FabCafe Taipei, AI-generative music x Traditional instruments x Live interactive performance

Review

After several AI in generative music meetups, there came about an opportunity for cross-disciplinary collaboration within the community. After receiving the invitation to perform “A Song of Wordlessness,” the Erhu musician WANG Ying-chieh started collaborating with Chi Po-hao, who has been investing in generative music for a long time, to experiment on a concerto piece which incorporates both traditional instruments and machine learning.

FabCafe coordinated this gathering before the official performance of “Song of Wordlessness” for collecting feedback and creativity from the community. Participants in the event included composers, performers, development creators of musical instruments, creators who work in both technology art and music, and researchers of AI voice recognition. In addition, by revealing the developing process of erhu and live electronic sound, the artists also hoped to find partners who are interested in this area to join the project.

Photo: The creators who came to participate asked several questions

At the beginning of the event, the erhu musician WANG Ying-chieh demoed an interaction with live electric sounds as a warm up. By letting the computer listen to the erhu’s bowing, instead of generating noise, the computer can synthesize small segments of sound such as wind and rain to interact with erhu. Erhu led to an impromptu performance with the machine and create a new chapter.

Photo: The erhu musician, WANG Ying-chieh

In the actual performance on the stage, the musicians have certain professional standards for themselves. However, based on the experience of previous implementations, the electronic sounds produced by machine learning or AI during live performances are relatively unpredictable, a fact that is in conflict with the musicians high expectations during the performance. This means that live performance with electronic sounds must reduce unpredictable factors (or “errors”).

On the day of the performance of “A Song of Wordlessness,” in addition to the erhu and electronic sounds, two artists from Japan and Korea will also perform the Japanese bamboo flute, the taiko drum, and the Korean changgo drum. At this point, the maturity of the system still can’t keep up with the requirements of the show. This leads us to the question: When we can only let system live interaction with erhu, how can we get closer to machine learning during this stage?

Photo: Discussion and feedback

The next challenge is to let the computer understand the various elements related to the playing of the erhu including its air sounds, glissando (sounds made by continuous slide upwards or downwards between two notes), spiccato (playing as rapidly as possible between two strings within incertain pitches in an extremely high position), and so on. If you are a good erhu player, why not join us to play a song alongside AI! Maybe the computer can also understand your bowing technique after training! We look forward to the performance of “A Song of Wordlessness” at Zhongshan Hall and the next gathering at FabCafe on April 23rd.

Photo: Participants in the meeting interacted with the electronic system with their own erhu bowing methods

About the AI in generative music meetup

This is a meetup that aims at the creation process and discussion of using AI for generative music. The meetups are intermittent with the hope to gather and mix creators and with AI developers who have worked on areas such as “parameter design,” “generative arts”. We hope to encourage a space to discuss, research and create how AI can be applied to music together.

About “A Song of Wordlessness”

“A Song of Wordlessness” is the first piece of the Yunshu Trio led by WANG Ying-chieh. It brings together traditional Chinese, Japanese, and Korean musical instruments to explore the lineages of three major Asian cultures, and to connect traditional performance techniques with contemporary music grammar and to explore toward innovative musical pieces. The new electro-acoustic artist Chi Po-hao is also invited to participate in the creation process. The first performance of this contemporary sound piece that integrates Chinese, Japanese, and Korean cultures will be held in the Zhongshan Hall on April 7th, 2019.

The original event report for this article can be found at the FabCafe Taipei blog.

For additional photos from this event: https://www.flickr.com/photos/fabcafetaipei/albums/72157707321836654

Erhu Interactive Electronic Sounds, “A Song of Wordlessness”

Author: Yipin Huang
Editer: Kelsie Stewart