Experimental Speech Science Laboratory Curriculum

Theoretical Foundations of Acoustic Speech Analysis

This reference manual outlines how digital signal processing algorithms analyze raw voice waves to isolate speech characteristics like resonance, intensity, and clarity.

1. Source-Filter Speech Production Model

In speech science, production is modeled using the foundational Source-Filter Theory. The sound source is created by the larynx as periodic air pulses, or by vocal tract constriction as turbulent noise.

The vocal tract acts as a dynamic acoustic filter, shaping the sound source. Resonating frequencies are amplified or dampened based on the physical position of features like the tongue and lips, producing distinct, identifiable speech sounds.

2. Core Acoustic Metrics & Visualization Systems

The diagnostic console analyzes voice inputs across several core parameters to trace speech traits:

Fundamental Frequency Contour (F0)

Measures vocal fold vibration speed in Hertz (Hz). This pitch track captures prosody patterns, emotional stress indicators, and language-specific tone variations.

Vocal Resonances (Formants F1/F2)

Identifies vocal tract resonance frequencies. F1 aligns with jaw opening height, while F2 maps front-to-back tongue placement, defining vowel identity.

3. Evaluating Sound Clarity: Harmonic-to-Noise Ratio

To analyze voice clarity and breathiness mathematically, the engine evaluates the balance between periodic and non-periodic energy wave attributes using the Harmonic-to-Noise Ratio (HNR):

HNR Equation Output Metric Value Scale Profile: dB = 10 × log10( Energy_Harmonics / Energy_Noise )

High HNR values indicate clear, stable vocal tones, where harmonic components dominate. Drop-offs into lower numbers signal increased airflow turbulence or friction noise, characteristic of hoarse, whispered, or unvoiced speech traits.

4. Applications in Clinical and Pedagogical Fields

L2 Pronunciation Diagnostics

Educators use formant graph visualizations to help non-native students adjust tongue placement and refine target vowel pronunciation accuracy.

Clinical Acoustic Speech Pathology

Pathologists analyze jitter, shimmer, and HNR metrics to monitor vocal cord health and track recovery progress across speech therapy sessions.

Ready to compute digital speech signal contours?

Inject sample audio recordings into our core DSP engines to isolate voice tracks interactively.

Launch Signal Evaluation Console Core