Making the 'Performer' Obsolete

As I had decided I wanted my system to be able to operate purely stand alone with the computer acting as the ‘performer’ carrying out functions on the various outputs based on the information it is getting from the inputs. I decided it would be best to go for a very similar flow to Di Scipio, I have included an image of this flow below. In my case however room should simply be replaced by the springs in the flow.

The main agent of change in this scenario has to be DSP as other elements of the system are static and not able to effect change in themselves (this is where the role of a human performer could come in). In order for the DSP processes to work effectively I need to consider which factors I want to measure as input to the DSP. Possibilities that could be measured include amplitude, pitch, density of events and many more possibilities. One possibility would be that when the low energy reaches above a certain threshold low frequencies are reduced by DSP. These control signals can then be routed to any number of variables within the DSP.

By making use of other signal variation possibilities such as pitching, delays, granulation, amplitude following et I may be able to create even more complex evolving sounds through the system. For example by using the control signals to control the length of a delay or use an amplitude follower to only bring in certain parts of the system when it reaches above a certain noise threshold. The DSP is the most exciting element of this whole process as it is this that will bring life to the concept and bring it above just solid and uncontrolled feedback.

For this processing I will either make use of Supercollider or MAX MSP as they are both very capable of real time audio processing. Another advantage of this is that I could also easily map effecting parameters to external MIDI controls allowing a ‘performer’ to interact with the systems ‘insides’ without having to make any physical interaction with the system.

Design Pt.1

When designing the interactive physical feedback system the original core concept was that I would use multiple springs attached a speaker at one end and individual contact mics at the other. When the contact mic is fed and amplified back through the speaker it would cause pitched feedback based on tension, length and density. By having multiple springs attached to the same speaker I was hoping to cause some sympathetic resonances in either spring that I could then either mix back in to the speaker for more feedback or send to a separate feed just for playback.

Later on this idea developed into using multiple versions of the set up pictured above hooked up through a more complex routing setup. The concept here was that I would then be able to route a feedback ‘path’ through multiple speakers and springs that would add to the sonic complexity of the system and also mean that all parts of the system were intrinsically linked and would react to changes in each other states and changes in the environment. In this scenario I envisioned having 3 separate speakers each mounted on separate platforms in the performance space.

The image above depicts this set up with just one of the many possible feedback ‘paths’ that could be followed when in use. As can be seen there is a lot of scope for very complex signal flows through this system and I was sure that there would be a lot of potential for the system. Other possibilities here were that as well as being able to function as a completely stand alone no performer required installation once DSP had been sorted. A performer would be able to interact with the system and any change at any point in the feedback path would result in system wide variation.

Linear Looping?

Above is a video showing a live performance using hardware. As may be clear this is obviously not a strictly linear process, there is a form of feedback loop going on in the ears and brain of the performer. As he takes in the sound he has created he is then using his brain to make judgments about what he should do next as a compositional decision. However this loop is purely optional, the system would still create sound even if the performer was not aware of the output material, for example if the output was fed to a speaker in another room! 

In order to move from the concept of interactive composing to composing interactions we must have ‘self-observing system’. That is, it must make changes to the output based on the input it receives from the environment surrounding it, of which it is an inherent part. This means that a feedback loop is created in which a human performer is not required but could instead be considered a part of the macro-environment in which the whole ‘performance’ takes place. 

A system of this nature could almost be considered self-aware, making changes to its output based on the input it’s receiving from its self. However a better description as used by Di Scipio is self-observing. as the computer does not know the relationship between the output sound and the input parameters it merely views them as separate sets of data.

Initial Thoughts

Feedback has always been of interest to me, coaxing sounds from the guitar as I move around the room, placing the it on the amp to increase the physical contact and induce yet more feedback. Being able to ‘play’ feedback I always enjoyed as it feels very primal and so close to just going out of control. it takes work to keep it in check.

A video posted by Matt Teal (@matthewteal) on Jul 2, 2014 at 4:13am PDT

I am considering creating an interactive physical feedback system using multiple springs connected to multiple speakers. I have taken inspiration from Agostino Di Scipio’s Audible Eco-Systemic Interface (AESI) project as well as Nicolas Collins Pea Soup.

I took particular interest in his idea of shifting from interactive composing to composing interactions. This leads to a shift from creating tools for composing in real time using a human agent as the driving force behind the development of the piece to tools that will react to changes in environment that they themselves have caused creating a feedback loop through the system in which it is not necessary to have a performer. 

Most live electronics follow a linear flow where the performer makes use of some sort of control device which has a effect on synthesis parameters and processing etc which then leads to sound creation(this is also observed in live electronic music performed with hardware where realtime parameter changes are effected by the performer and then made audible). This flow can be applied to conventional instruments as well, in this flow there is little room for change/innovation other than novel tactile control? (with regard to control over performance, not the sonic material itself).

It is also worth noting that without the performer a system of this nature will not produce any sound, the performer is the only agent of change.

INTERACTION DESIGN

Interaction Design Is an approach to design developed by Bill Moggridge and Bill Verplank. It is focused on the behaviour of the end user and how they would best interact with the product/object, or in my case instrument.

(Bill Verplank, 2000)

Above is a diagram illustrating the basics of interaction design taken from a lecture by Bill Verplank at CCRMA. First and foremost the do and feel are the most important areas of interaction design, essentially it's asking when a user interacts with an object what feedback will they receive from it to improve the interaction. For example in the case of levers where pulling the lever clearly moves it into a different space and also provides tactile feedback.

However simply acting on an object and feeling it's response is not enough for an effective design. The user needs some way of knowing what actions on the object will yield which results. In order to do this there are 2 main methodologies. One of these is to map out the interface so the user can see what actions will bring about the results they want and achieve them by taking any route through the object or software. Another way of doing this is with the path methodology essentially this means that the user has to follow a certain protocol to reach the desired output. 

Each method has their own advantages and disadvantages. The path methodology is better for new users as it provides step by step feedback. However this also lends itself to high efficiency when placed in the hands of expert users as there is less need for thought in the process. Maps however are based around the idea that there is not necessarily only one way to achieve the desired reaction and there may be alternatives available and different routes one could take to reach the same end goal. This can be seen all the time in software design where there are often multiple ways of achieving the same results.

References:

Mr. Bill Verplank (2000) INTERACTION DESIGN, Available at: http://www.billverplank.com/Lecture/ (Accessed: 28th November 2015).