All kinds of delay effects

What we’ve said about early delays can also illustrate the use of effects such as flanging, phasing and chorusing. For all of these effects, three things must happen: first, there must be a delay, second, it must be recombined with the original signal, and third, the delay time must vary. In the case of flanging and phasing, we have a special effect where the frequency of the comb filter is swept up and down by varying the delay time. Delays are very short: a 10 millisecond delay might be a suitable maximum, because it gets the fundamental frequency of the comb filter down to 100 Hz. (the delay time will equal the period of the comb-filter frequency). The minimum delay should be .05 milliseconds or so, which will bring the comb filter's frequency to 20,000 Hz. or the upper limit of audibility. The reason this effect is such a psychic bombshell is that it sonically mimics a situation where the wall is moving! (The psychologists call this cognitive dissonance - where our various senses give us conflicting information. It is a really useful tool for creating profoundly entertaining effects. In this case, while our eyes tell us that everything is stationary, our ears inform us that the environment is moving! Yeeha!!!)

Chorusing mimics something else: the sound of multiple voices. When you hear a sound and its room reflection, the amount of delay (and phase shift) remains constant unless (a) you move or (b) the source moves or (c) the wall moves or (d) you’ve ingested too much single malt scotch (that’s a joke - however, I’ve tested it and it is true!). This is because there is a fixed relationship between you, the source and the wall. The fixed quality of that relationship results in a pattern of delays that also have fixed phase relationships. When two or more voices sound in unison, however, they cannot vibrate in a “phase-locked” way. If the voices are oscillators, each with its own stable frequency, they will “beat” with each other with a constant “beat” rate. If they are musicians, they are constantly correcting and trying to hold exactly the same frequency, so they tend to drift around each other, always holding very close to the unison but never at exactly the same frequency. When a whole group of voices does this, it results in a very rich, vibrant and dynamically “happening” sound. The chorusing device simply alters the delay time randomly, and within a few milliseconds. Additional delays will mimic the sound of additional voices, and yield a richer chorusing effect.

It is possible to use a couple of delays, or even a single one, to effectively create the early delays to simulate a room around your multitrack sounds. The trick for using these effectively is to “think like a room.” Below is an illustration of an easily simulated room with side reflections only:

	+enlarge
	The simulated room for the first patch. The two delayed paths are not quite identical.

In this case the room has only two walls, left and right, and they are almost equidistant on either side of the direct track. Below are two basic patches for simulating this room. Take a look at the first patch.

	+enlarge
	Console flow for left and right reflection simulations. Auxiliary bus sends signal to delay lines which return to Left and Right stereo buses. Left and right delays should be slightly different. Delay lengths should be between 10 and 50 milliseconds and usually only a few milliseconds different. Level of delayed signals should be to taste, but typical 6 dB below direct signal. Longer delay return should have higher level to offset Haas effect pulling image toward shorter delay.

In this case, the direct sound (there can be many direct sounds, mixed, of course) is sent in mono to both stereo busses, and is also sent via the auxiliary bus to two separate delay lines. I typically start with delays of 30 and 35 milliseconds, but this is really an issue of taste, your own personal fantasy about the size of the room in which you imagine the music is happening, and the style and mood of the material you are recording.


You can also use equalization to enhance the character of the delays and implied ambience: high-frequency boost will add a hard, bright edge to the ambience, suggesting hard, reflective surfaces with no diffusion, and low-frequency boost will add a sense of spaciousness and envelopment.

In any case, the delays should be almost the same, but not quite, and the level of the longer delay should be greater, to offset the Haas effect, which will tend to pull the image toward the channel with the shorter delay. The overall level of the delayed sounds should be 6 to 10 decibels below the direct sound, for three reasons. The first is simply that in reality (if we “think like a room”), reflections, which have traveled further and attenuated more, will typically be 6 to 10 dB lower in level than direct sounds. Second, the effect can be a little overwhelming if it is overdone. Finally, the comb filtering problems will be reduced as the difference between levels increases, which means that you are going to run into less trouble when you sum the mix to mono. I find that this patch works really well on the lead vocal, maybe all the vocals, the lead instrument, and the electric bass. It gives a richness and presence to these lead parts that is quite compelling. I also love to “wake up” synth sounds with this kind of treatment. They usually get an awful lot better with a couple of judicious delays like this.

Below, I’ve shown a variation on this patch that only requires one delay line:

	+enlarge
	Alternate console flow for left and right reflection simulations. Multitrack equivalent of Middle-Side stereo microphone setup. Auxiliary bus send signal goes to single delay line which returns to both Left and Right stereo buses. Right channel return has signal inverted (polarity reversed). Left and Right return levels should be identical. Delay lengths should be between 10 and 50 milliseconds and usually only a few milliseconds different. Level of delayed signals should be to taste, but typical 6 dB below direct signal.

This patch mimics the Middle-Side stereo microphone placement, where the delay line serves the same function as the side microphone. It has a couple of advantages, and one big disadvantage (which causes me to not use it very much). The advantages are that it requires only one delay line, and it is absolutely mono compatible, which is also its big disadvantage. When the stereo signal is summed to mono, the left and right returns (simple polarity reversals) will cancel each other out, so the effect of the delays is completely removed in mono, which usually really changes the musical character and impact of the mix. On the other hand, the patch is very effective, and, because there are no time differences between left and right buses, balancing it is a cinch.

Naturally, if you have more delay lines, you can set up many rooms or ambiences for different instruments in your mix. Our ears are very tolerant of diverse information, and will cheerfully accept a whole bunch of overlaid ambiences and reverbs without much trouble or confusion. You can also mix dry and wet sounds simultaneously without much trouble and often with great effectiveness. This is an area where there is tremendous potential, and a great deal of experimentation is called for.

Have fun, but don’t delay too long (that’s another joke)!

COMMENTS

Post a Comment