Grove Audio PMRESO User manual

PMRESO –PM RESONATOR™
3 CHANNEL FORMANT FILTER
PRELIMINARY OPERATION MANUAL
April 11, 2016
Copyright 2016 - Grove Audio®
All Rights Reserved

Copyright 2016 –Grove Audio®
All rights reserved Page | 2
INTRODUCTION
The PM Resonator –3 Channel Formant
filter is a recreation of the formant filter in the
Polymoog™ polyphonic keyboard instrument.
The module consists of three resonant filters
with adjustable filter modes and Emphasis
(resonance) followed by a 3 channel mixer.
The filter modes are Low Pass, Band Pass,
Band Notch, and High Pass. All three filters
will have the same mode as set by the Mode
switch.
A toggle switch on the panel labelled Range
allows the ranges of the two lower channels,
LOW and MEDIUM, to be extended to cover
a better range of values for vocal formant
use. A second toggle allows the filters and
mixer to be bypassed so that the input signal
can be heard directly. This assists in
adjusting the signal characteristics so that the desired result can be achieved.
BASIC OPERATION
Simulate natural body resonances of musical instruments
Create human vocal formants corresponding to vowels
Serve as a simple parametric equalizer
The Frequency control (and the associated Range switch) are adjusted to the frequencies of the three
major partials of the resonance structure that is being simulated. Typically three frequencies are all that is
needed to give a good approximation of the body resonances of a typical instrument. For example, the
violin has a narrow peak at 300 Hz., another narrow peak at 700 Hz. and a broader peak at 3 KHz. The
relative amplitudes are 3.5, 3.5 and 10 respectively. This information can be obtained by searching for the
particular instrument to be simulated and “body resonance” or “formant”. The peak width is set on the
Emphasis controls and the relative amplitudes are set on the Gain controls. Typically, when using the
filter for formant simulation, the bandpass Mode setting should be used.
To get an accurate frequency setting, set the filter for bandpass, connect the input to an oscillator sine
wave output, set the oscillator to the desired frequency and then move the corresponding frequency
control on the filter back and forth until a peak in the filter output is detected. It may be necessary to
increase the emphasis control to help make the peak more prominent.

Copyright 2016 –Grove Audio®
All rights reserved Page | 3
FORMANT SYNTHESIS BASICS
Musical instruments and human voices are identified by the effect that the resonant cavity associated with
the body of an instrument or the vocal tract of a human has on the amplitude distribution of the harmonics
produced by the driving frequency source. The auditory perception of a human ear and brain is acute
enough to be able to make an identification of the type of instrument or the vowel produced in
vocalizations with only the first three formants being present.
Formants are fixed structures for any particular resonant system.
Discussion of difference between formants and harmonics.
This effect can be produced electronically by using three bandpass filters working in parallel with variable
frequency, gain and Q. Q is a dimensionless factor representing the ratio of the filter bandwidth measured
at -3dB power level (.7071 for voltage) to the center frequency. Formant Q’s are generally in the range
of 5 –50.
As an example, here are the first three formants for a violin. The fundamental frequency is 220 Hz.
F1 (Hz)
F1 (dB)
F1 (Q)
F2 (Hz)
F2 (dB)
F2 (Q)
F3 (Hz)
F3 (dB)
F3 (Q)
518
1613
2685
And here are the first three formants for a tenor voice (pitch range C3-C5) singing “a” as in “bob”. The
fundamental frequency is 120Hz.
F1 (Hz)
F1 (dB)
F1 (Q)
F2 (Hz)
F2 (dB)
F2 (Q)
F3 (Hz)
F3 (dB)
F3 (Q)
673
1097
2457
USE A HARMONIC RICH WAVEFORM
In order for formant synthesis to work effectively waveforms with sufficient harmonic content should be
used. Sine waves would not work well as the energy content of the wave is all concentrated at one
frequency. The best driving waveform for formant synthesis purposes is a sawtooth wave as this
waveform produces energy in all odd and even harmonics providing a source of energy to the resonators.
KEEP THE PITCH WELL BELOW THE FIRST FORMANT
If the driving waveform frequency is too high, there will not be enough harmonics that fall within the range
of the resonator filters to provide a realistic simulation of the formant. This effect becomes apparent when
simulating higher pitched vocal formants. The vowel being produced will be harder to identify if the driving
frequency approaches or exceeds the frequencies of the first or second formants.

Copyright 2016 –Grove Audio®
All rights reserved Page | 4
TECHNICAL SPECIFICATIONS –PM Resonator™
Low Frequency Range
Medium Frequency
Range
High Frequency Range
Input Range
Normal 60Hz - 300Hz:
Extended 180Hz - 900Hz
Normal 300Hz - 1.5KHz;
Extended 600Hz - 3KHz
1.5KHz - 7.5KHz
+-10 Volts max.
Input Impedance
100K Ohms
Module Power
+15 Volts 25 mA., -15 Volts 25
mA., +5 Volts 0 mA.
Module Size
Panel height - 8.75 inches (222
mm); Panel width - 4.250 inches
(108 mm); Depth (behind panel)
2.5 inches (96 mm)
PM RESONATOR “VINTAGE” MOD
If you ordered the vintage mod kit for your PM Resonator you will find a set of three op amps that can be
substituted for the TL074 op amps that are supplied installed. The LM348s are similar to the units
installed in the original Polymoog. You can remove the socketed op amps and substitute these to
experiment with different sound textures. Be sure to observe the correct orientation of the op amps
carefully matching the notch in the socket with the notch in the plastic body of the part. This diagram
illustrates the correct orientation.

Copyright 2016 –Grove Audio®
All rights reserved Page | 5