratio, and output gain settings. The threshold sets where you want the compressor to begin attenuating the signal; the ratio sets how much attenuation you want; and the output gain enables you to turn your signal back up after the compression has occurred to maintain optimal signal level. Some sample compressor settings are listed in Table 2.

Table 2: Sample Compression Settings

Purpose	Settings
Light speech compression (general usage)	Set threshold in the -15 to -10dB range, use a 3:1 to 6:1 ratio
Heavy speech compression (radio)	Set threshold in the -20 to -12dB range, use a 5:1 to 10:1 ratio
Light musical compression (even out disparate levels)	Set threshold in the -10 to -6dB range, use a 2:1 to 3:1 ratio
Heavy musical compression (FM pop radio)	Set threshold in the -15 to -10dB range, use a 6:1 to 4:1 ratio

A quick glance at Table 2 shows that voice is much more compression tolerant than speech. People are accustomed to listening to heavily compressed vocals, because all radio broadcasts are heavily compressed. In general, you should use moderate compression settings to even out the levels of your programming for consistency. You can use heavy compression, but be careful - heavy compression changes the sound and feel of your programming. As always, use your ears and err on the side of caution.

Closing Reminder

When using signal processing techniques, be sure to keep an eye on the meters to make sure your signal doesn't distort or go too low. For analog equipment, this means peaks in the -3dB to 0dB range; for digital equipment this means peaks in the -10dB to -6dB range. Being careless with your levels defeats the point of signal processing.

Next Month: Video Production - It's All About Light