Computational Analysis of Yaredawi YeZema Silt in Ethiopian Orthodox Tewahedo Church (EOTC) Chants

Mequanent Argaw Muluneh, Yan-Tsung Peng, and Li Su

Institute of Information Science, Academia Sinica, Taiwan

Department of Computer Science, National Chengchi University, Taipei, Taiwan

• Finding of True Cross: 27 September

• Ethiopian Epiphany: 18-19 January

EOTC chants are currently used for 10s of millions people.

The EOTC chants – Background

UNESCO Registered Holidays: chants are main components

The EOTC chant - Yaredawi Zema

Yaredawi Zema

Saint Yared

Ge’ez

Ezil

Araray

Silt

(Mode)

Unique to Ethiopian and Eritrean Orthodox Christian Churches in the world

NB: plural of silt is siltoch

Created by

The EOTC chant - Yaredawi Zema

Yaredawi Zema

Saint Yared

Ge’ez

Ezil

Araray

Silt

(Mode)

Unique to Ethiopian and Eritrean Orthodox Christian Churches in the world

NB: plural of silt is siltoch

• Challenges in computational Analysis of Yaredawi Yezema Silt:
• No public dataset for MIR tasks
• Limited previous works tailored to the chants.
• Shelemay et al. in 1993 is prestigious one so far.

Yezema Silt Classification

• We aim to investigate if computational approaches match the western perspective to the modes.

​

“Araray and Ezil both use the same pitch set … ambiguity in these modes … is the product of two factors…”

(Shelemay, Jeffery, and Monson, 1993)

Yezema Silt Classification

Findings from our CNN based classification:

1. Ambiguity exists between Ezil and Araray.
2. YeZema silt is a song-level property to be described in pitch histogram features .

​

• The Yaredawi YeZema Silt Dataset:
• 10.2 hours long,
• 369 recordings (75 Ge’ez, 176 Ezil, 118 Araray)

Computational Analysis

Findings:

1. the pitch sets of the three modes are distinctive!
2. a mode is characterized by its pitch variances.

• Pitch histogram features for further GMM based analysis