Development and Validation of Persian Speech Test Signal (PerSTS)

Area of Science:

• Audiology
• Speech processing
• Linguistics

Background:

• Digital hearing aids require precise fitting for effective speech processing.
• The International Speech Test Signal (ISTS) lacks Persian phonetic and prosodic features.
• A standardized Persian speech test signal is needed for accurate hearing aid verification.

Purpose of the Study:

• To create and acoustically verify a Persian version of the ISTS, named the Persian Speech Test Signal (PerSTS).
• To enhance the accuracy of digital hearing aid verification for Persian-speaking users.

Main Methods:

• Translated the "North Wind and the Sun" passage into Persian, ensuring natural syntax and phoneme distribution.
• Recorded the passage by six native female speakers at 44.1 kHz/24-bit resolution.
• Assembled a 60-second PerSTS by segmenting, enhancing, and stitching recordings to match the Byrne female long-term average speech spectrum (LTASS) and expert-rated naturalness.

Main Results:

• The PerSTS shows spectral peaks characteristic of female Persian speech (200-400 Hz and 3-4 kHz).
• Its long-term average speech spectrum (LTASS) closely aligns with the Byrne LTASS and ISTS.
• Acoustic analysis confirmed natural speech characteristics, with a crest factor of 16.7 dB and dynamic range of 27.1 dB; experts rated naturalness as 4.35/5.

Conclusions:

• The PerSTS accurately mimics acoustic and temporal features of natural Persian speech.
• It meets International Electrotechnical Commission (IEC) standards for audiology tools.
• Provides a standardized, effective tool for hearing aid verification in Persian-speaking populations.