top of page

Chapter 3 Contents

Digital Signal Processing

Digital signal processing (DSP) is one of the most important aspects to generating inventive sounds and inspiring film cues.  DSP and or analog processing can add great depth and scope to audio bringing it to a higher level of spatial and sonic appreciation.  It is crucial that students develop excellent facility with DSP so that their recordings stand out among others.


DSP processing alters some characteristic of a sound–timbre, envelope, space, and time.  They can be grouped into four categories:


• Spectrum—equalizers

• Time Processors—reverb, delay, time compression/expansion, 

   pitch shifting

• Amplitude—compressor and limiter (equalizer, de-esser)

• Noise Processors—Dolby and DBX


While a wave can be processed digitally by these DSP engines, it is best to use them in a side chain or in parallel to gain more control should the dynamics change as other sounds are added.

Illustration 3.21 Four categories of digital and analog signal processing.






One can boost or cut (peak or dip) or shelving a level of a signal at a specific point in the spectrum.  Generally, the frequencies are in full, half and third octave intervals.  There are four types of equalizers (EQs):  Fixed-frequency, Graphic, Parametric, and Filters (high, low, band-pass and notch)


Fixed-frequency operates with two or three (high, middle and low) set frequencies.  Standard consoles have concentric controls whereby the outer ring chooses the frequency and the inner ring increases or decreases level-boost or cut.  They have a pre-set bandwidth and the center control also affects adjacent frequencies dependent 

on the bandwidth of the device.  Stereo systems typically have fixed FQ EQs–a treble and bass.  Once the built in FQ maximum is reached, shelving, or flattening out, occurs so there is no further change to the sound.


Graphic equalizer is a type of fixed-frequency equalizer except that it features sliding instead of rotating controls and offers greater control than a two band fixed EQ.  They do this by attenuating–reducing FQ levels that have been chosen. The positioning of the sliders gives a graphic display of the frequency curve.  Generally, they come in 3 – 32 selectable frequencies.  


The parametric has continuously variable frequencies and bandwidths.  They provide greater flexibility and control of equalization.  A parametric equalizer combines fixed sliding controls with selectable parametric equalization.  One chooses the FQ and how much to attenuate or increase it.


Illustration 3.22  DSP Equalizer in Logic Pro X






Filters attenuate certain bands of frequencies.  They affect all frequencies above or below and determined by your Q setting.  Q determines how many FQ beside the chosen FQ are affected in a bell shape curve.  One can choose the neighboring two or more that will incrementally have a reduced effect of the chosen FQ.  Variance in loudness is pre-set and usually steep.  The most common filters are:


Illustration 3.23  Logic Pro X’s Parametric EQ bands








Duncan Metcalfe


Illustration 3.24  High-pass filter in Logic Pro X—attenuates all frequencies below a pre-set determined frequency.



Duncan Metcalfe


Illustration 3.25  Low-pass filter in Logic Pro X—attenuates all frequencies above a pre-set determined frequency.



Duncan Metcalfe



Illustration 3.26  Band-pass filter in Logic Pro X—sets high and low cut-off points allowing the frequencies between them to pass.  Used for corrective purposes.



Duncan Metcalfe


Illustration 3.27  Notch filter in Logic Pro X—also used mainly for corrective purposes.  Can cut a very narrow band allowing for frequencies on either side to pass, i.e. hum can be corrected without affecting adjacent frequencies.



Duncan Metcalfe


Be careful that a desired FQ is not lost in the process when using pass filters.


Illustration 3.28 of a DSP Filter.  In this instance, a special envelope filter from Omnisphere 2 is used.







Duncan Metcalfe

Psychoacoustic Processors

They add clarity, definition and overall presence to a mix by adding harmonics and narrow-band phase shifting.  They can also bring out the quieter detail without affecting louder sounds.  These processors are not typically found on digital systems and are primarily used for mastering in the analog domain.


Time Signal Processors 

They affect the time relationship of signals.  They include:

• Reverberation
• Delay

• Pitch-shifting and Time-shifting


Reverberation is the random, multiple repetitions of a sound signal.  It adds spatial dimension.  Reverb DSPs have infinite possibilities for emulating room or hall size and the reflective characteristics of these spaces.  To produce reverb a ‘dry’ sound must be sent and returned ‘wet’.  Two types of analog reverbs–acoustic reverb chambers and plate reverbs can be emulated by DSPs.  It is recommended that when recording analog sound, one records them dry so that there is more flexibility when adding reverb.  Also, too much reverb can make a recording sound muddy–limit its use and add different amounts to different sound based on their already existing envelope.


Digital Reverb

A digitized signal is delayed and recycled for several milliseconds.  This is repeated many times a second with a gradual decrease in amplitude to produce decay.  It is possible to control ADSR to produce a customized reverb effect.  They can provide an almost limitless variety of reverberation effects.

Illustration 3.29  Logic Pro X’s ChromaVerb Reverb DSP



Duncan Metcalfe



Delay is the repetition of a sound that typically takes place at least 50 milliseconds after the original sound.  In other words, the time between echoes is perceptible.  Digital delay functions the same way as a tape delay.  The signal is sampled and passed from one capacitor to the next until the signal reaches the output.  Variance in speed and delay times can be selected.  These units are less expensive and have a warmer sound.  However, the longer the delay the more noise is entered into the signal.

Digital Delay—converts analog to digital information, stores, and converts back to analog (for selected period of time and duration).  They have better signal-to-noise ratios than analog delays and have an extensive array of effects.


Uses of Delay:  doubling—35 milliseconds of delay create a natural, open ambience for dead or dry sounds.  Chorus—is created by recirculation of the doubling affect.  This can make a signal voice sound like many.  Creates a false sense of stereo.

Slap back echo—35 milliseconds or more creates the perceived discrete echo.

Delay with reverberation—Some units allow for the delay of signal before it gets to the reverb making the sound more natural.

Illustration 3.30  Logic Pro X’s Stereo Delay Delay DSP



Duncan Metcalfe



Exercise 3.5


Import an audio file, or create one in Logic Pro.  Choose a type of reverb and add it to one of the effect (FX) sends on the track strip in the mixing window.  Now experiment with the DSP and discuss the resulting outcomes when changing the FQs affected, coupled with the filters selected.  Also, explore the default setting.  Now do the same with delay and then combine the two.


Flanging is a time-delay effect that produces a ‘swishing’ sound with the delay time from 0–20 milliseconds.  The sound is a series of phase cancellations creating fluctuating peaks and dips while phasing is similar except it uses a phase shifter instead of time-delay circuitry.  The peaks and dips are more irregular resulting in sound whose timbre resembles that of being played under water. 


Morphing is the seamless transition of one affect into another.  Most morphing devices are delay based and are considered to be time signal processors, i.e. moving from phasing to delay.  With today’s DSPs this affect can be easily done and would be a great addition to the variety of a sound.


Exercise 3.6

Import an audio file, or create one in Logic Pro.  Choose a type of DSP and add it as an track effects insert in the mixer window.  Now experiment with the effect parameters over time by drawing the changes into the automated track arrangement window.  Discuss the resulting outcomes.


Amplitude Signal Processors

Amplitude processors affect dynamic range and include compressors, limiters, and expanders.


Compressors simply compress the signal—its output increases slower than its input.  The compression ratio establishes the proportion of change between input and output, i.e. 2 to 1 ratio means that for every 2db increase in input the output increases by 1db.  Compression threshold is the level at which the compression ration takes effect.  There is no way of knowing which level works best for any given sound;  it is a matter for experimentation.  The attack time is the length of time it takes for the compressor to react to an incoming signal.  If it is short it can bring out the percussiveness of a sound, if it is too long it can overshoot the transient of a sound and cut off portions of the wave.  The release time or recovery time is the length of time it takes the compressed sound to return to normal.  Its purpose is to make variations in loudness caused by compression imperceptible.  Negative outcomes with use include:  

• Fast release with close compression equals a perceived louder signal.

• Too short of a release and too wide a ratio caused pumping.

• A longer release smoothens the fluctuation of a signal.

• Too long a release time creates a muddy sound.


Limiters are compressors whose output level stays at or below a pre-set point regardless of input level.  It puts a ceiling on loudness but can eliminate high frequencies.  


Compressors and Limiters can:

• Minimize wide changes in loudness.

• Compression can smooth the variations of attack with strong transients.

• Compressor can improve the intelligibility of a vocal.

• Compressor can reduce apparent noise if the ratios are close.

• Compression has the effect of moving a sound back in the mix.

• Limiting prevents surges in signals.

• The combination of compression and limiting can add power to a sound.



An expander increases dynamic range rather than reducing it.  They have ratios much like a compressor and are triggered when a signal falls below a set threshold.  It can be used as a noise gate to reduce or eliminate low-level noise from analog sources.

3.22 - dsp eq.png
bottom of page