This effect combines the original signal with its delayed version, where the delay time has been modulated (this means that it varies constantly and the rate of its variations is defined by a function, for example, a sinusoid). Let's see what happens in such a situation if we consider a sinusoidal signal:
The upper waveform is the original signal which we will presume to be formed by a sinusoid and its second harmonic. The second waveform is identical to the first but has variable delay which at its highest is equal to half a semiwave. So let's now picture the second waveform oscillating on the horizontal axis between 0 and the position it has in the diagram. When it is at 0, the two waveforms are in phase and the result is a reinforcement of all the signal component frequencies. When it is in its delayed position (as in the diagram), we notice how the first harmonic (called the fundamental) gets cancelled and the second is reinforced. So, the frequency content of the original signal has been altered. All the intermediate positions act in different ways both on the first and second harmonic.
Summarizing: the phasing effect consists in summing to the original signal a delayed copy of itself in which the delay time is modulated by a certain waveform (if the delay weren't modulated we'd have static variation of the original signal's frequency-content. Through modulation the effect becomes more interesting).
The following is a sound to which a phaser effect has been applied.
We can simulate the phaser effect by using two microphones to pick up the same signal. By holding a microphone still whilst moving the other one cyclically backwards and forwards in relation to the sound source, what results is two copies of the same signal, one delayed compared to the other. The backwards and forwarding of the second microphone simulates modulation of the delay time.
The following diagram shows the logical scheme of a phaser:
We can see how the input signal is divided into two parts: the first part reaches the output without being manipulated whilst the second is passed through delay and then mixed with the input signal. Delay time is controlled by an LFO circuit (Low Frequency Oscillator). Such a circuit consists of an oscillator that generates low frequency waveforms, generally sinusoids (at 1 Hz or even less). These oscillators are usually used to control the parameters of other effects, as in this case, where the LFO modulates the delay time between the two signals (for example, if we modulate with a 1 Hz sinusoid, the two signals return in phase every second). We can see how one part of the signal routed to the output is picked up and sent back to the input. This artifice is used in many other kinds of effects and amplifies the effect even more.
These are the typical controls of a phaser:
Rate: the rate at which the delay time varies (the LFO modulator's frequency).
Mix: mixes the original signal and the manipulated signal.
Feedback: controls the amount of applied phasing.