Benchmark for Speaker Identification using Mel Frequency Cepstral Coefficients on Vowels Following the Nasal Continuants in Kannada (Record no. 17949)
[ view plain ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 02558caa a2200217 4500 |
| 022 ## - INTERNATIONAL STANDARD SERIAL NUMBER | |
| International Standard Serial Number | 2582-4961 |
| 100 ## - MAIN ENTRY--PERSONAL NAME | |
| Personal name | Suman Suresh |
| 245 #0 - TITLE STATEMENT | |
| Title | Benchmark for Speaker Identification using Mel Frequency Cepstral Coefficients on Vowels Following the Nasal Continuants in Kannada |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Date of publication, distribution, etc. | 2015 |
| 270 ## - ADDRESS | |
| Address | Lecturer in Speech Sciences, Department of Speech-Language, Sciences, All India Institute of Speech and Hearing, Mysore-06 |
| 270 ## - ADDRESS | |
| Electronic mail address | hemachari2@yahoo.com |
| 300 ## - PHYSICAL DESCRIPTION | |
| Extent | 63-75 |
| 490 ## - SERIES STATEMENT | |
| Volume/sequential designation | Volume 34 |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | Aim was to obtain the benchmark for speaker identication using Mel Frequency Cepstral Coecients (MFCC) on vowels following the nasal continuants in Kannada language. articipants chosen were twenty Kannada speaking male neuro-typical adults, in the age range of 20-30 years. Kannada meaningful words (30) with long vowels /a:/, /i:/, /u:/ following the nasal continuants /m/ and /n/ formed the material. Speech Science Lab Work bench, a Semi-Automatic vocabulary dependent speaker recognition software was used to extract MFCC for the trun- cated (PRAAT) long vowels. Results indicated higher percent correct identication for Condition I (live verse live recording). On comparison among the three vowels following the nasal continuant /m/, /i:/ is better followed by /a:/ and /u:/. Whereas for /n/ the vowel /a:/ is better followed by /i:/ and /u:/. On an average of percentage of correct speaker identication of three vowels compared between the nasal continuant, the vowels following the nasal /n/ (90%) and /m/ (90%) was similar. Condition II (Mobile verse Mobile) and Condition III (Mobile verse Live) was comparatively poorer than Condition I, thus the benchmark was ob- tained. Discussion concludes that during the transmission of voice signals through communication channels, the signals are reproduced with errors caused by distortions from the microphone and channel, and acoustical, electromagnetic interferences and noises aect the transmitting signal. Where speech coding algorithms that are part of Global System for Mobile compress speech signal before transmission, reduce the number of bits in digital representation but at the same time, maintain acceptable quality. |
| 653 ## - INDEX TERM--UNCONTROLLED | |
| Uncontrolled term | Hypothetical, Mobile Transmission, Semi-Automatic, Electromagnetic Interference |
| 700 ## - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Hema, N |
| 773 ## - HOST ITEM ENTRY | |
| Main entry heading | Journal of All India Institute of Speech and Hearing |
| 856 ## - ELECTRONIC LOCATION AND ACCESS | |
| Uniform Resource Identifier | http://192.168.102.7/jaiish/index.php/aiish/article/view/843/531 |
| 942 ## - ADDED ENTRY ELEMENTS (KOHA) | |
| Koha item type | Faculty Publications |
| -- | FJAS |
No items available.