Cantor Digitalis: chironomic parametric synthesis of singing

Abstract

Cantor Digitalis is a performative singing synthesizer that is composed of two main parts: a chironomic control interface and a parametric voice synthesizer. The control interface is based on a pen/touch graphic tablet equipped with a template representing vocalic and melodic spaces. Hand and pen positions, pen pressure, and a graphical user interface are assigned to specific vocal controls. This interface allows for real-time accurate control over high-level singing synthesis parameters. The sound generation system is based on a parametric synthesizer that features a spectral voice source model, a vocal tract model consisting of parallel filters for vocalic formants and cascaded with anti-resonance for the spectral effect of hypo-pharynx cavities, and rules for parameter settings and source/filter dependencies between fundamental frequency, vocal effort, and formants. Because Cantor Digitalis is a parametric system, every aspect of voice quality can be controlled (e.g., vocal tract size, aperiodicities in the voice source, vowels, and so forth). It offers several presets for different voice types. Cantor Digitalis has been played on stage in several public concerts, and it has also been proven to be useful as a tool for voice pedagogy. The aim of this article is to provide a comprehensive technical overview of Cantor Digitalis.

A Spectral Glottal Flow Model for Source-filter Separation of Speech

“…A close correlation has also been found between vocal effort and spectral tilt [8,9]. These results motivate the use of the 3-pole glottal flow model for voice quality modification [10,11] and expressive singing or speech synthesis [12,13].…”

A Mouth Opening Effect Based on Pole Modification for Expressive Singing Voice Transformation

“…Figure 6 shows the results for the different α values. As one can see, the answers highly overlap and span the whole [1][2][3][4][5] range for all sounds categories, with no difference between α values -0.5, 0, and 0.5. The mean rating is however a bit lower for higher transformations levels -1 and 1.…”

“…Past studies have already permitted to gather knowledge about how various voice spectral features may change during singing, which could be exploited in synthesis systems by defining a set of rules, as proposed in [1] with formant tuning for pitch change. But if such rules are easy to implement in formant synthesizers like [2], where all source and formants' parameters can be explicitly controlled, this task is much more complicated for systems based on signal transformations like [3] and [4].…”

A unit selection text-to-speech-and-singing synthesis framework from neutral speech: proof of concept

“…First-generation rulebased systems gave way to data-driven approaches mainly due to the difficulty of generating the control parameters to get high-quality results [1]. However, formant synthesis is still used nowadays in the context of performative singing synthesis [29], where flexibility and real time are the main issues.…”