Time-related sound processors

The advent of modern technologies has changed the expectations of sound quality. Today, more than ever, audiences are scrutinising "radio-sound". To meet these expectations, the recording and postproduction industries are becoming "all-digital". Already, stricter standards for audio quality and realistic (or hyperrealistic) productions are demanded of every link in the signal chain. Recording engineers and sound technicians are not exempt from this process. The more comprehensive the rig, the more difficult our work becomes.

The biggest part in the rig is occupied by sound processors. Sound processors are used to alter certain characteristics of sound.

• the spectrum-related processor such as the equalizer, affects the spectral balances in a signal;
  
• the time-related processor, such as reverberation or delay device, affects the time interval between a signal and its repetition(s);
• the amplitude-related processor, such as the compressor-limiter, affects a signal's loudness, and
• the noise-related processor, such as the Dolby system, does not really alter a signal but makes it clearer by reducing the noise added in analogue recording.

The invention of phasing cannot be credited to digital effect equipment. Phasing actually originated from pop music, played during the 1960s. At that time the effect was created manually, using three tape machines (with separate recording and reproducing heads). Machine 1 fed the signal to Machine 2, which was recording simultaneously. The playback from both was fed to machine 3, while playback of one machine was delayed, usually by applying some pressure to the supply reel to slow it down. The result: the music reached the reproducing head of the manipulated machine about 0.2 to 2 milliseconds later. This is still the time base for phasing effects in digital effect equipment. The final result of this kind of a phasing session was recorded on machine 3. The sonic result is a kind of swishing sound having many uses in the creation of special effects for music recording and sound effects.

Acoustical principle of phasing and flanging

Fig. 1
Two sinewaves superimposed, at different phase relations

Phasing can be either "positive", meaning that direct and delayed signals have the same polarity, or it can be "negative", meaning that these two signals have opposite polarity. Negative phasing can create strong, sucking effects like a sound being turned inside out. For example: a phaseshift of 180° of a 500Hz frequency is equal to 1ms. If you send a piece of music through a fixed delay of 1ms, and afterwards add the output of the delay unit to the original music, you create a tone quality which sounds flat at 500Hz (180° phase shift has cancelled amplitude at 500Hz). For the frequencies which are a little bit lower or higher than 500Hz, you will hear an increase of amplitudes, caused by a phaseshift other than 180°.

Typical phasing or flanging is not done with a fixed time, but with a varying delay time between 0.5ms and 2ms. The average value is at about 1ms. The result: with a continuous change other frequencies are always affected by either a more or less strong decrease or increase of intensity (this effect is also known as a "comb filter effect"). One effect which gets close to the phasing/flanging effect is a slowly moved WAH-WAH pedal or a parametrical equaliser with fully decreased gain for the middle frequency range, and a continuous change of the central frequency.

In practice, phasing and flanging cannot be divided into two different effects. Theoretically the phaser is working with a shorter delay time (0.1ms to 2ms) than the flanger (1ms to 10ms). According to this definition a phaser modulates more trebles. One could say that the sound is somewhat more "flowery".

Fig. 2
Time-related processors in relation to delay times

Parameters

The first parameter to be mentioned is the average delay time. It indirectly defines the frequency range, which is mostly affected by the decrease or increase in amplitude while phasing or flanging. To achieve phasing, this average delay time has to be modulated. Usually a so-called Low-Frequency-Oscillator (LFO) with an adjustable frequency (rate) is used, which starts to modulate the average delay time from 1Hz up to 10Hz (LFO-speed). Some phasers can even specify the LFO-waveform: sine or triangle.

Another parameter is the depth of the delay time modulation, which sets the phaser effect depth. In addition, some equipment controls the amount of effect via the input signal's amplitude. For example: an increase in volume of the input signal results in a higher intensity of the effect. This parameter is called resonance. The final parameter is the feedback (or regeneration), where the output signal of the phaser/flanger is fed back to the input of the device.

Guideline

In some cases, pure phasing or flanging effects can be used (no dry or direct signal mixed in the signal path). Be aware that 2ms of delay may already destabilize a certain groove. Mostly this problem appears when the beat comes from either a bass guitar or acoustical drums. Softer sounds with smoother impulses are less affected.

The flanger is followed by the chorus, which has longer delay times (5ms to 20ms). The chorus creates a kind of "ensemble" effect, which creates the effect that more musicians are playing a piece of music than actually are. The chorus uses an effect created in an ensemble where musicians start playing at slightly different times, and with different intonations. Pure chorus sound is very "indirect". Delay times are longer than flanger delay times and can only be used in a mixture with dry or direct sound. Timing would otherwise be very problematic.

Parts of chorus delay times are also covered by double tracking delay times (15ms to 50ms). Doubling can be done live by recording one track and then repeating the performance on a separate track in synchronisation with the first track (also called voice doubling effect). As it is not possible to repeat the same performance exactly, variations in pitch, timing and room sounds add fullness or openness to the sound. The delays are like early sound reflections and lend a sense of ambience to flat-sounding instruments or voices. A continuation of doubling would produce chorus sounds.

The terms echo and delay are often interchanged. Both terms are correct. For a long time, echo was produced with tapemachines. The speed of the tape and the distance between the recording and playback heads defined respectively the time or the length of the echo. Multiple echos were often produced by using echo machines with a tape-loop and several recording and playback heads. When the tape was running fast enough and the heads were close enough together, one could even produce a double tracking effect, although without any delay time modulation (unless the machine had a problem with wow & flutter).

Slap echo was also very popular - especially for Rock'n Roll, where the sound"slapped" back (ping-pong effect) with a delay time of about 30ms to 150ms. A loss of high frequencies due to the recording medium was typical for an echo created with tapemachines. The more echo-repetitions/feedback, the higher the treble-losses, but the sound was more natural as found today in some echo programmes, where even after so and a number of echoes, the high frequencies are still present.

The lowest delay time for echo starts at the point where one can achieve a separation between direct and "reflected" sound. This depends very much on the type of sound. Echo for percussive impulses can already be distinguished at around 25ms, while for string-instruments the delay time may go up to 500ms, 800ms or more. There is no limit for echo delay time in terms of sound, but there is in terms of costs: delay time = memory = costs. Processing equipment with 2000ms delay time is more expensive than equipment with a maximum of 300ms.

Analogue versus digital delay

The attributes of analogue and digital delay differ. Each has its advantages and disadvantages. Analogue units are less expensive and warmer sounding. They are especially good at creating sound effects, particularly flanging effects. However, a significant problem is that with delays of about 80ms and longer, the signal-to-noise ratio drops dramatically. Also, the longer the delay time, the worse the high-frequency response becomes. Digital delays are generally more expensive than analogue, but they can offer better signal-to-noise ratios, less distortion, and improved high-frequency response with higher delay times.

reverb / early reflection / gated reverb / reverse reverb / pitch

The first rooms used to create artificial reverb for studio productions were bare cellars, equipped with one loudspeaker and a microphone. The takes of a production, which had to be mixed with reverb, were reproduced via the loudspeaker. The microphone picked up the reverb. Later, in stereo technique, two microphones were used. Finally, the reverb was mixed with the original sound. This procedure was quite common until the 1970s.

During the 1960s, two electro-mechanical systems for reverb production were developed. The reverberation plate, a metal plate of about 2 square meters, which - in connection with a contact loudspeaker - started to vibrate. Two contact microphones picked up the reverb signal. An infinitely variable insulating panel made it possible to adjust the reverb time between 1 to 4 seconds.

The reverberation spring is still used in portable guitar amps. The principle was similar: the emitter of the reverb signal was installed on one end, the receiver on the other. Producing stereo reverb was also possible by using two separate springs. The springs manufactured for professional studio productions were certainly of much better quality than those used in guitar amps.

Reverberation plates as well as reverberation springs are no longer manufactured. The typical sound has nevertheless been rebuilt in digital effect processors. The digital reverberation programmes calculate the reverb according to three main criteria: natural rooms (halls, churches, cellars, toilettes, etc.), reverberation plate and spring.

Reverberation and early reflections

Many acoustical aspects have to be considered when precisely reproducing characteristical structures of reverberation. For example, a measurement of reverberation in natural rooms shows that before the actual reverberation starts, early reflections appear. These early reflections are important for defining the correct room-size.

Fig.3
The listener in a room with a source of sound. First, direct sound reaches the listener, then early reflections and finally late reflections or reverberation (circle).

Fig. 4
Typical structure of reverb in natural rooms. The first reflection is delayed in relation to the direct sound. The pre-delay time is adjustable. In this example, the reverb can also be delayed by adjusting the reflection/reverb delay. If this delay is negative, the reverb will be shifted to the early reflections.

In Fig. 4, we find the pre-delay before the early reflections. After the early reflections comes what is generally known as reverb. The reverb itself actually made of many reflections, as well. Their density is so high that the human ear cannot distinguish the sound as single reflections. The density itself depends on the construction of the room. A rectangular room with six even surfaces/walls produces less diffusion than a room where the surfaces are much more interlocked. Normally this effect is found in digital reverberation equipment under the parameter "diffusion".

Listening to reverb in a natural room, the human ear cannot distinguish between early reflections and the reverb. This distinction would only be possible by artificially constructing a room, as is done in effect processors. The processors even offer early reflection programmes with no reverb following at all. Early reflections sound more concrete, but less "romantic".

Non-linear reverb

During natural decay, the reverb fades down from the highest amplitude to zero. In effect programmes is not always the case. The best example is the gated reverb, a combination of reverb and a gate.

Fig. 5
Parameters used to adjust a gated reverb are: threshold, attack, hold, and release times. Hold and release times define the reverb length. The shorter the release time, the more the reverb is "cut off".

For the use of a gated reverb care must be taken in the relation between hold and reverb times. If the reverb is shorter than the hold time, the gate will be of no use. Depending on the angle of the release-time, the effect sounds either softer or harder.

Close to the gated reverb is the sound of the reverse reverb. The intensity of the reverb increases over time and at the end suddenly drops to zero. Gated and reverse programmes use the same parameters as linear reverb programmes, as with pre-delay, diffusion, etc.

A pitch shifter, in its function as a time compressor, can shorten audio material on film, video, and audio tapes with no editing and no alteration in frequency. Many broadcasting stations time-compress recordings to gain additional time for more music or advertisements. Commercials are often time-compressed to include more information without exceeding the contracted time. Time-compression also comes in handy when timing sound to images. If the nonsynchronous audio is longer than the video by a few seconds, it is possible to shorten the audio to the appropriate length without having to edit manually or electronically.

Time-related sound processors