Voice Grip
Content
AES 139 Design Competition – Undergrad
The Voice Grip
by James Pinkl
The Voice Grip
Abstract— The market for vocal performance effects is
currently minimal, and vocalists with an interest in
expanding their sound palette often resort to restrictive
alternatives. Guitar pedal rigs and laptops are powerful
tools to achieve new sounds, but they limit user stage
presence when repurposed for this application. The Voice
Grip is a finished project which allows performers to dial
in 3 real time effects wirelessly. Vocalists are now able to
maintain control of several standard audio effects without
being bound to bulky processing hardware.
I. DESIGN
The Voice Grip is designed to provide users with several
effects, all optimized for vocals and controlled wirelessly. It
starts with an interface that connects directly to the XLR
output of a dynamic microphone. It includes 3 potentiometers
and an XBee wireless module. The potentiometers correspond
to an independent effect and data on each potentiometer’s
position is transmitted wirelessly.
This data is then received by an XBee coordinator and fed
to an Arduino Due. The Arduino controls a parameter of each
effect based on the interface’s potentiometer postitions . The
input audio signal, bypassed from the interface, then gets
distorted, filtered, and delayed in real time.
Fig. 1
Fig. 2
Pre amp circuit simulated in LT SPICE
III. DIST ORT ION
The first effect implemented in the Voice Grip is distortion.
This circuit is a non-inverting amplifier with variable gain and
an asymmetrical diode network for clipping. Referring to
Figure 3, the unbalanced input signal is biased up to half
supply using a high value resistor. The feedback network
includes C2 to provide both unity gain at DC as well as a high
pass filtered output. An LM741 op amp was used to achieve a
thicker distortion sound [2]. Rheostat_1 is the first 10K ohm
digital potentiometer. Although most distortion circuits are
designed to achieve much higher gain (a maximum gain of 3
is particularly low), this behavior was intentionally avoided to
decrease the potential for feedback. Diodes with a low forward
voltage drop were also used (1N5817s with Vf = .3V). This
resulted in the desired distortion at lower amplitudes.
This stage is the only stage to feature a switch on the
interface. When the switch is activated, a relay gets flipped
and the diodes become part of the completed circuit. When the
switch is deactivated, the relay disconnects the diodes and
rheostat_1 becomes a short. The circuit then becomes a unity
gain buffer and the signal is not clipped.
Block diagram of the design
II. M ICROPHONE PRE A MP
It is important for the input stage to perform well since it is
common for dynamic microphones to have high output
impedance. The circuit used applies a gain of 1.8 before
converting the balanced signal to an unbalanced signal.
As the entire circuit operates on one 9V supply, it is
necessary for this stage (and all the stages that follow) to bias
up the signal to half supply. The pre amp design, shown in
Figure 2, ensured proper biasing and low noise performance
[1].
Fig. 3
Distortion circuit simulated in LT SPICE. Rheostat_1
adjusts the amount of gain.
IV. FILT ER
The distortion circuit is followed by a variable filter. This
allows the user to shape distorted vocals for a wide array of
timbres. The tilt topology, shown in Figure 4, is particularly
well-suited for this application as it operates by adjusting one
knob.
The circuit works by boosting and cutting bass and treble
bands. When one band is boosted, the other is attenuated.
Figure 5 displays the frequency response when
Potentiometer_1 (a resistance formed by the design’s second
10K ohm digital potentiometer) is at its minimum and
maximum values. A flat frequency response can be achieved
when Potentiometer_1 has equal resistance to Potentiometer_2,
which is another important benefit of a tilt circuit [3].
listening test was conducted to determine an appropriate
behavior of the effect. The delay time was experimentally set
to be about 200 milliseconds.
The mix control was done using a design based on a
summing amplifier, where the dry signal has fixed gain and the
delayed signal has adjustable amplitude. Through the use of
another 10K ohm digital potentiometer, the delayed signal
could be attenuated from unity gain down to no signal. This
stage also had to have a bias voltage appropriately applied.
Fig. 4
T ilt circuit simulated in LT SPICE. R_pot1 and R_pot2
adjust the amount of treble and bass present.
Fig. 6
Delay circuit simulated in LT SPICE. Both the input and
the outputRofpotentiomer
the delay are summed
via an op amp. R_pot 3 and
ADJUST
in diagram.
R_pot4 provide variable attenuation to the delayed signal.
VI. PROGRAMMING
Fig. 5
Frequency response of the tilt circuit at full bass and full
treble position
V. DELAY
The delay stage is the last audio effect and it utilizes the
PT2399 chip. This is a low noise, low distortion IC capable of
achieving delay times as high as 400ms [4]. Although multiple
knobs for regeneration and delay time can be considered
essential for a delay effect, a mix knob was determined to be
most important when limited to a single control.
The final circuit (Figure 6) was almost identical to the Echo
Application Circuit example in the part’s datasheet. The only
significant differences were altered capacitor values and fixed
resistors in lieu of the regeneration and delay potentiometers . A
To ensure smooth response to user input, both the XBee
and Arduino had to be programmed optimally. The XCTU
application was used to set up the two Xbees as a
coordinator and router on the same network. XCTU also
allows IO pins to be programmed as digital or analog
inputs. The router XBee utilized the 3 analog inputs and 1
digital input needed to create a fully functional interface.
The sampling rate was set up to be 25 samples per second.
The Arduino handled both the incoming XBee
transmissions and the digital potentiometer toggling. After
discarding bytes that were irrelevant to user actions on the
interface, the Arduino applied the data to a 256-position
AD55206 digital potentiometer. The SPI communication
interface was used for one way communication between the
master (Arduino Due) and slave (AD5206) [5]. While
prototyping, it was found that large increases or decreases
in digital potentiometer position resulted in audible pops.
Because of this undesired artifact, it was ensured that the
potentiometer position was updated by no more than 1
increment at a time.
VIII. CONCLUSION
The Voice Grip was a worthwhile design experience. The
project required applications of knowledge in all different
areas of electrical engineering. There were many takeaways
from this project.
The initial work was done entirely on a bread board. Many
of the circuits developed slight modifications over time and
that was based on the performance of the circuit on the
breadboard. Taking this approach was certainly beneficial, but
noise issues, specifically ground hum, were very common at
this stage.
Another challenge was writing the code for the Arduino.
There were some very audible pops that would occur if the
digital potentiometers did not transition smoothly. Writing
code to account for this was difficult. A zero crossing detector
circuit may be a very useful addition in a future build.
Fig. 7
Samples of the Arduino code, including the function for
SPI communication and the code for data interpretation from the
router XBee
Fig. 9
T he Voice Grip’s interface attached to a Shure SM58
Fig. 10
A prototype of the project on a breadboard
VII. LAYOUT
The final design was routed on a 70mm x 115mm 2 layer
PCB board. Clip leads were placed on the board to connect
with the external Arduino Due. DIP package ICs and SMD
parts of package size 0603 were used. The approach for
placement was to keep the audio path distant from the digital
pins on the AD5206. Ground and power planes were also
implemented to simplify the routing.
Fig. 8
Final design of the PCB
REFERENCES
[1] Gary K. Hebert, Designing Microphone Preamplifiers, 129th
AES Convention, 2010
[2] Brian Wampler, How to Modify Guitar Pedals, 2nd ed.,
unpublished, 2012
[3] Francesco Balena, Implement an Audio Frequency Tilt
Equalizer Filter
http://www.edn.com/design/analog/4368935/Implement-anaudio-frequency-tilt-equalizer-filter, 2012
[4] M erlin Blencowe, Small Time: A Delay / Echo with Tails
http://www.valvewizard.co.uk/smalltime.html
[5] Heather Dewey-Hagborg, Controlling a Digital Potentiometer
Using SPI
https://www.arduino.cc/en/Tutorial/DigitalPotControl, 2015
The Voice Grip
by James Pinkl
The Voice Grip
Abstract— The market for vocal performance effects is
currently minimal, and vocalists with an interest in
expanding their sound palette often resort to restrictive
alternatives. Guitar pedal rigs and laptops are powerful
tools to achieve new sounds, but they limit user stage
presence when repurposed for this application. The Voice
Grip is a finished project which allows performers to dial
in 3 real time effects wirelessly. Vocalists are now able to
maintain control of several standard audio effects without
being bound to bulky processing hardware.
I. DESIGN
The Voice Grip is designed to provide users with several
effects, all optimized for vocals and controlled wirelessly. It
starts with an interface that connects directly to the XLR
output of a dynamic microphone. It includes 3 potentiometers
and an XBee wireless module. The potentiometers correspond
to an independent effect and data on each potentiometer’s
position is transmitted wirelessly.
This data is then received by an XBee coordinator and fed
to an Arduino Due. The Arduino controls a parameter of each
effect based on the interface’s potentiometer postitions . The
input audio signal, bypassed from the interface, then gets
distorted, filtered, and delayed in real time.
Fig. 1
Fig. 2
Pre amp circuit simulated in LT SPICE
III. DIST ORT ION
The first effect implemented in the Voice Grip is distortion.
This circuit is a non-inverting amplifier with variable gain and
an asymmetrical diode network for clipping. Referring to
Figure 3, the unbalanced input signal is biased up to half
supply using a high value resistor. The feedback network
includes C2 to provide both unity gain at DC as well as a high
pass filtered output. An LM741 op amp was used to achieve a
thicker distortion sound [2]. Rheostat_1 is the first 10K ohm
digital potentiometer. Although most distortion circuits are
designed to achieve much higher gain (a maximum gain of 3
is particularly low), this behavior was intentionally avoided to
decrease the potential for feedback. Diodes with a low forward
voltage drop were also used (1N5817s with Vf = .3V). This
resulted in the desired distortion at lower amplitudes.
This stage is the only stage to feature a switch on the
interface. When the switch is activated, a relay gets flipped
and the diodes become part of the completed circuit. When the
switch is deactivated, the relay disconnects the diodes and
rheostat_1 becomes a short. The circuit then becomes a unity
gain buffer and the signal is not clipped.
Block diagram of the design
II. M ICROPHONE PRE A MP
It is important for the input stage to perform well since it is
common for dynamic microphones to have high output
impedance. The circuit used applies a gain of 1.8 before
converting the balanced signal to an unbalanced signal.
As the entire circuit operates on one 9V supply, it is
necessary for this stage (and all the stages that follow) to bias
up the signal to half supply. The pre amp design, shown in
Figure 2, ensured proper biasing and low noise performance
[1].
Fig. 3
Distortion circuit simulated in LT SPICE. Rheostat_1
adjusts the amount of gain.
IV. FILT ER
The distortion circuit is followed by a variable filter. This
allows the user to shape distorted vocals for a wide array of
timbres. The tilt topology, shown in Figure 4, is particularly
well-suited for this application as it operates by adjusting one
knob.
The circuit works by boosting and cutting bass and treble
bands. When one band is boosted, the other is attenuated.
Figure 5 displays the frequency response when
Potentiometer_1 (a resistance formed by the design’s second
10K ohm digital potentiometer) is at its minimum and
maximum values. A flat frequency response can be achieved
when Potentiometer_1 has equal resistance to Potentiometer_2,
which is another important benefit of a tilt circuit [3].
listening test was conducted to determine an appropriate
behavior of the effect. The delay time was experimentally set
to be about 200 milliseconds.
The mix control was done using a design based on a
summing amplifier, where the dry signal has fixed gain and the
delayed signal has adjustable amplitude. Through the use of
another 10K ohm digital potentiometer, the delayed signal
could be attenuated from unity gain down to no signal. This
stage also had to have a bias voltage appropriately applied.
Fig. 4
T ilt circuit simulated in LT SPICE. R_pot1 and R_pot2
adjust the amount of treble and bass present.
Fig. 6
Delay circuit simulated in LT SPICE. Both the input and
the outputRofpotentiomer
the delay are summed
via an op amp. R_pot 3 and
ADJUST
in diagram.
R_pot4 provide variable attenuation to the delayed signal.
VI. PROGRAMMING
Fig. 5
Frequency response of the tilt circuit at full bass and full
treble position
V. DELAY
The delay stage is the last audio effect and it utilizes the
PT2399 chip. This is a low noise, low distortion IC capable of
achieving delay times as high as 400ms [4]. Although multiple
knobs for regeneration and delay time can be considered
essential for a delay effect, a mix knob was determined to be
most important when limited to a single control.
The final circuit (Figure 6) was almost identical to the Echo
Application Circuit example in the part’s datasheet. The only
significant differences were altered capacitor values and fixed
resistors in lieu of the regeneration and delay potentiometers . A
To ensure smooth response to user input, both the XBee
and Arduino had to be programmed optimally. The XCTU
application was used to set up the two Xbees as a
coordinator and router on the same network. XCTU also
allows IO pins to be programmed as digital or analog
inputs. The router XBee utilized the 3 analog inputs and 1
digital input needed to create a fully functional interface.
The sampling rate was set up to be 25 samples per second.
The Arduino handled both the incoming XBee
transmissions and the digital potentiometer toggling. After
discarding bytes that were irrelevant to user actions on the
interface, the Arduino applied the data to a 256-position
AD55206 digital potentiometer. The SPI communication
interface was used for one way communication between the
master (Arduino Due) and slave (AD5206) [5]. While
prototyping, it was found that large increases or decreases
in digital potentiometer position resulted in audible pops.
Because of this undesired artifact, it was ensured that the
potentiometer position was updated by no more than 1
increment at a time.
VIII. CONCLUSION
The Voice Grip was a worthwhile design experience. The
project required applications of knowledge in all different
areas of electrical engineering. There were many takeaways
from this project.
The initial work was done entirely on a bread board. Many
of the circuits developed slight modifications over time and
that was based on the performance of the circuit on the
breadboard. Taking this approach was certainly beneficial, but
noise issues, specifically ground hum, were very common at
this stage.
Another challenge was writing the code for the Arduino.
There were some very audible pops that would occur if the
digital potentiometers did not transition smoothly. Writing
code to account for this was difficult. A zero crossing detector
circuit may be a very useful addition in a future build.
Fig. 7
Samples of the Arduino code, including the function for
SPI communication and the code for data interpretation from the
router XBee
Fig. 9
T he Voice Grip’s interface attached to a Shure SM58
Fig. 10
A prototype of the project on a breadboard
VII. LAYOUT
The final design was routed on a 70mm x 115mm 2 layer
PCB board. Clip leads were placed on the board to connect
with the external Arduino Due. DIP package ICs and SMD
parts of package size 0603 were used. The approach for
placement was to keep the audio path distant from the digital
pins on the AD5206. Ground and power planes were also
implemented to simplify the routing.
Fig. 8
Final design of the PCB
REFERENCES
[1] Gary K. Hebert, Designing Microphone Preamplifiers, 129th
AES Convention, 2010
[2] Brian Wampler, How to Modify Guitar Pedals, 2nd ed.,
unpublished, 2012
[3] Francesco Balena, Implement an Audio Frequency Tilt
Equalizer Filter
http://www.edn.com/design/analog/4368935/Implement-anaudio-frequency-tilt-equalizer-filter, 2012
[4] M erlin Blencowe, Small Time: A Delay / Echo with Tails
http://www.valvewizard.co.uk/smalltime.html
[5] Heather Dewey-Hagborg, Controlling a Digital Potentiometer
Using SPI
https://www.arduino.cc/en/Tutorial/DigitalPotControl, 2015
