by leonie.dunke -

19 | Prototype – Visual Representation

In my last blog entry, I experimented with whether the vibration of the larynx alone could be used to read the pitch. Since this worked and I know that this technique is the basis for the concept, I have continued to work on the visualisation.

For my prototype I used Touchdesigner to compare my singing with the pure tone frequency and to see whether the information from the tuner was correct. As a result, you could hear that it worked.

The next step was to build a 3D model that visually represents my prototype. Part one is a small device that is placed close to the larynx to measure the vibration frequency. The analysed frequency is transferred to either a wristband or a clipper on the notebook and tells you if you have sung too high or too low.

With a simple interactive visualisation I show how the whole thing could work. The sung note is first compared with the pure frequency. If both match, the blue light in the centre lights up and if, as in my recording, the pitch was too low, the lower sphere lights up red. If the pitch is too high, the upper light lights up red.

The final part of my prototype is an app prototype with the idea to upload the choir score digitally qnd connect it to the vibration sensor. The sung notes are compared with the sheet music, and any inaccuracies are highlighted with a blue or red line, indicating where you are in the song and whether the pitch and timing are correct. I aim to keep the visualisation in the app simple, yet professional enough to provide a detailed vocal analysis.

Problem

One problem I had was reaching my choirmaster. Since I couldn’t go to rehearsals after the test and didn’t receive any feedback on my emails, I unfortunately couldn’t get any deeper insights to be able to use them in time to push the prototype forward.

by leonie.dunke -

18 | Prototype – Test

I tested the string tuner I had bought by singing at different pitches and trying to find the best possible position where the vibration could be read most easily.

Test Result

  • The frequency meter should be positioned close to the centre of the larynx, as this is where the vibrations are strongest. This means that the device does not have to be pressed so hard against the neck.
  • Functionality: Detection by vibration basically works, but there are strong fluctuations in the display.
    Tone change: When the position of the device was changed, different tones were displayed at different points, although the same pitch was maintained. Half-tones are also recognised.
  • Sensitivity: The device needs to be more sensitive for smooth use in order to provide more precise results.
  • Current use: The device is currently used for string and side instruments.
by leonie.dunke -

15 | Singing Aid

During my exploration of digitalisation in theatre, I noticed that I was not making progress and had difficulties in developing concrete solutions. In our other project „Projection Mapping“, where we create stage visualisations, our group places great emphasis on real-time visualisation and audio reactivity. Therefore, my thoughts often revolved around sound, music, audio, and real-time feedback. This made me aware of an issue in my everyday life that, while not directly related to theatre, is still artistically relevant.

At the beginning of the semester, I decided to join a choir and start singing again. I quickly realised that little remained of my past choral experience and my voice was equally rusty. During rehearsals, I had great difficulty hearing myself to determine if I was hitting the notes and the rhythm, especially in four-part singing. It is enormously difficult to hit the correct note precisely after a long break. I often wished to receive direct feedback during rehearsal to understand if my self-perception matched the actual singing and to adjust the pitch immediately.

I have developed a small prototype of how I envision such assistance. Initially, it was important to determine which aspects make up singing and where direct feedback is useful:

  • Pitch
  • Beat rhythm
  • Melodic rhythm
  • Volume
  • Emphasis, pronunciation

For the prototype, I focused on feedback regarding pitch and melodic rhythm.

Idea

My idea is based on a small device that provides direct feedback through vibration about correct and incorrect intonation. If sung too high or too low, it gives corresponding vibration feedback. An extension of this tool could also relay the rhythm of the piece through vibration feedback.

Concept

The tool consists of a small microphone and a vibration motor. This device can be clipped near the mouth on one’s clothing, capturing the individual’s singing and vibrating if sung incorrectly.

Another type of silent feedback could be sending a small light signal. With this approach, it would be possible to connect a small type of clip with the microphone that can be attached to the music score. This too would signal errors by lighting up if the singing is too high or too low.

Since analysing mistakes after rehearsal is also crucial, an AR app would be a possible implementation to display errors directly on the music score. If the score is also available digitally, this could be directly entered in the app. This allows for preparation for the next rehearsal.