Posts about “TEI w:4”

  1. Instrumental knowledge

    Sansa: A Modified Sansula for Extended Compositional Techniques Using Machine Learning

    McLean J. Macionis and Ajay Kapur 2018

    This feels like a school project, a good school project, but not one with very high academic standards or contributions. It feels like something that might be the result of our final project in TEI.

    I think it is a bt thin in motivating why they did it. There are many instruments that are easy to pick up and understand and a lot of them could probably house electronics. Why choose the sansula?

    Most of the features are just the result of electrifying an instrument, not particularly their implementation. The parts that are more unique feel a bit bolted on, would making gestures with the whole instrument have any positive effect for the musician. On top of that, those examples don't really use the instrument as such at all. So why not just use a sensor in any ol object then?

    Kontrol: Hand Gesture Recognition for Music and Dance Interaction

    Kameron Christopher et al 2013

    Kontrol seem to be in the same vein as Sansa in terms of the type of text. It seems a bit less cohesive as some of the examples of the Kontrol seem to be totally different devices.

    I think the saxophone example shows a lack of understanding of the instrument and context. When talking about the Guqin they go into great lengths to describe how the tone is generated and what makes it special but with the saxophone they seem to think that the important part is the hands. Thinking you can respond to the music by just knowing hand movement. That would be like just knowing what string was played on a guitar or the guqin.

    Furthermore I don't think most saxophonists that use filters would use them with a laptop, I think they would have pedals.

    Trying to cram some buzzwords into the text in the end does not do it any favors. What do they mean by "incorporation of nanotechnology"?

    Sensori-Motor Learning with Movement Sonification: Perspectives from Recent Interdisciplinary Studies

    Frédéric Bevilacqua et al 2016

    This text was better. It goes through some earlier research into the use of sonification in different fields.

    At times it feels like it rushes through what they did but there are some nice insights in it. The parts about what sounds are expected and how that can be used to increase the effect of the sound is interesting.

    Most of the research seem to be inconclusive and they want further research to explore the same issues. Some of the experiments seem to be replications of older experiments and have the same results. It's hard to see the contribution of the paper other than saying that this is a field that could be studied further. I think they are right in that though. This field seems interesting and it should be studied more.

  2. Photo by Franck V. on Unsplash

    Computer, make music

    This week we explore computer assisted music with the help of machine learning in Wekinator and some Processing.

    Before the demo today I ws pretty sure I was going to do this project in Node and the browser and use TensorFlow PoseNet there, but after seeing how easy it was to get going with Wekinator and Processing I'm not so sure anymore. I think we will just doi it with the tools Daniel demoed today and I can create some kind of web server that connects to Wekinator in my spare time.

    A web server bridge could still be useful as it could be used for some distributed systems like art installations that the audience can interact with. I think this is a good project for the future.

    Today also showed how much progress my class has made since last year with programming. It also showed in the last course but I ws impressed to see how people just started editing the Processing examples even though we have barely worked in Processing before. It really shows how experience builds confidence and how understanding of one language translates to other languages.

    Assignment: ML Music

    Brief: Design an instrument that uses a computer to generate sound

    Materials: Wekinator and Processing or other stuff

    Team: Cathrine, Denisa, Simon and me

    We started by looking through some examples and started poking in the MIDI output example but we moved on after a while as our understanding of MIDI is limited and we don't see the great payback in figuring out all of that in a week to get Logic Pro to work with it. We think we can make more with just Processing and playing samples.

    Simon made a nice example where he extended the drum machine example. He imported some nicer samples and and this will probably be a basis when continue work with Motion Leap tomorrow.