AUTHOR=Yang Ye , Zeng Fan-Gang 

TITLE=Syllable-rate-adjusted-modulation (SRAM) predicts clear and conversational speech intelligibility

JOURNAL=Frontiers in Human Neuroscience

VOLUME=Volume 18 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2024.1324027

DOI=10.3389/fnhum.2024.1324027

ISSN=1662-5161

ABSTRACT=Objectively predicting speech intelligibility is important in both telecommunication and humanmachine interaction systems. The classic method relies on signal-to-noise ratios (SNR) to successfully predict speech intelligibility. One exception is clear speech, in which a talker intentionally articulates as if speaking to someone who has hearing loss or is from a different language background. As a result, at the same SNR, clear speech produces higher intelligibility than conversational speech. Despite numerous efforts, no objective metric can predict the clear speech benefit at the sentence level. Here we proposed a Syllable-Rate-Adjusted-Modulation (SRAM) index to predict the intelligibility of clear and conversational speech. The SRAM used as short as 1s speech and estimated its modulation power above the syllable rate. We found that not only did the SRAM outperform a similar index calculating the modulation power over the entire frequency range, but also three reference metrics (envelope-regression-based speech transmission index, hearing-aid speech perception index version 2 and short-time objective intelligibility) and five automatic speech recognition systems (Amazon Transcribe, Microsoft Azure Speech-To-Text, Google Speech-To-Text, wav2vec2 and Whisper). The SRAM can potentially help understand the characteristics of clear speech, screen speech materials with high intelligibility, and convert conversational speech into clear speech.Recent advances in automatic speech recognition (ASR) have made it a viable alternative for predicting speech intelligibility (Feng and Chen, 2022;Karbasi and Kolossa, 2022). We are not aware of the application of ASR to predict clear speech benefit. Here we compared the SRAM performance against five popular modern ASR systems.