Effects of rain noise from metal roofs on speech identification

Background. The aim of this study was to investigate the individual effects of envelope enhancement and high-pass filtering (500 Hz) on word identification scores in quiet for individuals with Auditory Neuropathy. Method. Twelve individuals with Auditory Neuropathy (six males and six females) with ages ranging from 12 to 40 years participated in the study. Word identification was assessed using bi-syllabic words in each of three speech processing conditions: unprocessed, envelope-enhanced, and high-pass filtered. All signal processing was carried out using MATLAB-7. Results. Word identification scores showed a mean improvement of 18% with envelope enhanced versus unprocessed speech. No significant improvement was observed with high-pass filtered versus unprocessed speech. Conclusion. These results suggest that the compression/expansion signal processing strategy enhances speech identification scores—at least for mild and moderately impaired individuals with AN. In contrast, simple high-pass filtering (i.e., eliminating the low-frequency content of the signal) does not improve speech perception in quiet for individuals with Auditory Neuropathy.

Download Full-text

Perceptual compensation for effects of reverberation in speech identification: a computer model based on auditory efferent processing

10.21437/interspeech.2010-662 ◽

2010 ◽

Author(s):

Amy V. Beeston ◽

Guy J. Brown

Keyword(s):

Computer Model ◽

Model Based ◽

Auditory Efferent ◽

Speech Identification

Download Full-text

Speech identification using a sequence-hased heuristic

47th International Symposium ELMAR, 2005. ◽

10.1109/elmar.2005.193683 ◽

2005 ◽

Author(s):

G. Heinrich

Keyword(s):

Speech Identification

Download Full-text

Modern Speech Identification Model using Acoustic Neural approach

International Journal of New Practices in Management and Engineering ◽

10.17762/ijnpme.v6i03.58 ◽

2017 ◽

Vol 6 (03) ◽

pp. 01-06

Author(s):

Mr. Ashish Uplenchwar

Keyword(s):

Speech Processing ◽

Vital Role ◽

Reservoir Computing ◽

Continuous Speech Recognition ◽

Efficient Communication ◽

Machine Interface ◽

Recognition Systems ◽

Speech Identification ◽

Identification Model ◽

Sophisticated Technology

Modern day technology demands sophisticated technology to give input commands to computational devices. Prominent techniques has to be introduced to make human machine interface smooth and compatible especially speech signals. Establishing efficient communication between computer and machine plays a vital role in speech processing. This article uses one of current technologies in the Continuous Speech Recognition systems which is Reservoir Computing based Neural Network followed by likelihood conversion. Our aim is to build a stand-alone system which understands the terminology of languages. Throughout the development, measures will be taken to keep the memory requirement and the processing time of the software as small as possible.

Download Full-text