Sound Generator

The Sound Generator is a tool to explore fundamental aspects of acoustics and of voice science by means of reconstructing sounds from scratch. It also allows you to generate synthetic sounds for calibrating measurements.

The workflow is to adjust some settings, and then to click on the Generate button to add the new sound to the current recording. To get started, click on the Saved Profiles list on the right, pick one of the predefined profiles, and click on Generate. Try out the different predefined profiles to see what kind of sounds they generate.

Fundamental and Overtones

Sound Generator - Fundamental and Overtones

Figure 3.11. Sound Generator – Fundamental and Overtones


The fundamental is the basic waveform from which the generated sound is composed. In can be one of the following:

Sine Wave

Your sound will be composed of one or more sine waves added together.

White Noise

Your sound will be pure white noise. This fundamental sound type has no additional parameters.

Coherent Noise

Coherent Noise is somewhere between sine waves and white noise. It is a waveform that is more irregular than a sine wave, but still has a frequency.

White Noise (Breathiness)

This setting allows to mix a percentage of pure white noise to the other fundamental waveform types, which can be used to simulate the breathiness of a voice, or the background noise of a recording.


This section determines if the pitch stays constant over the duration of the sound, or if it moves, for example by forming a triad or a scale.

When the pitch type is not constant, the pitch will move from the from to the to note or frequency. You can check return to start to make the pitch move back to the starting note.


Pitch type can be one of the following:

Constant Pitch

The pitch will be constant over the entire duration of the sound.

Single Step

The pitch will smoothly move from the start to the finish note.

Major / minor triad

The pitch will form a major or minor triad.

Major / minor scale

The pitch will form a major or minor scale.

Sweep (linear)

The pitch form a straight interpolation along the linear scale.

Sweep (log)

The pitch form a straight interpolation along the logarithmic scale.


This setting defines if the pitch changes gradually from one tone to the next (high smoothness), or with a more sharply defined step (low smoothness).


This section determines the number, spacing, and intensity of the overtones of the sound, which will shape its timbre.


The number of overtones can be anywhere from one (which means the sound only has the fundamental) up to the maximum that is determined by the sampling rate of the sound. In digital audio, a waveform can represent frequencies of up to half the sampling rate. For example, with a sampling rate of 44100 Hz, the highest frequency that can be represented is 22050 Hz. Therefore, overtones of up to a maximum of 22050 Hz can be included in the sound.

Specify amount

Enter the amount of overtones. Overtones above the maximum frequency will be ignored.

Up to maximum

Include all overtones below the maximum frequency, which is half the sampling rate.

Up to frequency

Include all overtones below the entered frequency (or half the sampling rate, whichever is lower).


The spacing of the overtones determines the distance or the ratio between successive overtones. This allows to create sounds that are based on the natural harmonic series, as well as other sounds that have different frequency ratios between the overtones.

Harmonic Series

Each overtone will be a whole multiple of the fundamental. For example, if the fundamental is 100 HZ, the overtones will be 200 Hz, 300 Hz, 400 Hz, 500 Hz, etc. In this case the overtones will be the harmonics of the fundamental.


Each overtone will have twice the frequency of its predecessor. For a fundamental of 100 Hz, the overtones will be 200 Hz, 400 Hz, 800 Hz, 1600 Hz. In this case the overtones will all be octaves of the fundamental.

Multiply by

Each overtone will have the frequency of its predecessor multiplied by the entered factor. If the factor is 2.0, this is the same as selecting Octaves.

In mathematical terms: frequency(n) = factor*frequency(n-1)

Fundamental factor

Each overtone will have the frequency of the fundamental, multiplied by a factor and by the number of the overtone. If the factor is 1.0, this will generate the harmonic series. In other words, the factor sets the distance between consecutive overtones.

In mathematical terms: frequency(n) = n*factor*fundamental


This section determines the amplitude, or intensity, of successive overtones. The values are measured in decibel (dB). A value of -6 dB corresponds to a reduction of the intensity by half. If the dropoff by octave or overtone is zero, all overtones will have the same intensity. If it is negative, higher overtones will be quieter than the lower ones (this is usually the case in natural sounds).

Tilt (dB/Octave)

The tilt sets the intensity drop per octave. With a value of -6 dB, each octave will be half the intensity of the previous octave.


This sets the intensity drop per overtone. Since the overtones follow a linear frequency scale, this setting will create a linear intensity dropoff (unlike the tilt, which will create a logarithmic dropoff).

Half intensity / Octave

With this setting, the intensity will drop exactly in half with each octave, which is equivalent to having a tilt of -6,0206 dB per octave.

Each overtone has an intensity that is one over N, where N is the frequency of the overtone divided by the frequency of the fundamental.

In mathematical terms: intensity(n) = 1/(frequency(n)/fundamental)

Apply Resonances

When this is checked, the generated sound will be filtered through the resonances currently defined by the vowel chart and the Resonances part of the Sound Generator. This is intented to simulate sounds made by the human voice.


Sound Generator - Modulation

Figure 3.12. Sound Generator – Modulation

The modulation page offers several ways in which the generated sound can be changed. In particular, it has ways to affect (modulate) the intensity of the sound, and its frequency.

The words intensity, magnitude, amplitude, and loudness, all refer to the same thing, which is how loud the generated sound is. Each term has a slightly different meaning. When referring to individual points in time of a waveform, amplitude is the maximum deviation of the waveform away from its center, so it can be positive or negative. Magnitude is the absolute value of the amplitude. When referring to a longer waveform rather than individual points in time, amplitude and magnitude are often used synonymously, and in that case they both mean the absolute value of the maximum deviation from the center.

Intensity is a synonym for magnitude.

Amplitude, Magnitude, and Intensity can be measured in linear values, or in decibel. A value of 0 dB is the maximum intensity that can be represented in a digital waveform. It corresponds to a linear amplitude of 1.0. A value of -6 dB corresponds to a linear amplitude of 0.5, a value of -12 dB corresponds to a linear amplitude of 0.25, a value of -18 dB corresponds to a linear amplitude of 0.125, etc.

Loudness is a subjective measure that refers to the how loud a sound is perceived as by the listener. This depends on many factors such as the used speakers, their volume settings, and the ears of the listener. Therefore the software cannot say much about the perceived loudness of the sound, but it can indicate the relative intensity of the sound in qualitative terms, and it can indicate the mathematical magnitude of the sound.

The frequency affects the pitch of the generated sound.


This setting determines the overall intensity of the generated sound.

Normalize after filter

The final intensity of the sound will be normalized to the entered value after the resonances have been applied.

Normalize before filter

The generated sound will be normalized before it is filtered through the resonances. This can be used to compare the relative strength of different resonance settings. If the initial sound is too loud, this will generate clipping, so start with very low values, such as -40 dB.

Frequency Modulation (FM)

This setting enables a periodic change of the fundamental frequency (or pitch), for example to create a vibrato effect in singing.

FM Frequency

The frequency of the pitch modulation. In singing, typical values of a singer’s vibrato are around 6 Hz. That is the frequency with which the pitch of the fundamental will get faster and slower.

FM Amplitude

This is the amount in cent by which the fundamental varies. A value of 100 cent will mean that the fundamental pitch will go up and down by one semitone.

Frequency Modulation (FM)

This setting enables a periodic change of the intensity of the generated sound.

AM Frequency

The frequency of the intensity modulation.

AM Amplitude

This is the amount in dB by which the magnitude of the generated sound varies. A value of 6 dB means that the intensity of the sound doubles with each oscillation.


This section allows to make each of the preceding aspects of the sound generation less regular by adding noise to it. This can prevent sounds that appear too “computerized” and unnatural by being too perfectly regular. The added noise is of the type “Coherent Noise” that can also be used as a fundamental waveform. It is not a truly random noise, but has a frequency and an amplitude.

To add noise to a parameter, select it in the Apply noise to field, and then define the properties of the noise generator on the right side. There is some overlap between the different ways of adding irregularity to a sound, because ultimately all parameters will either change its frequency or its amplitude. This can be done directly by adding noise to the fundamental, or indirectly, by adding irregularity to the FM or AM modulation.

Also, this section will add coherent noise to the generator, which is somewhat random, but still periodic. To add pure white noise, use the setting on the Fundamental and Overtones page.


The fundamental frequency of the noise generator. For human singers, a value of 5 Hz is a good starting point for many irregularities in the voice. You can also set values that change extremely slowly by entering frequencies such as 0.1 Hz, or even 0.01 Hz.


The range, in percent, by which the affected parameter will increase and decrease. If the target is a frequency modulation, this will affect the pitch, in cent. If the target is an amplitude modulation, the percentage value will affect the intensity.


The number of overtones for this noise generator. If it is set to 1, only there will be only one level of noise at the given frequency. With more than one overtone, higher frequencies of noise will be added.


This determines the spacing between overtones. More precisely, it is the frequency ratio of successive overtones. This setting is equivalent to the Multiply by setting for overtone spacing. A setting of 2.0 means that all overtones are octaves of the fundamental.

dB Delta

The intensity reduction, in dB, for each successive overtone. This is equivalent to the dB / Overtone setting for the overtone amplitude.


Sound Generator - Resonances

Figure 3.13. Sound Generator – Resonances

This section allows to pass the generated sound through a series of resonant filters akin to the vocal tract. Each filter has a frequency and a bandwidth that define the frequencies that this resonance amplifies / lets through.

This section of the Sound Generator is closely linked to the Vowel Chart and the Frequency Scale. You can create Resonance Path Nodes on the Vowel Chart, and you can change the frequencies of each resonance by moving the corresponding ruler on the Frequency Scale.

Simulate Vowel Adjustments

By default, the Generator will use the same resonances throughout the whole duration of the generated sound. However, the Sound Generator can also simulate how the resonances in the vocal tract change over time, for example when transitioning between different vowels, or when making more fine-grained adjustments to the current vowel when changing pitch to a different vocal register.

To get started with this, open the Vowel Chart, click on the icon, and then click on the Vowel Chart to create two or more nodes.

Now you can see the created nodes on the Resonance Path Timeline. Click on a node to be edited, and then adjust the resonances for this point in time. When generating the sound, the resonances will change over time according to the defined path. When you click on any point in time on the timeline (but not on a node), the current values for the resonances at this point in time will be shown.