Edgar Berdahl∗ and Alexandros
Kontogeorgakopoulos†
∗Audio Communication Group
Technical University of Berlin
Sekretariat EN-8
Einsteinufer 17c
10587 Berlina
Alemania
eberdahl@mail.tu-berlin.de
†Cardiff Metropolitan University
Cardiff School of Art and Design
Llandaff Campus
Western Avenue
Cardiff, CF5 2YB, Wales
Reino Unido
akontogeorgakopoulos@cardiffmet.ac.uk
The FireFader: Simple,
Open-Source, y
Reconfigurable Haptic
Force Feedback for
Musicians
Abstracto: The FireFader is a simple haptic force-feedback device that is optimized for introducing musicians to haptics.
It is based upon a single-degree-of-freedom potentiometer fader coupled to a DC motor, also known as a “motorized
fader.” A light is connected in parallel with the motor to help communicate the force’s strength visually. The FireFader
consists of only open-source hardware and open-source software elements. Como consecuencia, it is relatively easy for users
to repurpose it into new projects involving varying kinds and numbers of motors and sensors.
An open-source device driver for the FireFader allows it to be linked to a computer via USB so that the computer can
perform the feedback control calculations. Por ejemplo, the computer can simulate the acoustics of a virtual musical
instrument to concurrently synthesize sound and calculate the motor force as a function of the fader position. El
serial connection over USB increases the latency of the control signal compared to embedded implementations, pero
the serial connection facilitates easier programming via the computer, and the force feedback can be automatically
disabled when the user is not touching the fader. Some new devices derived from the FireFader design are presented.
Introducción
Although traditional musical instruments provide
haptic touch-oriented feedback, this kind of
feedback is lacking in many digital musical
instruments. This is one reason why the computer
music community is interested in endowing digital
musical instruments with haptic feedback (Cadoz,
Luciani, and Florens 1984; Hayes 2012). Haptic
feedback can take many forms (Birnbaum 2007),
and one important form is force feedback, cual
can enable a user to interact kinesthetically with a
virtual mechanoacoustical system.
Overview
Considerable research (Cadoz, Luciani, and Florens
1993; Florens, Cadoz, and Luciani 1998; Cadoz et al.
Computer Music Journal, 37:1, páginas. 23–34, Primavera 2013
doi:10.1162/COMJ a 00166
C(cid:2) 2013 Instituto de Tecnología de Massachusetts.
2003) has been carried out on force feedback for
musical systems since 1978. This research has not
trickled down into systems that are widely available
to musicians, sin embargo. One of the main reasons
is that most prior force-feedback systems designed
for musical applications were too expensive for
many musicians. Por ejemplo, the force-feedback
devices from Ergos Technologies are designed with
the highest quality musical applications in mind:
The devices have a position resolution of 1 μm,
bandwidth of 10 kHz, and maximum force output of
hasta 200 norte (Florens et al. 2004). Certainly, it would
be appealing to teach workshops with them, pero
porque, they cost tens of thousands of U.S. dollars,
they are prohibitively expensive for most musicians
and artists (Florens, Cadoz, and Luciani 1998).
Force-feedback devices from Sensable Technolo-
gies are less expensive. The least expensive device
from Sensable is the Phantom Omni, which costs
approximately US$ 1,000. Regardless of cost, the devices from Sensable are designed with medical applications, such as surgery, in mind. A musician Berdahl and Kontogeorgakopoulos 23 l D o w n o a d e desde h t t p : / / directo . mi t . e d u / c o m j / lartice – pdf / / / / 3 7 1 2 3 1 8 5 5 8 0 9 / c o m _ a _ 0 0 1 6 6 pd . j f b y g u e s t t o n 0 7 septiembre 2 0 2 3 on stage might be less interested in sitting at a table and performing with a pen-like object. For instructional purposes, several universities have made simpler haptic force-feedback devices that are less expensive. The series of “Haptic Pad- dles” are single-degree-of-freedom devices based upon a cable connection to an off-the-shelf DC motor (Okamura, Ricardo, and Cutkosky 2002). Such designs are problematic, sin embargo, because of the unreliable supply of surplus high-performance DC motors (Gillespie 2003). The authors also contacted several companies selling DC motors with integrated sensors, but we could not find a low-price motor of this type that was always available new. Surplus motors can be obtained, por ejemplo, from stocks of old hard disks (Oboe and De Poli 2002; Verplank, Gurevich, and Math- ews 2002). Removing motors from disk drives and installing them into a haptic device requires significant manual labor, sin embargo. Además, using surplus motors frustrates the creation of a standardized device, both because different motors have different mechanical and electrical charac- teristics, and because they require different kinds of cables, connectors, and sometimes even control algorithms. New DC motors that incorporate a position sensor can be ordered from companies such as Maxon motors, but they usually cost more than US$ 200 cada. A diferencia de, the iTouch device
at the University of Michigan instead contains a
voice-coil motor, which is hand-wound by students
(Gillespie 2003). Making a large number of devices
is time-intensive, and the part specifications are
not currently available in an open-source hardware
format.
In prior work, we have used what was then the
least expensive commercial general-purpose force-
feedback robotic arm, the NovInt Falcon. Over the
past several years, its price has ranged between
US$ 100 and US$ 250. The Falcon is designed
primarily for gaming, so it is accessible to musicians
and is inexpensive (Berdahl, Kontogeorgakopoulos,
and Overholt 2010). We recently demonstrated a
musical composition for two Falcon devices at
the International Computer Music Conference in
Ljubljana (Berdahl and Kontogeorgakopoulos 2012a).
Conveniently, an open-source driver compatible
with Mac OS X, Windows, and Linux is available for
the Falcon (Berdahl, Niemeyer, and Smith 2009a);
the appearance and size of the device, sin embargo,
might be less appealing to some musicians. También,
the Falcon’s geometry cannot easily be modified
because of the complex interconnections of the
mechanical cables, circuit boards, and other parts.
These considerations have influenced the design of
the FireFader, which can be used in, and abstracted
a, a wide variety of configurations.
This article describes the FireFader design in
detail and is a revised version of an earlier conference
paper (Berdahl and Kontogeorgakopoulos 2012b).
Requirements
In consideration of prior work and our experiences,
the following requirements for the device design
have been derived and have guided the design.
In order to promote the sharing of program code,
the device should implement a new standardized
protocol for force-feedback control. To make the
device accessible, it should be compatible with Mac
OS X, Linux, and Windows. The device should be
relatively inexpensive, yet still perform well enough
for intriguing musical applications, including live
actuación. The device should also be controllable
via double-precision floating-point physical models
(Castet, Courouss ´e, and Florens 2007) and have
access to high-quality audio converters so that it can
be used in applications requiring high-quality audio.
The device should be simple enough that users can
understand, build, and reconfigure it. Además,
the device should consist only of open-source
hardware and open-source software so that users can
easily reconfigure it for other applications. Finalmente,
the device design and the demonstration programs
should inspire musicians to take an interest in
haptic force feedback.
Hardware Design
The hardware design focused primarily on meeting
the requirements while keeping cost low.
24
Computer Music Journal
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 1. FireFader
prototype (abajo) con
MIDI keyboard (arriba).
Motor
To significantly lower the cost, we decided to use
a motorized fader, which is a linear potentiometer
coupled to a DC motor. Motorized faders are mass-
produced by multiple companies, including ALPS
Electric and Penny+Giles, so the competition
tends to drive the price downward. Además,
motorized faders can be found in many audio mixing
consoles, so most professional musicians are already
familiar with them. Por esta razón, we assumed that
musicians might, on average, be more interested
in force-feedback interaction with a motorized
fader than with something more foreign-looking
like a haptic paddle. Además, during laboratory
exercises, this tendency has been noticed among our
estudiantes.
It should be mentioned that the idea of using
a motorized fader for musical applications is not
nuevo. Bill Verplank has maintained a stock of them
at the Center for Computer Research in Music
and Acoustics (CCRMA) at Stanford University
for several years, and other human–computer
interface researchers have experimented with them
(Rodriguez et al. 2007), even for audio applications
(Gabriel et al. 2008; Andersen et al. 2006; Verplank
and Georg 2011). The present work differs, sin embargo,
in its open-source perspective, focus on low cost,
specific design features motivated by the musical
application, and firmware design enabling feedback
control via floating point computations.
Concept
For musical applications, it could be interesting to
combine the FireFader with other more common
user interfaces that lack force feedback. Undoubt-
edly, building a keyboard with all force-feedback
keys is remarkable (Cadoz, Lisowski, and Florens
1990; Gillespie 1992; Oboe and De Poli 2002; Gille-
spie et al. 2011) and could be implemented by
connecting several FireFaders to a Universal Serial
Bus (USB) hub and orienting the faders vertically.
One can also obtain interesting interactions by com-
bining fewer force-feedback degrees of freedom with
standard musical controllers. Por ejemplo, Cifra 1
presents an example configuration combin-
ing a FireFader prototype with a small MIDI
keyboard.
Lights
Although force feedback can easily change the
nature of interaction from a performer’s perspective,
the audience may not perceive the presence of the
force feedback in many performance contexts. De este modo,
we believe the device benefits from a method for
communicating the force level to the audience, como
a first step toward coherent multimodal feedback
(Cadoz et al. 2003).
For each motor on the FireFader, there is a light
indicating the force level applied to that motor. Incluso
if audience members may not all realize that the
lights communicate the force level, the lights can at
least draw attention to the force itself. Además,
this feature may seem exciting and inspiring to
músicos, emphasizing the distinction between a
common digital musical interface and an interface
Berdahl and Kontogeorgakopoulos
25
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 2. Template for
making the top plate using
a laser cutter.
Cifra 3. Completed
FireFader.
Cifra 4. The Arduino
Nano 3.0 microcontroller
board plugs into the
2MOTOR H-bridge board,
which plugs into a
solderless breadboard.
with force feedback. En efecto, a bright light helps
suggest the metaphor of fire, suggesting power,
excitement, and extraordinariness. This metaphor
explains the device’s name, which helps to further
distinguish it from other user interfaces that lack
active haptic feedback.
Encima 50 different lights have been tested in order
to find a bright light that looks appealing over the
motor H-bridge driver’s dynamic range. A FireFader
with a 5-W, 12-V halogen lamp provides the bright
glow shown in Figure 1. Twelve-volt replacement
lamps constructed from LEDs are more efficient
but may have light-radiation patterns that are
more focused, or they may necessitate additional
components such as diode bridges.
Cifra 3
Cifra 4
Electronics
Enclosure
A laser cutter is used to cut the top plate of the
enclosure. This approach is appealing because of its
low cost, ease of reconfigurability, and the wide va-
riety of materials available for laser cutting. Cifra 2
shows the open-specification template for the top
plate. The laser cutter cuts all the way through the
top plate to make the screw holes and the slits for
the fader shafts, but the laser cutter only engraves
the text and icon. The top plate mates with the
commercially available Strapubox 2003 SW plastic
box with dimensions 160 mm × 83 mm × 52 mm.
The template could also easily be modified to fit
another mass-produced box. A completed FireFader
incorporating two motorized faders is shown in
Cifra 3.
The FireFader is based on the Arduino platform,
which is an open-source platform for prototyping
electronics. Arduino aims specifically to make it
as easy as possible for novices to make interactive
projects (Banzi 2009). The Arduino Nano 3.0 micro-
controller board runs the firmware for controlling
the faders, and the 2MOTOR H-bridge board from
Gravitech powers the motors. These two boards can
be conveniently plugged into one another and a sol-
derless breadboard (ver figura 4) for rapid hardware
prototyping. Alternativamente, in finalized projects,
the motor and other components can be hard-
wired directly to the 2MOTOR H-bridge board. El
schematics for both boards are publicly available
on the manufacturer’s Web site (www.gravitech.us),
and the interconnecting wires for interfacing the
26
Computer Music Journal
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 5. FireFader
schematic for controlling
two motorized faders and
two halogen lamps.
Cifra 6. Signal flow for
the control loop.
Cifra 5
Cifra 6
remaining FireFader components are given in the
schematic in Figure 5.
Software Design
The FireFader communicates with a general-purpose
computer as shown in Figure 6 in order to gain access
to fast real-time floating-point computation and
high-quality audio converters. The use of a general-
purpose computer eases programming the FireFader,
because code efficiency is less of a concern.
Arduino Firmware
A special firmware program written by the first
author must be installed on the Arduino Nano.
Although the firmware is generally designed to
work for many applications, some users might be
interested in customizing the firmware, por ejemplo
for compatibility with other sensors or motors.
The main loop in the firmware repeats the steps
detailed in Figure 7 in an infinite loop. Data are sent
over the serial interface using packets of 8-bit bytes.
Each packet begins with the sentinel byte 255. Cada
following byte in the packet, which represents a
single physical variable, may take on any byte value
other than 255. Because each physical variable is
encoded using only a single 8-bit byte, no bit shifting
is required in the source code, which makes it easier
to read. For further details, please see Figure 7.
Data obtained via the analog inputs are averaged in
order to reduce the effects of noise. The top of the
flowchart in Figure 7 indicates that the firmware
loop does not repeat until at least one new packet
arrives over the serial input from the computer.
This feature keeps the firmware in sync with the
driver running on the computer, thereby making
it easier to debug the driver and firmware than if
the firmware ran the entire loop or part of the loop
freely at its own speed.
The forceEnable variable helps in protecting
the device. If the user is not touching a given fader
knob, the force for the fader will be turned off (ver
Cifra 7). También, if the fader knob is at the end of its
travel, the force is turned off to prevent the device
from overheating, which could for example happen
Berdahl and Kontogeorgakopoulos
27
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 7. Flow chart
describing the operation of
the firmware. The doubled
arrows show the data
paths and the single
arrows show the
transitions through the
flowchart states.
Serial input
Read input packet from the serial interface.
s
d
norte
a
metro
metro
oh
C
mi
C
r
oh
F
Is another
input packet already
arriving?
Sí
No
D4,D5,D6,D7
Scale the last force commands by
forceEnable
and write the scaled forces to the H−bridge pins.
D2,D9
Perform capacitive sensing for each of the faders and
digitally lowpass−filter the results.
User chooses
analog inputs
For two auxiliary «user» pins to be used with generic
sensors, sample the corresponding analog input pins
multiple times and average the results.
A2,A3
For each fader, sample the corresponding analog input
pin multiple times and average the results.
Serial
producción
Send the sentinel character 255 over the output serial
interface to indicate that a new packet is being sent.
Send the position of each fader, clipping it to the range
0 a 254 to avoid confusion with the sentinel.
Send the values from the auxiliary analog inputs,
clipping each value again to the range 0 a 254 to avoid
confusion with the sentinel.
Send the capacitive sensing values for each of the
faders, clipping each value to the range 0 a 254 a
avoid confusion with the sentinel.
s
norte
oh
i
t
i
s
oh
PAG
s
r
mi
t
mi
metro
a
r
a
pag
y
r
a
i
yo
i
X
tu
A
s
yo
a
norte
gramo
i
s
mi
s
norte
mi
s
h
C
tu
oh
t
s
mi
yo
b
a
i
r
a
v
mi
yo
b
a
norte
mi
mi
C
r
oh
F
Calculate a first−order model of the temperature of
each motor using the history of the force commands.
Temp.
estimate
forceEnable
Calculate the variable for each fader so that the fader force can be
disabled if the user is not touching the fader knob, if the fader is pushed all the way
to either extreme, or if the temperature model indicates that the fader might
be getting too hot. Lowpass−filter the variable for smooth operation.
forceEnable
28
Computer Music Journal
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 8. Order of channels
for generating a
standalone application for
controlling the FireFader.
if the software got stuck trying to push the fader
knob beyond the end of travel. Finalmente, a simple
temperature model uses the history of the force
commands to produce an approximate temperature
estimate (ver figura 7), which is used to turn off
the force if the fader motor might be getting too hot
because of a prolonged period of large force exertion.
Position 1
Position 2
1
2
Faust
DSP
Code
Fuerza 1
Fuerza 2
Left−channel audio out
Right−channel audio out
1
2
3
4
Host Software on the Computer
The dynamics of the FireFader can be programmed
using a variety of host software programs.
Max/MSP
In Max/MSP, the FireFader appears as an external
object, with audio outlets transmitting the fader
positions to the patch and audio inlets receiving
force signals generated by the user patch. Desde
a haptic signal-processing library with extensions
for physical modeling has been developed by the
autores, it is straightforward for users to develop
physical models for programming the FireFader
(Berdahl, Niemeyer, and Smith 2009a; Berdahl,
Kontogeorgakopoulos, and Overholt 2010). Este
programming paradigm is particularly convenient
for rapid prototyping of new force-feedback control
algoritmos, because the patches can be edited in real
tiempo.
For lower-latency driver performance, el
Max/MSP scheduler interval should be set to
1 msec in the Preferences window under “Sched-
uler,” and the audio signal vector size must be set
a 1 so that the physical models are stable (Berdahl,
Kontogeorgakopoulos, and Overholt 2010). En el
case of the firmware and driver for only one fader,
we used an oscilloscope to determine that the la-
tency around the entire control loop ranges between
aproximadamente 2.5 msec and 6 msec for a MacBook
Pro running OS X 10.6. Different latencies will be
achieved using different firmware versions, drivers,
operating systems, and even operating system up-
fechas. The jitter in the latency is due to various
timing uncertainties, relating primarily to jitter in
the way that the host application services interrupts
at high frequencies.
Computationally Efficient Externals
for Physical Modeling
Computationally efficient physical models for
controlling the FireFader can be generated using
the Synth-A-Modeler software package (Berdahl and
Herrero 2012). Synth-A-Modeler outputs Faust DSP
code that can then be compiled into a variety of
objetivos (Orlarey, Fober, and Letz 2009). The physical
models can therefore be compiled into externals for
Max/MSP or many other audio host applications, o
they can be compiled into custom applications.
Standalone Application
We have written a Faust architecture file that allows
a Faust DSP file generated by Synth-A-Modeler
to be directly converted into a standalone JACK
Audio Connection Kit (JACK) audio application.
The application has a Nokia Qt-based graphical
user interface and communicates directly with
the FireFader, where the communication with the
FireFader happens in between audio vectors. Para esto
reason, choosing an audio vector size of 64 muestras
en 44.1 kHz or similar is appropriate, given the
specifications for the FireFader. The ordering for the
channels in the Faust DSP file is shown in Figure 8.
Unsupported Applications
Example drivers have also been written so that Pure
Datos (Pd) (Puckette 1997), Chai 3D (Conti et al.
2003), and custom C/C++ applications can commu-
nicate directly with the FireFader device; the future
development of these drivers to support upgrades
to Pd, Chai 3D, etc., will depend on the support re-
turned by the community. Users of other computer
music programming environments can either adapt
the existing drivers to their environment, or they
can use the JACK audio server to input audio from
Berdahl and Kontogeorgakopoulos
29
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 9. User interacting
with one FireFader
channel rendering a spring
with stiffness 0.5 N/mm
without the touch sense
enable feature. (Top:
Position of the FireFader
knob. Bottom: Capacitive
touch sensor signal is
elevated if the user is
touching the knob.)
a Faust-generated standalone application to their
ambiente.
Touch Sensing
Capacitive sensing can be used to determine if
a given fader knob is being touched by a user or
no. Por esta razón, the FireFader uses electrically
conductive fader knobs and motorized faders that
include a “touch sense” pin (ver figura 5). Cuando el
user touches a fader knob, the user capacitively loads
the touch sense pin, which in turn makes it take
significantly longer for the 1-M(cid:2) pull-up resistor to
pull the capactive input pin (D2 or D9 in Figure 5)
alto. Depending on how the user is connected to
electrical ground and other electronic devices, él
may take a particularly long time for the capacitive
input to go high if the user is touching the knob.
En este caso, the firmware will simply stop waiting,
reset the pin to 0 V, and continue to other parts of
the program to avoid introducing additional latency.
Como consecuencia, the fidelity of the capacitive sensing
signal from the firmware is limited. The firmware,
sin embargo, can reliably detect whether or not the user
is touching the knob by comparing the capacitive
sensing signal against a constant threshold. El
following section describes how the FireFader uses
the capacitive sensing signal to enable the force
feedback only when a user is touching a given fader.
Enabling the Touch Sensing
If haptic feedback control with a large gain is desired,
which can be the case when modeling a stiff spring
or strong damper, then the force feedback can cause
a haptic force-feedback device to become unstable,
especially if the latency around the control loop is
significantly long (Diolaiti et al. 2005). Por ejemplo,
consider using one channel of the FireFader to model
a relatively stiff spring with stiffness 0.5 N/mm.
Although the feedback system is stable for certain
kinds of manipulations, such as when the user
holds onto the fader knob and wiggles the device
back and forth slowly (see the thick lines for the
time period 0–4.5 seconds of Figure 9), the system
can become unstable and oscillate erratically if
the user briefly taps the knob to the side (see two
boxed portions of Figure 9)—in each case a brief
tap causes the feedback system to go unstable,
making the fader knob move rapidly back and forth
like a released spring, until the user touches the
knob again. Por eso, although the system is stable
when the user is continuously touching the device,
thereby increasing the mechanical load on the
device (Kuchenbecker, Parque, and Niemeyer 2003),
the feedback system can become unstable if the user
is not continuously touching the device.
The touch sense enable feature improves the
stability of the FireFader by disabling the feedback
control for a knob when the user is not touching
the knob. This enables the implementation of stable
models that have increased stiffnesses, damping
parámetros, etc.. Consider then the simple spring
model with the touch sense enable feature—in this
caso, the device behaves like a spring with stiffness
0.5 N/mm only while the user is touching the device,
otherwise the force is zero. The device performance
is then stable as demonstrated in Figure 10. As in the
prior example, the user initially wiggles the device
back and forth to feel the spring (see thick lines for
time period 0–4.2 seconds in Figure 10). Próximo, el
user taps the fader knob ten times. For each tap, el
user briefly touches the fader knob resulting in a
brief spike in the touch sensor signal, the fader knob
moves, and then the force feedback is disabled as the
user is no longer touching the knob, which causes
the fader position to stop at a new value, remaining
constant until the next tap (see time period 4.5–10.5
seconds in Figure 10).
30
Computer Music Journal
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 10. User interacting
with a FireFader rendering
a spring with stiffness
0.5 N/mm with touch
sense enable feature.
(Top: Position of the
FireFader knob. Bottom:
Capacitive touch sensor
signal.)
Cifra 11. Sound Flinger.
]
metro
C
[
norte
oh
i
t
i
s
oh
PAG
1
0
−1
h
C
tu
oh
t
r
oh
s
norte
mi
s
2
4
6
8
10
Time [segundo]
The touch sense enable feature is also useful for a
beginning user, particularly if he or she inadvertently
programs an unstable model. In our experience, a
beginning user’s first instinct may be to let go of the
knob if it becomes unstable. En este caso, the touch
sense enable feature will cause the force feedback
to turn off, which prevents the device from moving
erratically, becoming damaged, or possibly even
upsetting the user.
Testing
The FireFader has been tested via instructional
escenarios, augmentation with additional sensors,
and integration into completed projects.
Instructional Prototypes
En octubre 2011, the first author evaluated the
FireFader from an instructional perspective using
five FireFader prototypes. Students in Stanford Uni-
versity’s Music 250A course, Physical Interaction
Design for Music, used the prototypes to complete
a musical design-oriented exercise, which was
motivated by a previous exercise by Bill Verplank
(Verplank 2005). In the new exercise, students pro-
grammed by specifying physical models instead of
writing explicit program code (Berdahl, Florens, y
Cadoz 2011). The results indicated that specifying
physical models directly enabled students to focus
on calibrating the essential model parameters rather
than getting distracted by the implementation de-
tails of writing and debugging explicit program code.
Estimating Downward Pressure
Adding sensors to a project tends to be less expensive
than adding motors. Besides incorporating generic
analog sensors via user-chosen pins as supported by
the firmware (ver figura 7), users may be interested
in the application of sensing the downward pressure
that a user applies to the FireFader knob. Incluso
though the standard FireFader implementation
allows for actuation only along the axis of the fader,
certain haptic illusions can be created by modulating
the fader force as a function of the downward
pressure, providing the illusion of additional axes of
feedback control (Verplank, Gurevich, and Mathews
2002; Lederman and Jones 2011). A conference paper
(Berdahl and Kontogeorgakopoulos 2012b) explains
how to estimate the downward pressure for the
FireFader using relatively inexpensive sensors.
Example Projects
Example projects have provided more opportunity
to refine the FireFader’s design. Cifra 11 shows the
Sound Flinger by Chris Carlson, Eli Marschner, y
Hunter McCurry, which was designed for several
users standing at the center of a square of four
speakers (not shown). The Sound Flinger projected
an input sound signal at a time-varying angle out of
the loudspeakers toward the users in the horizontal
plane (Carlson, Marschner, and McCurry 2011). Este
angle corresponded to the position of a virtual mass
that slid along the edges of the square formed by
the four faders (ver figura 11). Using the faders, el
Berdahl and Kontogeorgakopoulos
31
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 12. String-U-Topia.
Cifra 13. Exoskeletal
haptic pincher from
PROJECT SQUEEZE.
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
users could add and remove inertia from the virtual
mass, which gave the sonic impression of flinging
the sound around the room.
The first author constructed the String-U-Topia
using an early FireFader prototype. The motorized
fader held in the user’s right hand enabled him or
her to pluck three virtual strings, while the user’s
left hand pressed buttons to adjust the pitches of
the three virtual strings (Berdahl, Niemeyer, y
Smith 2009b). The entire instrument sat neatly in
the user’s lap (ver figura 12).
For PROJECT SQUEEZE in the CS277 Exper-
imental Haptics class at Stanford University, joel
Sadler and Shruti Gupta constructed a single-degree-
of-freedom exoskeletal haptic pincher derived from
FireFader parts. The enclosure was made from
laser-cut acrylic, including a piece which links
the fader handle with the thumb (ver figura 13).
The goal was to investigate the utility of an
inexpensive prosthesis.
Conclusions
The design of the FireFader has been an iterative
process and was first announced in an abstract
published by the Acoustical Society of America
(Berdahl 2011). The present article provides the full
open-source disclosure of the hardware and software
and describes how musicians have been using the
FireFader.
The hardware component specifications, Arduino
firmware, pictures, and drivers for Max/MSP, Faust-
generated applications, Pd, Chai 3D, and sample
generic C/C++ applications can be found in an
archive linked from the project Web site (URL
provided subsequently). The device is simple, abierto-
source, and reconfigurable, so we hope that it will
appeal to musicians as well as the broader do-it-
yourself (DIY) comunidad. In order to help galvanize
the DIY community, we have started an online sup-
port discussion group, which is also accessible from
the project Web site: http://www.openhaptics.org.
We believe that the FireFader is the only com-
pletely open-source hardware and software system
for building haptic musical instruments with force
comentario. It is relatively inexpensive, with the total
32
Computer Music Journal
cost of the parts adding up to about US$ 150 for two
faders. The touch sense enable feature allows for the
force feedback to be enabled only when the user is
touching the knob, facilitating force-feedback con-
trol with larger control gains despite the relatively
long feedback control latency.
In future work, we plan to create a library of
physical models for controlling the FireFader to aid
in teaching users how to integrate force feedback
into their own musical instrument designs. Alguno
of these physical models will demonstrate a haptic
approach for controlling digital audio effects in
which the user can feel the sound as it is processed.
Por ejemplo, with the “tremolo” haptic audio effect,
the user can change the volume of an input audio
signal by adjusting the force that is applied. En el
mismo tiempo, the user feels vibrations of the audio
signal that are coherent with the output sound. en un
second example, known as the “switch” haptic audio
efecto, the interaction is similar but more strongly
nonlinear, resulting in significant distortion of the
input sound signal that the user can palpably feel
(Kontogeorgakopoulos and Kouroupetroglou 2012).
We hope that the computer music community
will make use of the FireFader resources by building
projects derived from them and by sharing their
own modifications with the community. Besides
maintaining the resources, we plan to continue
using them in future projects to come, incluido
building more elaborate physical interfaces based on
the FireFader.
Expresiones de gratitud
The authors would like to thank the Alexander
von Humboldt Foundation for supporting this work
and would like to acknowledge the researchers
who have influenced, motivated, and enabled this
trabajar, including Bill Verplank, Claude Cadoz, Stefan
Weinzierl, Jean-Loup Florens, Annie Luciani, y
Chris Chafe.
Referencias
andersen, t. h., et al. 2006. “Feel the Beat: Direct
Manipulation of Sound during Playback.” In First IEEE
International Workshop on Horizontal Interactive
Human-Computer Systems, páginas. 123–126.
Banzi, METRO. 2009. Getting Started with Arduino. Sebastopol,
California: Make Books, O’Reilly Media.
Berdahl, mi. 2011. “FireFader: A Single Degree-of-Freedom
Force-Feedback Device for Multimodal Interaction with
Physical Models.” In Proceedings of the 162nd Meeting
of the Acoustical Society of America, pag. 2508.
Berdahl, MI., J.-L. Florens, and C. Cadoz. 2011. “Using
Physical Models is Necessary to Guarantee Stabile
Analog Haptic Feedback for Any New User and Haptic
Device.” In Proceedings of the 8th Sound and Music
Computing Conference (pages unnumbered).
Berdahl, MI., y un. Kontogeorgakopoulos. 2012a.
“Engraving–Hammering–Casting: Exploring the Sonic-
Ergotic Medium for Live Musical Performance.” In
Proceedings of the International Computer Music
Conferencia, páginas. 387–390.
Berdahl, MI., y un. Kontogeorgakopoulos. 2012b. "El
FireFader Design: Simple, Open-Source, and Reconfig-
urable Haptics for Musicians.” In Proceedings of the 9th
Sound and Music Computing Conference, páginas. 90–98.
Berdahl, MI., A. Kontogeorgakopoulos, y D. Overholt.
2010. “HSP v2: Haptic Signal Processing with Ex-
tensions for Physical Modeling.” In Proceedings of
the Haptic Audio Interaction Design Conference, páginas.
61–62.
Berdahl, MI., GRAMO. Niemeyer, y j. oh. Herrero. 2009a. “HSP:
A Simple and Effective Open-Source Platform for
Implementing Haptic Musical Instruments.” In Pro-
ceedings of the Ninth International Conference on New
Interfaces for Musical Expression, páginas. 262–263.
Berdahl, MI., GRAMO. Niemeyer, y j. oh. Herrero. 2009b.
“Using Haptic Devices to Interface Directly with
Digital Waveguide-Based Musical Instruments.” In
Proceedings of the Ninth International Conference on
New Interfaces for Musical Expression, páginas. 183–186.
Berdahl, MI., y j. oh. Herrero. 2012. “An Introduction to
the Synth-A-Modeler Compiler: Modular and Open-
Source Sound Synthesis using Physical Models.” In
Proceedings of the Linux Audio Conference (paginas
unnumbered).
Birnbaum, D. 2007. “Musical Vibrotactile Feedback.”
Tesis de maestría, Universidad McGill, Montréal, Canada.
Cadoz, C., l. Lisowski, and J.-L. Florens. 1990. “A Modular
Feedback Keyboard Design.” Computer Music Journal
14(2):47–51.
Cadoz, C., A. Luciani, and J.-L. Florens. 1984. “Responsive
Input Devices and Sound Synthesis by Simulation
of Instrumental Mechanisms: The Cordis System.”
Computer Music Journal 8(3):60–73.
Berdahl and Kontogeorgakopoulos
33
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cadoz, C., A. Luciani, and J.-L. Florens. 1993. “CORDIS-
ANIMA: A Modeling and Simulation System for
Sound and Image Synthesis—The General Formalism.”
Computer Music Journal 17(1):19–29.
Cadoz, C., et al. 2003. “ACROE – ICA: Artistic Creation
and Computer Interactive Multisensory Simulation
Force Feedback gesture transducers.” In Proceedings
of the International Conference on New Interfaces for
Musical Expression, páginas. 235–246.
Carlson, C., mi. Marschner, and H. McCurry. 2011. "El
Sound Flinger: A Haptic Spatializer.” In Proceedings
of International Conference on New Interfaces for
Musical Expression, páginas. 138–139.
Castet, J., D. Courouss ´e, and J.-L. Florens. 2007. “A
Real-Time Simulator for Virtual Reality Conceived
around Haptic Hard Constrains.” In Proceedings of the
International Conference on Enactive Interfaces, páginas.
49–52.
International Computer Music Conference, páginas. 400–
403.
Kontogeorgakopoulos, A., y G. Kouroupetroglou.
2012. “Low Cost Force-Feedback Haptic Interaction
with Haptic Digital Audio Effects.” In E. Efthimiou,
GRAMO. Kouroupetroglou, and S.-E. Fotinea, eds. Lecture
Apuntes sobre Inteligencia Artificial: Gesture and Sign Lan-
guage in Human-Computer Interaction and Embodied
Comunicación, volumen. 7206. Berlina: Springer Verlag,
páginas. 48–57.
Kuchenbecker, K., j. Parque, y G. Niemeyer. 2003.
“Characterizing the Human Wrist for Improved Haptic
Interaction.” In Proceedings of the International
Mechanical Engineering Congress and Exposition,
Symposium on Advances in Robot Dynamics and
Control, páginas. 591–598.
Lederman, S., y yo. jones. 2011. “Tactile and Haptic
Illusions.” IEEE Transactions on Haptics 4(4):273–294.
Conti, F., et al. 2003. “The CHAI Libraries.” In Proceedings
Oboe, r., y G. De Poli. 2002. “Multi-Instrument
of EuroHaptics, páginas. 496–500.
Diolaiti, NORTE., et al. 2005. “The Effect of Quantization and
Coulomb Friction on the Stability of Haptic Rendering.”
In Proceedings of the First Joint Eurohaptics Conference
and Symposium on Haptic Interfaces for Virtual
Environment and Teleoperator Systems, páginas. 237–246.
Florens, J.-L., C. Cadoz, y un. Luciani. 1998. “A Real-
Time Workstation for Physical Model of Multi-Sensorial
and Gesturally Controlled Instrument.” In Proceedings
of the International Computer Music Conference, páginas.
518–526.
Florens, J.-L., et al. 2004. “ERGOS: Multi-Degrees of
Freedom and Versatile Force-Feedback Panoply.” In
Proceedings of EuroHaptics, páginas. 356–360.
Gabriel, r., et al. 2008. “BounceSlider: Actuated Sliders for
Music Performance and Composition.” In Proceedings
of the Second International Conference on Tangible
and Embedded Interaction, páginas. 127–130.
Gillespie, B. 1992. “The Touchback Keyboard.” In
Proceedings of the International Computer Music
Conferencia, páginas. 77–80.
Gillespie, B. 2003. “Haptic Interface for Hands-On In-
struction in System Dynamics and Embedded Control.”
In 11th Symposium on Haptic Interfaces for Virtual
Environment and Teleoperator Systems, páginas. 410–
415.
Virtual Keyboard: The MIKEY Project.” In Proceedings
of the Conference on New Instruments for Musical
Expression, páginas. 137–142.
Okamura, A., C. Ricardo, y M. Cutkosky. 2002. “Feeling
is Believing: Using a Force-Feedback Joystick to Teach
Dynamic Systems.” ASEE Journal of Engineering
Educación 92(3):345–349.
Orlarey, y., D. Fober, and S. Letz. 2009. “FAUST: Un
Efficient Functional Approach to DSP Programming.”
In Assayag, GRAMO. y un. Gerzso, eds. New Computational
Paradigms for Computer Music. Sampzon, Francia:
Edition Delatour, páginas. 65–96.
Puckette, METRO. 1997. “Pure Data.” In Proceedings of
the International Computer Music Conference, páginas.
224–227.
Rodriguez, J., et al. 2007. “One-Dimensional Force Feed-
back Slider: Digital Platform.” In IEEE VR 2007 Work-
shop on Mixed Reality User Interfaces: Especificación,
Authoring, Adaptation, páginas. 47–51.
Verplank, B. 2005. “Haptic Music Exercises.” In Proceed-
ings of the International Conference on New Interfaces
for Musical Expression, páginas. 256–257.
Verplank, B., and F. Georg. 2011. “Can Haptics Make New
Music?—Fader and Plank Demos.” In Proceedings of
the International Conference on New Interfaces for
Musical Expression, páginas. 539–540.
Gillespie, B., et al. 2011. “Characterizing the Feel of the
Piano Action.” Computer Music Journal 35(1):43–57.
Hayes, l. 2012. “Performing Articulation and Expression
through a Haptic Interface.” In Proceedings of the
Verplank, B., METRO. Gurevich, y M. Mathews. 2002. "El
Plank: Designing a Simple Haptic Controller.” In
Proceedings of the International Conference on New
Interfaces for Musical Expression, páginas. 1–4.
34
Computer Music Journal
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
mi
d
tu
/
C
oh
metro
j
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
/
/
/
3
7
1
2
3
1
8
5
5
8
0
9
/
C
oh
metro
_
a
_
0
0
1
6
6
pag
d
.
j
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3