IDMIL - Digital Orchestra Toolbox Max Object

http://www.idmil.org/
" This “toolbox” is a collection of Max/MSP objects that we have found useful in creating gesture processing patches for digital musical instruments. Each patch is accompanied by a help patch to demonstrate its use. "

Could be useful

A few links

To some researcher's pages:

Jorg Piringer - http://joerg.piringer.net/index.php?href=home.xml&mtitle=home
Michel Waisvisz - http://www.crackle.org/

Real time Arduino

UPDATE:

Very good description of how clockrate and prescalers work as well as code on giving 77KHz sampling rate!

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1208715493/11

The forum post in full for the tutorial given at the bottom

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1224777046/all

Its possible!!

http://www.youtube.com/watch?v=Y241UMFNxDY

http://www.youtube.com/watch?v=MqJAWH8IP6I&feature=related

Some forum entries:

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1246918085/5

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1288880698/6

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1237148726

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1263493844/5

Which all pretty much lead back to this document:

http://interface.khm.de/index.php/lab/experiments/arduino-realtime-audio-processing/

Curve Fitting

A bunch of random resources from the web found by Tom Mitchell regarding finding how close curves are together and stuff... TO BE UPDATED

http://en.wikipedia.org/wiki/Taylor_series

http://mathworld.wolfram.com/Polynomial.html

http://en.wikipedia.org/wiki/Newton's_method

http://www.google.co.uk/search?hl=en&client=safari&rls=en&q=bezier+curve+fitting+applet&aq=f&aqi=&aql=&oq=&gs_rfai=

http://www.serc.iisc.ernet.in/~amohanty/SE288/lagrange/curves.html

http://www.cs.princeton.edu/~min/cs426/jar/bezier.html

http://www.theparticle.com/applets/nyu/BezierApplet/

http://www.cse.unsw.edu.au/~lambert/splines/Bspline.html

http://www.cs.uwaterloo.ca/~r3fraser/splines/bspline.html

http://saltire.com/applets/advanced_geometry/spline/spline.htm

SOS Korg Wavedrum Entry

http://www.soundonsound.com/sos/apr10/articles/korgwavedrum.htm

FYP Order Form

1 - Arduino Uno - RS - 715-4081

1 - Texas Instruments OPA350UA op amp - Farnell - 1106197

1 - 1m USB 2.0 A to B Lead - Maplin - NO8JA

4 - piezo xducer 27/1.8 - Maplin - YU87U

Arduino And Max

http://www.arduino.cc/playground/Interfacing/MaxMSP

See Arduino2Max

http://cycling74.com/docs/max5/tutorials/max-tut/communicationschapter02.html

Serial communication tutorial from Arduino to Max

MIDI Equipment

Here are some sources for MIDI equipment in case I decide to stick with the Arduino and communicate via MIDI:

Female MIDI socket component with breadboard compatible pins - SparkFun Electronics

Various breadboard components from CuriousInventor.com. MIDI IN/OUT modules towards the bottom. Also available is a breadboard minijack socket which might be useful. *Both currently out of stock*

MIDI IN/OUT to USB converter from Amazon. Also see Maplins

Arduino and Audio

This post is dedicated to research into how best to set up an Arduino environment for audio.

http://interface.khm.de/index.php/lab/experiments/arduino-realtime-audio-processing/

http://www.ladyada.net/make/waveshield/libraryhc.html

As far as I understand it audio can be processed using the Arduino. When I say processed here I just mean ADC and pass to the computer and DAC from the computer and pass to a power amp. The first link above seems to be the way to do this but it seems rather problematic and tedious and the results have a very limited sample rate and depth (for audio) and so are unlikely to provide adequate quality in the long term.

I'm not looking into the possibility of having a dedicated device to perform the conversions and simply using the Arduino as it is designed to be used, to control data. Thinking about it now however even using a dedicated ADCDAC device the Arduino would still need to sample the output from the device and so would impose its crappy sampling rate and depth. Or would it if the audio input to it was in digital form?? The only analogue input would then be the FSR and the arduino is fine for that...questions!!!

Hardware Choices

For the Arduino I think I good place to go is Arduino starter packs. They contain the board itself as well as loads of other stuff that I'd probably end up having to buy or root around for at uni. These include basic things like wires, USBA/B cable, bread boards etc to more specific bits of kit which nevertheless would be useful for experiments such as transistors, resistors, LEDS, motors, photodiodes, etc. One UK based company that provides a good choice is Oomlout, offering a range of Arduino products here. They offer an all inclusive kit here and a slightly toned down version here. I'd probably opt for the latter since its cheaper but includes all the main components.

UPDATE

After borrowing an Arduino Diecimila I think the Uno would be better. Its very similar but one new feature is that it automatically selects where to draw power from (USB or Jack) based on what its plugged into. This would therefore be one less thing a user would need to know about when playing the instrument.

This piezo seems to be a good choice, though it is more expensive than others. This may be indicative of a better quality and since I hope to use the raw audio signal to not only trigger sounds but to influence them based on its characteristics, quality seems important. The piezo presents a reasonable frequency response (as reasonable as a piezo mic could be I guess) in the required range for knocks (100Hz to 5kHz).

This is a nicely sized square FSR from UK based Cool Components. I think a bigger size for the FSR, rather than the much smaller circular models, is a good idea since i think it would be easier to use ergonomically if the player didn't have to aim too much to use it could just lunge in its general direction without too many problems. It is important to note however that the FSR will need to form one resistor in a voltage divider in order to convert the resistance change of the FSR to a voltage change for input the arduino analogues.

I keep coming back to this loudspeaker driver so I suppose its appropriate. It certainly seems that way.

Arduino

I have pretty much made up my mind that I will use Arduino to input sensor data to the computer. There are several reasons. Its reasonably priced. It has a good reputation. There seems to be a good community and lots of tutorials. It comes with its own coding language and environment with plenty of help and examples. People have created similar projects. It uses a simple USB interface which also powers the board.

The Arduino Uno seems like a good choice. It is their newest product and is being billed as setting the standard for Arduino. Here is the Arduino Uno available from RS. The Duemilanove is also a popular modal so may be worth comparing. Here it is from RS.

Here is an application of a (frustratingly) basic drum machine using two piezo elements a speaker and an arduino. Not a lot of tech details but just for interest.

Here is an arduino forum entry by a guy building some thing that looks horribly similar to some of my very first ideas (very similar to the nukulele). They have a conversation regarding preamps for his piezo pickups which may prove useful later.

Preamp

Here I will consider the preamp that I (think) I will need to increase the output of the piezos. I reckon I should only need one preamp for two (for example) piezos since its IC has multiple input and outputs.

Preamps seem relatively inexpensive. Here are some available from Farnell. Here is one from Digikey.

Power Amplifiers

Here I will consider my choice of power amplifiers for the loudspeaker drive unit.

This seems to be a more problematic decision to make since a lot of power amplifiers are big bulky and heavy. I could of course use an off the shelf model external to my instrument, but this would mean running extra cables from the instrument and would also go against the sense of "self containedness" that I would like to aim for. Marshall created custom amplifier circuits for his viblotar so perhaps this is some thing I could look into. It may transpire though that this could be a project in itself and that I will be forced to choose a commercial model. In this case I could cite the development of an internal custom circuit as potential for future work.

For the time being, here are some power amplifiers available on RS. Obviously, power amplifiers are readily available from many sources (though not necessarily for cheap!). Number of speakers is also a consideration here since I reckon if I only have one speaker I will only need a mono amp!

Piezo Electronic Transducers

Here I will consider my choice of piezos for the detection of strikes on the Cajon.

A range of Piezos are available from RS here. I'm not sure what are the most important features of a good piezo but according to the literature* the full bandwidth for various knocks amd taps os 100Hz-5kHz so a piezo with a good response in this range would be a good start. This piezo may be an option. Here is its spec. A piezo element is available from Farnell but it does not seem to have a spec. Piezos seem to be dirt cheap!

Here is an entry in sensorwiki for piezos with some technical and practical details. Here is a youtube video of an engineer attaching piezo pickups to his acoustic guitar which provides a few practical hints, for example the use of a hot glue gun to attach the piezos.

Loudspeaker

In this post I will consider potential choices for the loudspeaker to be included in the design of my DMI. It is important to note that technical it is the loudspeaker drive unit that I will be including, this is useful to know when searching google etc.

I figure I need a loudspeaker which focuses on frequency response, but which also has a reasonable amount of power ie for playing in the studio. External speakers would naturally be used for filling live performance spaces. Size is also an issue since recent designs of my Cajon have been small enough to be played between the knees or on the lap. The speaker therefore needs to be of a reasonable weight and size to allow this.

The main contender so far is this speaker from Visatron, available through RS. The technical spec. is available here. Its 6.4cm/2.5inches and is described as a full range speaker driver. It has a typical impedance of 8ohms (the literature suggests this is the norm) and a power of 8W. Is this a good power level?

Sources for Electronic Components

The following are some resources for electronic components suggested by David Creasey since the university has accounts with the companies and is able to purchase modestly priced items on behalf of students:

Farnell (http://uk.farnell.com/)
RS (http://www.rswww.com/)

Farnell in particular has a range of components from power amplifiers and loudspeakers to sensors and transducers.

Also of interest is Studio Spares (http://www.studiospares.com/page/home/)

Updated Controller Design Spider Diagram

Above is an updated version of the previously posted spider diagram to show key aspects, issues, ideas and potential solutions regarding the design of the controller. The key area that has been changed is the "physical design" branch which has been updated following the completion of the cajon comparsa prototype. It therefore discussions some potential solutions and issues presented by the new design and compares some of its aspects with those of the original cajon prototype.



Key:
Orange - potential solutions
red - potential issues

Cajon Comparsa Prototype

This post introduces the second prototype drum to be created for this project. This model is based on the Cajon Comparsa. The dimensions for the drum were taken from this website.

Although some lessons were learned from building the Cajon (such as not using a nail gun) and the new comparsa model is undoubtedly improved from the cajon, there were still some mistakes and plenty of tricky areas. Again my father helped me in the construction of this drum and he has years of experience with woodwork and a collection of the necessary tools and plenty that makes things easier (ie belt and orbital sanders, routers, jigsaws). This is making me think that perhaps if possible I would like to use one of the prototypes that we create to at least begin testing out the electronic and synthesis systems and possibly for inclusion in a final deliverable artifact. This is because if I were to attempt to build a drum on my own in order to gain credit for the design it would likely be time consuming, difficult and potentially expensive and it seems like perhaps, although clearly the ergonomics of the controller are important, for this project I need to focus more on other potentially more difficult areas. This is especially taking into account the legendary difficulty of the mapping stage. Mark Marshall for example said in his e-mail that in his DMI projects he has spent by far the most time fine tuning the mapping.

The comparsa is perhaps a little too tall to play on one's lap since for me, my arms are forced to bend upwards at the elbow from my waist to reach the tapa which is not particularly comfortable.

These issues suggest that perhaps a design closer in dimension to the Cajonito or even the Palito would be more suitable. I have some reservations about this though since I would like as much of the electronics to be housed internally as possible and obviously the smaller the controller is the less this is possible. Naturally however more internal electronics is going to mean more weight which could be a problem if the player is trying to hold the instrument with their legs.

Here are some pictures of the Comparsa:

As can be seen in the first few pictures the comparsa (according the demonstration videos presented on the Schlagwerk website) is typically held with opposing corners gripped between the knees and the tapa angled slightly downwards, with the sound hole facing away from the player. Also visible in the last few pictures are the marks from the wood filler used to fill up holes and chips created in the wood, evidence that putting the thing together is a little tricky.

Design details:

• 9mm thick MDF for sides and bottom
• Joined with wood glue and screws, screws have been counter-sunk
• 3.6mm plywood for tapa
• Joined with screws, counter-sunk
• 5mm sound hole
• All sides curved using router
• Coated with wood stain for an attractive finish

It was decided to use MDF for the body of the instrument since its cheaper than plywood. Also important to note that the quality of the wood is unimportant from a sound point of view since the sound will be synthesized, therefore cheap but sturdy wood can be used. Having said this the MDF presented as many problems with splitting as did the plywood. Also the MDF was thinner than the plywood at 9mm instead of 12mm. This size was chosen to make the instrument lighter, considering it was to be held between the legs.

Commercial Electronic Hand Percussion

In an attempt to find a little inspiration I spent a little time searching for commercial products that were of a similar concept as my curent Cajon-inspired DMI idea. I was mostly interested in aspects of ergonomics, design and control as well as sensing technologies. I first found the Roland HPD-10:

Roland HPD-10 product demonstration
Official product page

The HPD-10 is based around a set of 10 drum pads. Each pad is pressure sensitive and are able to detect the position of the strike on their surface.

Of more interest to me is the Korg Wavedrum:

Korg Wavedrum product promotion
Korg Wavedrum official product page

The Wavedrum is especially interesting for two main reasons. Firstly the instrument does not use a typical drum pag style interface. The instrument actually uses a real drum skin as the main portion of its interface. The skin is surrounded by a rim with a control interface at the top. The instrument consists of four discrete triggers; two are part of the rim, one is just below the control interface and the fourth is the drum skin itself. The use of the skin should instantly give the instrument great tactile feedback when playing with hands or sticks. The second reason is the sensing method. The skin and rim are equiped with pickup sensors allowing the vibrations of the instrument in response to strikes to be sensed. These signals are then used as the input to the synthesis algorithms. I can imagine this allowing an amazing sense of connection and intimacy with the instrument since any conceivable interaction with its playing surface will have an audible effect which is unique each time for each player as well as a rich amount of nuance (given an appropriate and effective software strategy which I'm positive is used here). As well as this, the skin is equiped with a pressure sensor to pick up extra detail from the player's attacks. This is similar to what I had in mind for my own DMI with my idea to use piezo microphones to monitor the vibrations of the instrument body (inspired by a conversation with Hugues Genevois). The addition of pressure sensing is also an appealing idea.

There are two things however that both these controllers seem to lack and they are vibrotactile feedback and sound localization. Neither controller (as far as I can tell) has any kind of built in sound production and would rely on external speakers, removing the immediate sense of presence of an acoustic instrument. Also, neither instrument provides any kind of vibrotactile feedback system even if only as a result of active internal sound feedback. Although I can imagine this situation being better in the Wavedrum since it incorporates a much more freeling vibrating membrane (as apposed to pads) I still can't imagine it compares with being in bodily contact with say a cajon and feeling it vibrate in response to your gestures.

The Wavedrum especially then provides some interesting sensor applications as well as some evident areas that I can attempt to work on to make my DMI a little more individual

The Vodhran DMI by Mark Marshall et al

Three articles describing The Vodhran are available towards the bottom of Mark Marshall's publications page:

Controlling the Virtual Bodhran - the Vodran - R. Bresin, S. Dahl, M. Marshall, M. Rath and B. Moynihan

The Virtual Bodhran - Design and Development of a Virtual Musical Instrument - M. Marshall, B. Moynihan and M. Rath

The Virtual Bodhran - The Vodhran - M. Marshall, B. Moynihan and M. Rath

I include these articles here because I thought the Vodhran may be conceptually similar to my Cajon DMI but on reading the articles its possible that they're not so similar, however they're still examples of an interesting DMI. The most usual article seems to be the first one, which gives the most concrete description of the various applications of the Vodhran from the software to the controller.

One application of the Vodhran uses Max Mathew's radio baton to measure the distance and position of the each end of the beater from the antenna. Another uses a gesture tracking system to extract details about the player's gestures such as velocity and acceleration. The application conceptually closest to the ideas I have at the moment seems to be the one involving the ddrum (a MIDI drum-pad with velocity sensitivity and aftertouch parameters):

" In addition to the control devices described above, the ddrum [6]
was used to play the sound model of the Bodhran (see Figure 2 A).
The ddrum is a commercial available electronic drumpad and the
MIDI velocity out from the control unit was used to excite the
sound model. For the pad used there was also MIDI poly-aftertouch,
which was used to control the dampening of the model. The ddrum
is a nice interface to play the model because of it’s tactile feedback
to the player and the lack of cables to the sticks used for playing. "

Note the advantages stated regarding the ddrum's passive tactile feedback.

These articles seem to demonstrate three main ways that I could approach handelling gestural interfacing for this project:

1. Monitor the position of the beater/s (hands, sticks) with respect to the controller itself
2. Use gesture tracking to extract indirect features of player's gestures for control (ie velocity)
3. Monitor the state of the controller itself for strikes

Project Deliverables

Key:
red = essential
green = desirable
blue = optional

Above is a diagram showing the deliverable items for this project. This has been created in order to break the project into its constituent parts and, more importantly, to show which areas should be concentrated on most heavily and which areas can be left in case of difficult deadlines. Not only this but it also helps to identify areas of the project which could be developed if time allows and also suggests some ideas for future work/projects, which will be useful when writing the report.

As it stands this diagram is only a first draft and I will likely extend and add to it as I begin to work on each of the individual parts it shows.

Control Parameters for Musical Instruments - Daniel Levitan

Controll Parameters for Musical Instruments: a foundation for new mappinds from gesture to sound - Daniel Levitan et al (third result)

This is a brilliant paper. It offers up a new and detailed way of disecting a particular musical tone into its constituant parts, defined as selective cognitive preconditions, beginning, middle, end, terminus. It presents in details the choices we can make of how to execute each portion and what sonic results they will have and then goes onto combine these principals into the manners in which we can execute a tone as a whole, showing a decision path. According to the authors:


" The essential points are to
employ controls that are both consistent with how the
sound might have been created in a traditional musical
instrument (so as to exploit principles of cognition and
the mind), and to inspire designers to think about the
control of sound in ways that go beyond those employed
in traditional musical instruments (which are constrained
by physical, but not always cognitive factors). "

The concept of designing DMIs that operate in such away that their sonic responses to our gestures correspond to our pre-existing cognitive expectations is an interesting one. Another interesting thing about this article is one of the motivations of the authors:

" Our approach is the
consequence of one bias that we should reveal at the
outset: we believe that electronically controlled (and this
includes computer-controlled) musical instruments need
to be emancipated from the keyboard metaphor;
although piano-like keyboards are convenient and familiar,
they limit the musician’s expressiveness (Mathews
1991, Vertegaal and Eaglestone 1996, Paradiso 1997,
Levitin and Adams 1998). This is especially true in the
domain of computer music, in which timbres can be created
that go far beyond the physical constraints of traditional
acoustic instruments. "

Physical Interfaces in the Electronic Arts - Bert Bongers

Physical Interfaces in the Electronic Arts - Bert Bongers (first result)

A fairly informative and interesting paper split into three main parts. The paper starts with an introductory history of the ergonomical aspects of the design of instruments. Several models are then presented of interaction in performance arts along with some practical examples. The final part gives a fairly detailed account of various kinds of sensors, including some technical aspects to their design and operation, and lists examples of DMIs and installations which have made use of them.

Tactile Audio Feedback - Chris Chafe

Tactile Audio Feedback - Chris Chafe

Article arguing that vibrotactile feedback can help to improve performer control in musical instrument playing situations

Principles for Designing Computer Music Controllers - Perry Cook

Principles for Designing Computer Music Controllers - Perry Cook (third result)

This paper presents a list of guidelines/advice for designers of computer music controllers. They are demonstrated through descriptions of several controllers of various categories.

Remutualizing the Musical Instrument - Perry Cook

Remutualizing the Musical Instrument: Co-design of Synthesis Algorithms and Controllers - Perry Cook (third result)

Particularly interesting in this article is a presentation of a brief history of the concept of separating control interface from sound production mechanism (going back to the first organs) along with a proposition of reasons why this has lead to lack of intimacy:


" The major flaw in the controller/synthesizer paradigm is
the loss of intimacy between human player and instrument.
I pose three primary reasons (lacks) for this intimacy loss:
• Lack of haptic feedback from the controller/instrument to
the player. Haptic (combined senses of touch, including
skin vibration and pressure, and the muscle senses of
motion, position, and force) feedback has been increasingly
addressed in musical interface research projects.
Commercially, the most successful haptic systems are electronic
keyboards that copy (passively, through weights and
levers) the feel of piano keys.
• Lack of fidelity in the connections from the controller/
sensor to the generator, primarily delays and distortions
in response to gestures. “Distortion” here refers to any
response that does not meet some usual, learnable, or
repeatable expectation.
• Lack of any sense that sound comes from the instrument
(the controller) itself. More generally, this is a subset of a
larger feeling that no meaningful physics goes on in the
controller. Trends toward larger concert venues, greater
amplification, and larger loudspeakers have consistently
worked to diminish the importance of the actual acoustical
instrument sound. The aesthetic influence of this has
been great, most importantly for the player and composer,
and shows profoundly in the musical results (Trueman,
1999). "

Finally a presentation of some DMIs which have been designed in an attempt to remedy these issues. Notably these instruments have followed a process where the synthesis and controller designs have been created in parallel with each other and have informed each other throughout. Included is the previously seen BoSSA system, some hand percussion physical model controllers and the SqueezeVoxen, a series of accordian-inspired instruments (Bart, Lisa, Maggie and Santa's Little Helper) which seem to have designed as a family, like the T-Sticks.

Of particular interest is the Nukulele and the Nukulele'elua which bear striking resembelence to some of the very first design ideas I had for my DMI.

Controller Design Spider Diagram

On the advice of David Creasey I've created a spider diagram which attempts to break up the task of creating the actual physical controller into constituant issues and problems. Above is a first draft of this diagram. Hopefully through revising the diagram I can begin to address a few of the problems it identifies. I may have some more answers to the questions once I've tried the new controller designs mentioned in the previous post. I will be able to address many questions once I begin playing with electronics at university.

Alternative Cajon design ideas for DMI

David Creasey has pointed out that perhaps the Cajon in its well known "box that you sit on" form is slightly awkward to play. I think that this is perhaps true especially considering that players will typically lean back and balance on the instrument in order to better reach the playing surface, which still requires bending down some what. This also certainly confirms some worries I had regarding the placement of any transducers I should choose to include, which I would likely place on the front playing surface (the really any of the surfaces would be about as awkward). The problem is I think that reaching down in order to strike an area of the tapa is OK but reaching down to perform a gesture on say, a linear position sensor, would be much more awkward especially if one was trying to achieve really fine control. Of course an important advantage often cited regarding DMIs is that we are not necessarily obligated to certain designs when constructing the instrument. Specifically, the seperation of control interface and sound production mechanism inherent in DMIs means that the sound production mechanism does not dictate the physical properties of the controller. For example an acoustic stringed instrument needs to be of a specific length in order to hold strings of the correct length to achieve their desired pitch (more correctly it is a balance between string length, thickness and tension). With DMIs there are no such obligations, we can produce any sound from any thing. (although actually I'm beginning to think we are not necessarily completely free of obligation since for example inclusion of one or several loudspeakers immediately requires an object of a certain size, the same for housing internal electronics and for providing transducers of a certain shape, length, size. It seems to me that obligations are perhaps reintroduced when thinking about designing GOOD or EFFECTIVE or EXPRESSIVE DMIs given the results of research. That's another argument for another post though.)

Keeping with the original inspiration for a Cajon-inspired controller (easy construction, loudspeaker placement etc) David has suggested I might try looking into other types of drum with the same family. Specifically the Cajinto. From the same website however I have found a section called "Cajon Compact" under which are the Cajon Comparsa and the Cajonito. Both of these drums closely follow the design of the Cajon in that they are both boxes with a hole in them but they are much smaller and are shown held between the legs like a bongo. The Cajonito is the smallest of the two. I think I would opt for one of these two over the Cajinto since I like the way they encourage a lot of bodily contact with the instrument, which I think would really allow the vibrotactile feedback element to do its job well. I'm not sure which of the two I would choose however. The Comparsa is nice because there's plenty of room for all the electronics on the inside and plenty of room to play on the top but the Cajonito, with its smaller size, is perhaps comfier to hold and easier to play.

All of these instruments are listed on the percussion website shlagwerk.com. This is a brilliant website for my needs since it has an extensive selection of instruments from the Cajon family, it has video demonstrations of all of them and it gives their dimensions which is one less concern when it comes it building!

There are two other options which come from the percussion company Meinl. The first is the Bongo Cajon. This is a smaller rectangular Cajon designed to be played either between the legs or on the lap. The bottom side is left open for sound output and interestingly the inside of the drum is split into two sections with a piece of wood, one being bigger than the other. This allows the player to get a low and a high sound out of the tapa. I think this could be quite a nice playing style to emulate in my DMI by some one measuring playing position from left to right. This could not only affect pitch but also some aspects of timbre which would give a nice one-to-many mapping associated with the best DMIs. Would also be interesting to think about the y-axis position.

The second option from Meinl is the Palito. This instrument seems similar to the Schlagwerk Cajonito, though seems to give a different sound. Also the Cajonito seems to be larger. I think I would probably opt for one of the above Schlagwerk designs above the Palito since I need to think about practical space considerations regarding the internal electronics, particularly the loudspeaker (not to mention amplifier and power supply which are another worry all together).

Listed below are some basic resources i've found so far:

This is a project page made by a guy who built his own approximation of a Bongo Cajon. There's some basic instructions along with construction pictures and a demonstration video. I don't really like the finished article. It seems rather big and awkward to play. Also the method of producing two seperate tones from the drum does not seem as effective as the more usual technique mentioned above.
Demonstration video of the Meinl Bongo Cajon
A second very similar demonstration of the same instrument by the same guy
Demonstration video of the Meinl Palito

Cajon Drum Prototype

Over the last few days my father and I have been working on a prototype cajon drum using designs and tutorial videos found on the internet (the sources for which are listed in a previous post), as well as some improvisation and customization on our part based on the materials available to us as well as some mistakes made.

The drum we've made is very much a practise run and there are certain aspects of its design that did not go so well at the time of building. But, importantly, there have been lessons learned from the mistakes and we plan to improve on the design and build a superior version of the drum.

In that respect building this initial version has been extremely beneficial. Although the drum is designed to be purely acoustic according to the traditional design, and so other issues such as electronic have not yet been addressed, its building has highlighted that creating a box of this kind is not as straight forward as first thought and there's plenty of room for mistakes.

Despite all that, the box functions perfectly well as a drum. It is possible to find different tones across the playing surface (tapa) and to combine these into interesting rhythms. Most importantly, its a lot of fun to play.

Here are some pictures of the design and basic playing position:

In this last one you can make out that the box is slightly tipped backwards during playing to allow more comfortable access to the tapa.

Design details:

• 12mm thick plywood for sides, back, top and bottom
• Joined with wood glue and nails (using nail gun)
• 3.6mm thick plywood for tapa
• Joined with screws

I'll write up some detailed design details for the next version since its likely to be the finally and should have no mistakes.

I now need to decide on some design details relating to the electronic version including:

• Loudspeaker size, since this will inform the required diameter of the "sound hole" in the back where the loudspeaker will be mounted.
• I know loudspeaker size will also affect the required size of the box (in terms of volume) but I'm not sure how. I may have to ask David Creasey about this...
• I need to make sure that I have easy access to the inside of the box to place the electronics, this should be possible through unscrewing the tapa.
• I need to think about issues around transducers and their placement for my DMI since I haven't actually got a concrete plan yet.

The Importance of Parameter Mapping in Electronic Instrument Design - Andy Hunt; Marcelo M. Wanderley; Matthew Paradis

The Importance of Parameter Mapping in Electronic Instrument Design - Andy Hunt, Marcelo Wanderley, Matthew Paradis

A rather brilliant article for several reasons. To start this article serves as some thing as a reference point for many major research projects into mapping in DMIs and as such will be very handy to refer back to later should I need inspiration for this stage. Also the article uses both objective and subjective treatments to argue that many-many (complex) strategies provide much more intimate and expressive music making experiences. One interesting example is that of the "accidental theremin". The paper also provides a better description (in my opinion) of the results and conclusions presented in Instrumental Gestural Mapping Strategies as Expressivity Determinants in Computer Music Performance by Rovan et al.

Most Important/Interesting Publications on DMIs

In this post I've collected what I consider to be the most important publications on DMIs that I've found with regards to my project. I've posted them here for quick reference and also in order to filter out the most useful items for printing and annotating.

• Gestural Control of Music - Marcelo M. Wanderley
• On the Choice of Transducer Technologies for Specific Musical Functions - Marcelo M. Wanderley et al
• Towards a Musician's Cockpit: Transducers, Feedback and Musical Function - Roel Vertegaal (second entry)
• Gesture - Music, Claude Cadoz
• Mapping Performer Parameters to Synthesis Engines - A. Hunt and M. Wanderley
• Expressiveness and Digital Musical Instrument Design - Daniel Arfib et al
• Physical Interface Design for Digital Musical Instruments - Mark Marshall
• Digital Lutherie - Crafting Musical Computers for New Musics' Performance and Improvisation.
• Towards a Model for Instrumental Mapping in Expert Musical Interaction - A. Hunt, M. Wanderley, R. Kirk
• Sensor Choice for Parameter Modulations in Digital Musical Instruments: Empirical Evidence from Pitch Modulation - Max T Marshall et al
• Musical Vibrotactile Feeback - David M. Birnbaum
• A Consort of Gestural Musical Controllers - Joseph Malloch (first result)

Cajon Drum

I've been aware of this instrument for a while not but I've only just gotten round to looking it up and finding out its name. There are a few different versions of this drum (pronounced cahon drum) within the family:

Perhaps the most well known is the cajon drum itself. The cajon drum is simply a rectangular wooden box. Five of the sides are thick plywood and the side which forms the playing surface is a thin piece of plywood. The back side typically has a circular hole to allow air to escape and project the sound. The typical playing position is to stand the cajon drum on its base so that the player can sit on top and hit the drumming surface with his fingers/palms/fists between his legs. I've also seen/read that players may tilt the drum backwards whilst playing. This implies to me that the bottom of the drum is left uncovered to allow air to escape like an african drum but I'm not sure whether this is common design practice given the hole in the back.

It is common to tension steel strings across the playing plate on the inside of the drum to give it a snare like effect in that position. The strings are often actual snare drum strings or occasionally steel guitar strings.

Another version is the Cajon Bongo. This is a smaller version of the drum which typically sits length ways across the lap. One long side of the drum is left open, this side faces downwards between the legs. The opposite side is the playing surface. The inside of the drum is commonly split into two compartments, one big (longer or deeper) and one small (smaller or shallower). This allows the production of two main tones from the drum, a deep one for the big compartment and a higher one for the smaller.

- Here is a practical demonstration of both drums (focusing on the bongo but demos of the big brother are available by the same author)

- Bela Fleck and The Flecktones - Over the Wall: The Cajon drum included in a band context. See around 4.26 for a percussion solo demonstrating the drum.

- cajondrum.org

- cajondrumming.com

- cajon-drums.com (seems fairly authoritative)

- Wikipedia

The reason I mention this instrument is in relation to some design problems I've been having with my DMI. Because I decided that I would like to try and implement some vibrotactile feedback in my instrument, most of my recent (basic) designs have involved one or several loud speakers, taking inspiration from Mark Marshall's DMI The Viblotar which uses loudspeakers to localize the sound source to the instrument itself and also to provide a degree of vibrotactile feedback (with options for some clever synthesis implementations). My problem however is that most of designs are coming out as square shapes in order to accommodate the speakers!

This seems like a problem for several reasons. For a start, the designs are not particularly innovative. Most of them look like actual loudspeakers with a few sensors stuck onto them. Also a lot of my DMIs have been designed to hold and play which may not be easy as things stand due to the necessary size required for a loudspeaker enclosure, the awkward shape of a big box and also the potential weight of the instrument.

After looking into the Cajon Drum I'm beginning to think that may be a big box may not be such a bad idea, if I take some inspiration from the drum:

- The Cajon Drum is an existing instrument with an existing gesture vocabulary, there by giving first time users perhaps some idea of the kind of techniques that can be used with such a DMI. The controller would thus be classified along the lines of a instrument-like or instrument-inspired controller (perhaps dependent on the direction I went in with the software, ie imitating a drum sound or some thing a little different...)

- The Cajon Drum is a reasonably popular percussion instrument therefore there will be a good repertoire of material that could be tried on the DMI. This meets some of the ideas behind approaches to DMI design that I mentioned in a recent post

- Embracing the box design of the drum means that speakers can be reasonably easily housed in an instrument design which already exists as an expressive acoustic instrument which should address any worries regarding playing style / efficiency

- A box shape (with justification!) is perhaps actually the best design choice since it would make the woodwork and construction fairly simplistic. This is perhaps especially advantageous in this project given my time limit, the notorious difficulty of software and mapping and also outside commitments. I think its perhaps important to remember at this point that I am in no way attempting to rewrite or improve the book on DMI design. Rather I am trying to use what I have learnt from the book to produce a DMI which could be considered to have a good "feel"

- The closed box design will also allow all the electronics of the DMI to be hidden inside. It has been suggested in studies that DMIs which have visible external wiring are often seen by performers as being "experiments" and being "fragile" which perhaps restricts the level of involvement a performer could have with an instrument

- Building an instrument designed to allow the player to sit of top of it will naturally result in it being very robust which, allow with internal electronics should reduce the impression of DMIs as being fragile experiments and allow the performing to really get involved in their playing without worrying about damaging the instrument. This should also have other obvious benefits such as providing a good amount of protection to sensitive electronic components and also being reasonably resilient to bumps caused by transport etc

- One interesting thing about the "box" designs I had previously come up with is that they often encouraged the used to hold them and wrap their arms around them in order to reach the playing surface. Although granted this playing position would have been too uncomfortable it did ensure that a large proportion of the instrument body was in contact with the players body. This would be especially advantageous for the vibrotactile feedback element of the DMI. This Cajon design should still hold this advantage since it encourages the user to sit on top of it!

I think the above statements make a compelling argument for using the Cajon drum as inspiration. I think at this point and after all this research, that finalizing the fundamental design choice in this manner is the next step to take and it would be good to advance in this way. There are however, some important points that I would have to consider if I took this direction:

- Gesture capture system. The main method that I had in mind was to use a piezo microphone mounted on the inside surface of the playing plate to monitor the vibrations of the plate and the rest of the instrument. Some thing to consider about this approach is that it would involve a certain degree of signal processing which could become fairly complex. For example I would have to distinguish between the vibrations caused by the loudspeaker feedback system and the vibrations caused by the actual player. There are possibly some interesting things to be done here such as distinguishing between amplitudes to trigger different effects and possibly distinguishing playing location and technique. On a related note here's an example of a commercially available Cajon Drum which includes a "DST pickup microphone" which I believe is a contact mic of some kind. The mic is mounted on the inside and outputs through a 3/4 inch jack plug with volume control.

Meinl Cajon with Microphone pickups

- If the same gesture vocabulary as the original instrument is used then the DMI is going to give primary acoustic feedback from the players hands hitting the wood whilst also giving the secondary acoustic feedback from the synthesizer. The question is do I want this primary feedback and if not, how shall I remove it? In the DMI the speaker will stand in for the air hole so this should provide some damping. I might also consider packing the inside with soft material, much like stuffing a bass drum with pillows.

- What kind of software approach should I take. I'm thinking more of what are the sounds that I would like this instrument to make. They needn't be percussive. Interestingly, I think that if I used percussive sounds exclusively then this DMI would perhaps be classed as instrument-like since it is designed to look, be played and sound like a cajon. If i used some other type of sound though it would perhaps be instrument-inspired since it is based on the concept of a cajon drum but is designed for a different purpose.

- One element that could inform the sound choices I make is the choice of transducer technologies I include with the DMI. I think placement of transducers here will be an especially important decision given the fairly unique playing position of actually sitting atop the instrument and bending down over it to play. Looking at some videos the arms tend to be fairly stretched at any given moment and their left-right movement is limited by the position of the legs. Two Cajon designs could possibly help with this:

The Cajon Bongo since the instrument is placed on the players lap and the player typically interacts with the top surface on which could be placed the transducers for easy access. With this design the controller would not be too dissimilar to Mark Marshall's Viblotar

The angled surface Cajon since the front plate playing surface is angled to slope downwards away from the player in order to allow slightly easier access to the various tones available from the edge of the plate to its center. In the same way it would allow easier access to any transducers which were placed on the plate. On a side note I'm not fairly convinced that typically the Cajon Drum does not have an open bottom and the reason for tilting the drum backwards during play is again to allow easier access to the full playing surface.

Here are some Cajon materials which are perhaps slightly more authoritative and better references than those given at the top of the post. I'll keep all Cajon stuff within this post for now until I take an active decision to follow this line of design. Could do with some even more authoritative works and perhaps some scientific studies. So far after a little google searching I can find no examples (aside from simple amplification) of digital or analogue modified Cajons.

• Frequency and Dynamics Analysis of Bass Tone of Cajon Box Drum - Peter Kicak - This article is fairly bad from what I've read especially since the English is rather poor, but the first section does give some background on the Cajon which I presume is reasonably reliable.
• Studies for Cajon - Martin Rottger - just a book preview entry found on google scholar but again there is a history section and also some information regarding damping the Cajon and about the tilted playing style

Commotion + Contingence

This section of David Birnbaum's blog presents two data-glove style DMIs created by David - Commotion + Contingence. Commotion is a drum track sequencer type instrument and Contingence is an FM synthesis controller providing a pentatonic scale. A head mounted camera is positioned in front of the mouth to monitor the mouth's shape (and possibly some other things to do with the tongue) and is used to control a delay effect. The instruments are designed to be able to produce sound texture pieces rather than pieces based on melody, harmony and rhythm. The page includes details of both instruments as well as a recording of the only composition to be written for the instruments named Alignment. Also included is an interview (non-technical) with David Birnbaum in which he gives a small demonstration of the instruments and briefly describes how to play them.

David Holt's Kinda Music

A series of clips from Folkways, an American TV Show that multi-instrumentalist and old-time country/blues/folk music enthusiast David Holt used to run. The clips present various well known names from the hayday of those musical styles in order to commemorate their contribution to music. I've included here just one's of some slightly more unusual examples of musician's and their instruments in an effort to illustrate that Human's can make music and have fun with almost any thing. I'd suggest that we've been doing so in one form or another for a probably surprising amount of time and will continue to do so indefinitely. From a couple of old bones to emerging DMI controllers taking advantage of the most advanced digital synthesis algorithms available today.

Washboard Bill's Washboard and David Holt playing spoons

Percy Danforth playing bones and David Holt on Guitar, Banjo, Harmonica and paper bag (!)

David Holt playing a whole bunch of instruments from guitars to his own body. Of particular relevance is the Electric Pajamas presented last. This last piece although some thing of a novelty really reminded me of my thinking that the development new age digital music and instruments needs to be guided by our own physical and emotional (not technological) connection with the music. This is some thing that I'm striving to keep in mind and why I'm becoming so interested in topics concerning DMI haptic design and vibrotactile feedback amongst other DMI design principles relating to the connection between the musician and the resulting sound.

David Birnbaum's Blog

Tactilicio.us - Research David Birnbaum's blog. Includes the previous two articles as well as others and descriptions of various projects including topics on DMI design as well as his involvement with the McGill Digital Orchestra. Some potentially interesting resources here as well as some banging electronic tunes.

Musical Vibrotactile Feedback - David M Birnbaum Master Thesis

Musical Vibrotactile Feeback - David M. Birnbaum

David M. Birnbaum's Masters Thesis at McGill University

Especially interesting in this thesis is chapter 2, which presents a very convincing argument as to why tactile interaction and feedback is so important in musical interactions citing many disciplines from philosophy to physiology, as well as chapter 3, which presents a physiological overview of the sense of touch in humans and identifies key aspects of the sense with regards to a DMI. Chapters 4 and 5 present the software and hardware (respectively )implementations of 2 flute-like DMIs (the former being a prototype for the latter) which utilize a fairly sophisticated approach to vibrotactile feedback based on the findings presented in chapter 3.

An interesting point is also made in section 1.4 - Methodology regarding creating DMIs with the intention of playing a more traditional repertoire of music, rather than creating modern experimental sound scape type pieces.

After naming a number of instruments born from technological advancements (acoustic and electronic) which were originally conceived to perform existing roles/functions/pieces in a new way but after some time contributed directly to the innovation of new styles and musical roles Birnbaum suggests that:

"Working within known musical styles may be more likely to generate instruments that invigorate widespread creative movements than design theories that seek to subsume the "expressive" characteristics of acoustic instruments with digital technology. Rather than starting with the goal of creating a new kind of music, new instrument design can reasonably be guided by enabling existing music to be played in a new way. Instead of limiting expressive potential, this approach may even enhance it."

It seems to me to be a good suggestion that we use the high expressive potential of existing acoustic instruments as a bench mark for our new DMIs. Not in an attempt to match or better it but simply to serve as a reference by which we could say if our approach to DMI design was producing new instruments which were at least heading in the direction of expressive capability. DMIs are a relatively new concept and it seems logical to start at the beginning and see if we can produce instruments we can judge based on our existing concept of a "good" instrument given existing repertoires and continue from there, rather than starting by reinventing the wheel. (NOTE I'm sure that a similar opinion was expressed in one of these papers I've posted recently. Would be worth looking that up - it was: Problems and prospects for intimate and satisfying sensor-based control of computer sound by Mathew Wright)

Potentially interesting is the related project page at IDMIL (Input Devices and Music Interaction laboratory) .org which presents an over view of the Touchflute and the Breakflute presented in the thesis. See the article - Touch Flute: Exploring roles of vibrotactile feedback in music performance - referenced at the bottom of the page.

A Systematic Approach to Musical Vibrotactile Feedback

A Systematic Approach to Musical Vibrotactile Feedback - David M. Birnbaum and Marcelo M. Wanderley

DMI Seminar at McGill University

According to joseph Malloch's thesis.

Approximately every two years an interdisciplinary seminar on Digital Musical Instruments

is offered at The Schulich School of Music. In contrast to seminars on HCI and sensing

offered as part of the Music Technology program, this seminar is offered equally to performers,

composers, and music technology students, and is taught jointly by Marcelo Wanderley

from the Input Devices and Music Interaction Lab (IDMIL) and Sean Ferguson, who heads

the Digital Composition Studio (DCS). Students are asked to form groups in which each

discipline is represented, and are tasked with designing, constructing, composing for, and

performing on a new digital musical instrument over the course of one semester. This

extremely accelerated schedule has nevertheless resulted in some very interesting projects,

from both technological and artistic perspectives. Video recordings of some of the past

seminar performances are available on the DCS website [52].

The Schulich School of Music is part of McGill University. The video recordings mentioned can be found on this page at the top, links to a small clip and the full presentation. No doubt some interesting things to be found. Further down the page are some images of some of the instruments presented along with their names for further reference.

A Consort of Gestural Musical Controllers - Joseph Malloch

A Consort of Gestural Musical Controllers - Joseph Malloch (first result)

Joseph Malloch's Master of Arts Thesis with McGill University on the subject of the creation of the TStick family of instruments

The TStick: From Musical Interface to Musical Instrument - Joseph Malloch

The TStick: From Musical Interface to Musical instrument - Joseph Malloch and M Wanderley

Mapping Strategies for Musical Performance

"Mapping Strategies for Musical Performance" - Andy Hunt and Ross Kirk

This is a brilliant study looking at how different kinds of interfaces and mapping complexity effects one's ability to perform musical tasks of varying complexity. The paper goes in depth on a number of principals and presents test results in detailed graphical form, eventually concluding that:

1. Real-time control can be enhanced by the multiparametric interface

2. Mappings that are not one-to-one are more engaging for users

3. Complex tasks may need complex interfaces

4. The "mouse interface" is good for simple tests and for little practice time

5. Some people prefer to think in terms of separate parameters

Importantly the tests carried out in this paper attempting to take into account learning time for a user. Results were taken from practice sessions across different ranges of time from short (3 sessions) to longer term (9 sessions). Interesting to note however that I have not found a study yet that looks at the effects of learning across time periods that are closer to those required to develop even a basic proficiency at a given acoustic instrument (ie perhaps 1-2 years)

Problems and prospects for intimate and satisfying sensor-based control of computer sound

"Problems and prospects for intimate and satisfying sensor-based control of computer sound" - Matthew Wright

This paper presents several challenging areas of DMI design. I quite like this paper because throughout the descriptions of problems are complemented by some of Wright's personal believes and principals when it comes to designing instruments and many of his sentiments I agree with. A lot of his statements also backup what other author's that have inspired me have said.

Wright makes an early point as to why it may be advisable to compare our new DMIs to traditional acoustic instruments and even why we may attempt to emulate aspects of their design:

" Systems along the lines of Figure 1 naturally invite comparison with traditional acoustic
instruments. Of course the point of the comparison is not to
attempt to decide whether acoustic or electronic instruments
are “better”; each have their strengths and weaknesses. My
point is that the finest acoustic instruments set standards for
nuance, intimate fine control, dynamic range, temporal
responsiveness, and the possibility for a performer to develop
virtuosity, and that electronic instruments have definitely not
surpassed acoustic instruments in all of these areas. The
challenge to meet or come near these standards with our
electronic instruments is a powerful force driving
development, and this development will naturally include
some solutions that can be generalized to non-musical uses of
sensors. "

Importantly though, he does not claim that acoustic instrument designs should be the be-all and end-all of DMIs:

" I believe that the emulation of
acoustic instruments is an interesting goal only to the extent
that it increases our general knowledge of sensor-based
instruments or results in hybrid instruments that combine the
strengths of electronic sound production with the features of
acoustic instruments. "

Else where in the paper Wright gives some frankly intimidating figures regarding latency and jitter and makes some interesting points about economy of motion and gesture transparency. Finally he concludes by stating:

" Complex real time sensor-based electronic music systems have
a staggering range of possible ways to fail, and it is depressing
how much time the performers of these instruments spend
troubleshooting them. "

It is interesting to see how many authors make points about instrument reliability and how performers should spend as much time playing and as little time repairing as possible, with some DMI creators going to extra lengths to make sure repairing their DMIs is as simple as changing a guitar string and requires no specialist knowledge. This all goes to show the amount of things one must consider if one is trying to design a DMI which could truely be called a musical instrument.

Instrumental Gestural Mapping Strategies as Expressivity Determinants in Computer Music Performance

"Instrumental Gestural Mapping Strategies as Expressivity Determinants in Computer Music Performance" - Joseph Butch Rovan, Marcelo M Wanderley, Shlomo Dubnov and Philippe Depalle

This paper sets out to test a hypothesis that convergent (many-to-one) type mapping schemes should give the most intimate control over a synthesis algorithm since it is most like typical instrumental physical phenomena. The example used to test the theory is a Yamaha WX7 wind controller controlling a rather complex additive synthesis system via several different mapping strategies (from one-to-one to a revised many-to-one). Issues are encountered with the synthesis system however and it seems that thorough testing and subject and objective results are never completed (or at least presented) although the author's conclude that:

"the mapping layer is a key element in attaining expressive
control of signal model synthesis." (interesting to note their hypothesis is only applied to signal models. I suppose we can mostly write off physical models since the mapping is largely built into the system and any mapping from controller sensor values to model inputs are typically one-to-one)

As well as:

"In an instrumental approach, the conver-
gent mappings demonstrated in this paper have the
potential to provide higher levels of expressivity to
existing MIDI controllers."

Interestingly the authors point out that:

" Without the need to de-
velop new hardware, off-the-shelf controllers can be
given new life via coupling schemes that attempt to
simulate the behaviors of acoustic instruments"

Which is a point that I don't think many other authors have directly made, considering that perhaps we can make existing commercial controllers perform better without building completely new controllers.

Mark T Marshall E-mail Response

Below is the main advice given by Mark T Marshall:

Regarding your project, I think it is an excellent idea. Be aware however that designing a DMI that is as good as an acoustic instrument is quite difficult. One of the major issues, particularly with a year long project, is that you (or any performers) have a limited time to learn the instrument. This means that it needs to be quite easy to pick up, but it should also be interesting to play. This is a point made by Wessel and Wright in [1]. The other major issue is proving that the instrument is any good. The best way to do this is to have musicians (ideally a number of them) play the instrument. Musicians know what has musical potential and what doesn't. Feedback from them will help guide the design and also can be used to show that the instrument is actually any good. This area (evaluating a DMI) is something I am working on at the moment, and is not an easy thing to do.

There have been lots of projects to build new DMIs. The NIME proceedings are full of them for example. However, very few of them have really been designed to compete with acoustic instruments. One excellent thesis dealing with this is by Newton Armstrong ([2] in the reference list below). I would strongly recommend you read it. It's a bit hard going at times, and I really don't think the final instrument follows the guidelines he sets out, but it is the best discussion of the design process I've seen.

One of the few really successful instruments I've seen when it comes to engaging performers is Joe Malloch's T-Stick. There's a bunch of information about it on http://www.idmil.org in the projects section. I suggest you have a read of some of the papers written about it also.

Finally, you need to put a lot of effort into the mapping. When we worked on the McGill Digital Orchestra instruments, this was the part that we found needed the most work. It helps if this is done along with any performer's who will be playing the instrument (if there will be any other than yourself). Have a look at [3] and [4] for information on how we went about doing this. Also on the subject of mapping, check out the papers in the special issue of Organised Sound on mapping [5].

My main advice for the project itself is to start playing the instrument as soon as possible. While you can use the existing literature to guide you, it's only once you've tried to play the instrument that you can tell if it will work at all.

[1] Wessel and Wright, "Problems and prospects for intimate musical control of computers"
[2] Newton Armstrong, "An enactive approach to digital musical instrument design"
[3] Sean Ferguson and Marcelo M. Wanderley. "The McGill Digital Orchestra: Interdisciplinarity in Digital musical Instrument Design"
[4] Xenia Pestova, Erika Donald, Heather Hindman, Joseph Malloch, Mark T. Marshall, Fernando Rocha, Stephen Sinclair, D. Andrew Stewart, Marcelo M. Wanderley, and Sean Ferguson. "The CIRMMT/McGill Digital Orchestra Project".
[5] Marcelo M. Wanderley (guest editor). Mapping Strategies in Real-time Computer Music. Organised Sound, 7(2), August 2002.

Mark T Marshall Suggested Reading

A list of key articles on various subjects within the domain of DMI design suggested by Mark T Marshall:

Wessel and Wright - "Problems and prospects for intimate musical control of computers"

Newton Armstrong - "An enactive approach to digital musical instrument design" (first result)

Sean Ferguson and Marcelo M Wanderley - "The McGill Digital Orchestra: Interdisciplinarity in Digital musical Instrument Design"

Xenia Pestova, Erika Donald, Heather Hindman, Joseph Malloch, Mark T Marshall, Fernando Rocha, Stephen Sinclair, D. Andrew Stewart, Marcelo M Wanderley and Sean Ferguson - "The CIRMMT/McGill Digital Orchestra Project"

Marcelo M Wanderley (guest editor) - "Mapping Strategies in Real-time Computer Music." Organized Sound

Most Useful Articles in NIME Archives

This post will be an attempt at organizing and categorizing some of the most useful/interesting article entries in the NIME convention archives, presented here. This post will likely be continuously updated for the near future so it may be worth checking it every now and again. I'll attempt to categorize the articles into different subject groups:

INSTRUMENT LIKE/INSPIRED CONTROLLERS

AUGMENTED INSTRUMENTS

• On Making and Playing an Electronically-augmented Saxophone '06
• Using the Augmented Trombone in "I Will Not Kiss Your F.ing Flag" '06
• Augmenting the Cello '06
• The Augmented Violin Project: Research, Composition and Performance '06
• Composing for Hyperbow: A Collaboration Between MIT and the Royal Acadamy of Music '06
• The Augmented Djembe Drum - Sculpting Rhythms '06

NOVEL CONTROLLERS

• The G-Spring Controller '06
• Tenori-On '06

MAPPING

PHYSICAL CONTROL/INTERACTION

• Paper FSRs and Latex/Fabric Traction Sensors: Methods for the Development of Home-Made Touch Sensors '06
• Vibrotactile Feedback in Digital Musical Instruments '06

GENERALLY INTERESTING/DISCUSSIONS

• The 'E' in NIME: Musical Expression with New Computer Interfaces '06
• Towards a Gesture Description Interchange '06
• Towards a Coherent Terminology and Model of Instrument Description and Design '06

Exploring the Effect of Mapping Trajectories on Musical Performance

Exploring the Effect of Mapping Trajectories on Musical Performance - Doug Van Nort and Marcelo M Wanderley.

The report considers mapping as a part of the DMI system and presents a discussion on how the mapping strategy employed in a DMI effects the "feel" (in terms of expressivity and ease-of-use) of that DMI.

Gestural Control at IRCAM - M. Wanderley

The article can be found here, first result. The article provides a brief over view of the activities in the area of gestural control of music carried out at IRCAM and by its associates between the years 1996 and 2001.

Sensor Choice for Parameter Modulations in Digital Musical Instruments: Empirical Evidence from Pitch Modulation

Sensor Choice for Parameter Modulations in Digital Musical Instruments: Empirical Evidence from Pitch Modulation - Max T Marshall et al

Article presenting various results of research carried out into appropriate sensor choices for musical applications in DMIs focusing on the task of pitch modulation (vibrato)

The HTML version can be found above along with a link to download the PDF.

CIRMMT

CIRMMT - the Centre for Interdisciplinary Research in Music Media and Technology

The McGill Digital Orchestra Project

The McGill Digital Orchestra Project

The McGill Digital Orchestra is a research/creation project supported by the Appui à la recherche-création program of the Fonds de recherche sur la société et la culture (FQRSC) of the Quebec government. Team members include Professors Denys Bouliane and Sean Ferguson of the Composition Area, Professors Marcelo Wanderley, Gary Scavone and Philippe Depalle of the Music Technology Area, and Professor André Roy of the Performance Department. All participants are members of the Centre for Interdisciplinary Research in Music, Media and Technology (CIRMMT). The grant has a duration of three years and will culminate with a performance of new works during the 2008 MusiMarch Festival.

Especially interesting is the projects page which lists the DMIs involved in the orchestra and provides links to more information.

Also see this paper for more detail about the orchestra and the pieces that have been written for it.