4. Analysis and quality conversion of nonacoustic signals: the physiological microphone (PMIC)
-
Seyed Omid Sadjadi
Abstract
There are a number of scenarios where effective human-to-human speech communication is vital, yet either limited speech production capabilities due to pathology or in noisy environmental conditions limit intelligible information exchange and reduce overall quality. Traditionally, front-end speech enhancement techniques have been employed to alleviate the effect of environmental noise. Speech-processing normalization of speech under pathology has also been employed to increase quality of pathological speech. A recent alternative approach to deal with these scenarios is the use of nonacoustic sensors, which are essentially independent of the acoustic sound propagation characteristics of the environments where human communication is needed. The physiological microphone (PMIC), as a nonacoustic contact sensor, has been shown to be quite useful for speech systems under adverse noisy conditions. It also could provide an alternative signal capture mode for speech under vocal fold pathology. However, due to alternative pickup location and the nonacoustic principle of operation, captured signals appear muffled and metallic to the listener with variations to the speakerdependent structure. To facilitate more robust and natural human-to-human speech communication, in this chapter we present a probabilistic transformation approach to improve the perceptual quality and intelligibility of PMIC speech by mapping the nonacoustic signal into the conventional close-talk acoustic microphone speech production space, as well as by minimizing distortions arising from alternative pickup location. Performance of the proposed approach is objectively evaluated based on five distinct measures. Moreover, for subjective performance assessment, a listening experiment is designed and conducted. Obtained results confirm that incorporating the probabilistic transformation yields significant improvement in overall PMIC speech quality and intelligibility. This solution offers an alternative to individuals with severe vocal fold pathology or areas where traditional speech production is not an option.
Abstract
There are a number of scenarios where effective human-to-human speech communication is vital, yet either limited speech production capabilities due to pathology or in noisy environmental conditions limit intelligible information exchange and reduce overall quality. Traditionally, front-end speech enhancement techniques have been employed to alleviate the effect of environmental noise. Speech-processing normalization of speech under pathology has also been employed to increase quality of pathological speech. A recent alternative approach to deal with these scenarios is the use of nonacoustic sensors, which are essentially independent of the acoustic sound propagation characteristics of the environments where human communication is needed. The physiological microphone (PMIC), as a nonacoustic contact sensor, has been shown to be quite useful for speech systems under adverse noisy conditions. It also could provide an alternative signal capture mode for speech under vocal fold pathology. However, due to alternative pickup location and the nonacoustic principle of operation, captured signals appear muffled and metallic to the listener with variations to the speakerdependent structure. To facilitate more robust and natural human-to-human speech communication, in this chapter we present a probabilistic transformation approach to improve the perceptual quality and intelligibility of PMIC speech by mapping the nonacoustic signal into the conventional close-talk acoustic microphone speech production space, as well as by minimizing distortions arising from alternative pickup location. Performance of the proposed approach is objectively evaluated based on five distinct measures. Moreover, for subjective performance assessment, a listening experiment is designed and conducted. Obtained results confirm that incorporating the probabilistic transformation yields significant improvement in overall PMIC speech quality and intelligibility. This solution offers an alternative to individuals with severe vocal fold pathology or areas where traditional speech production is not an option.
Chapters in this book
- Frontmatter I
- Foreword V
- Acknowledgments IX
- Contents XI
- List of contributors XIII
- Introduction 1
-
Part I: Comparative analysis of methods for speaker identification, speech recognition, and intelligibility modification in the dysarthric speaker population
- 1. State-of-the-art speaker recognition methods applied to speakers with dysarthria 7
- 2. Enhancement of continuous dysarthric speech 35
- 3. Assessment and intelligibility modification for dysarthric speech 67
-
Part II: New approaches to speech reconstruction and enhancement via conversion of non-acoustic signals
- 4. Analysis and quality conversion of nonacoustic signals: the physiological microphone (PMIC) 97
- 5. Non-audible murmur to audible speech conversion 125
-
Part III: Use of novel speech diagnostic and therapeutic intervention software for speech enhancement and rehabilitation
- 6. Application of speech signal processing for assessment and treatment of voice and speech disorders 153
- 7. A mobile phone-based platform for asynchronous speech therapy 195
Chapters in this book
- Frontmatter I
- Foreword V
- Acknowledgments IX
- Contents XI
- List of contributors XIII
- Introduction 1
-
Part I: Comparative analysis of methods for speaker identification, speech recognition, and intelligibility modification in the dysarthric speaker population
- 1. State-of-the-art speaker recognition methods applied to speakers with dysarthria 7
- 2. Enhancement of continuous dysarthric speech 35
- 3. Assessment and intelligibility modification for dysarthric speech 67
-
Part II: New approaches to speech reconstruction and enhancement via conversion of non-acoustic signals
- 4. Analysis and quality conversion of nonacoustic signals: the physiological microphone (PMIC) 97
- 5. Non-audible murmur to audible speech conversion 125
-
Part III: Use of novel speech diagnostic and therapeutic intervention software for speech enhancement and rehabilitation
- 6. Application of speech signal processing for assessment and treatment of voice and speech disorders 153
- 7. A mobile phone-based platform for asynchronous speech therapy 195
