Sono and Leria 3 in the Test of Arabic Singing Performance: Simulation or Imitation?

Artificial intelligence algorithms “Sono” and “Leria 3” dominate the tech scene with their attempts to simulate Arabic singing, challenging the complexities of Eastern maqamat and the uniqueness of quarter-tone notes. This technological competition raises a fundamental question about machines’ ability to embody the soul of Arabic music.

Challenges of Simulating Arabic Singing

Specificity of Eastern Maqamat

AI systems face significant obstacles in comprehending the intricacies of Arabic maqamat, renowned for their emotional uniqueness. The algorithms lack the emotional perception needed to embody the musical “pathos” characteristic of human performance, limiting their capacity to convey genuine human emotions.

The Quarter-Tone Dilemma

Quarter-tone notes represent one of the most complex technical challenges, requiring extreme precision in vocal generation. Current models often fail to achieve smooth transitions between these microtonal notes, resulting in performances lacking the natural fluidity inherent to human singers.

Algorithm Performance Comparison

Sono’s Vocal Advantages

The Sono model excels in maintaining vocal quality during extended passages, demonstrating remarkable stability in vocal tone. This advantage gives it superiority in performing extended vocal improvisations (mawawil) and songs with prolonged musical phrases, despite limitations in conveying emotional precision.

Leria 3’s Technical Accuracy

Leria 3 shows notable progress in technical accuracy, particularly in Arabic letter pronunciation and phonetic articulation. Its strength lies in detailed vocal simulation, yet it remains incapable of replicating the natural “authority” (sultana) that distinguishes professional instrumentalists and singers.

Technological Limits and Development Horizons

Digital Programming Constraints

Despite tremendous advancements, algorithms remain confined by programming limitations that cannot accommodate improvisational elements in Eastern art. Arabic musical performance relies on spontaneous interaction difficult to translate into algorithmic equations, creating a gap between human performance and mechanical simulation.

Future Technical Development

Current developments indicate potential for significant progress in coming years, with developers focusing on improving emotional aspects of vocal performance. Current emphasis centers on integrating cultural context elements into algorithmic training, potentially bringing mechanical simulation closer to authentic artistic expression.