Implementation of a 1200 bps low bit rate speech coding algorithm
2001
0 views
0 downloads
Advisor: Doç. Dr. Buyurman Baykal
Abstract (EN)
IMPLEMENTATION OF A 1200 bps LOW BIT RATE SPEECH CODING ALGORITHM (Phd. Thesis) Selma ÖZAYDIN GAZİ UNIVERSITY INSTITUTE OF SCIENCE AND TECHNOLOGY ABSTRACT Speech coding at bit rates below 4 kbps is widely used in applications such as digital wireless and wireline communications, satellite communications and digital voice storage systems. Such services require the use of high quality low bit rate coders to efficiently code the speech signal before transmission or storage. The majority of such coders employ algorithms that are based on linear predictive coding (LPC). In this thesis, we present a new speech compression technique designed for good quality speech coding at 1.2 kpbs based on the LPC algorithm. A 1.2 kbps matrix based coder (MBC) has been simulated using the C programming language. We have verified our performance results using subjective listening tests, in which 2.4 kbps MELP (mixed excitation linear prediction) coder and 2.4 kbps LPC coder has been used for comparison. 1.2 kbps MBC coder has produced good quality coded speech with a high degree of intelligibility and naturalness when compared to the coding schemes above operating at 2.4 kbps. We explore a method to decrease the bit rate of LSF (line spectral frequency) parameters coding. We found a LSF residual vector using the correlation between adjacent frame LSF vectors. A new coding model based on the residual multi stage matrix quantization (R-MSMQ) method is used to decrease the bit rate of quantization. LSF parameters of two adjacent frames are grouped and jointly quantized in order to obtain a high quality efficiency The effectiveness and limitations of this method are investigated by calculating the spectral distortion and outliers for each designed codebook. Results indicate that codebook spectral distortion can be decreased as low as 1 dB at 1 8 bits/frame with R- MSMQ technique. With using the linear discriminant voicing classification method and pitch period coding table, only 1 bit is used to code voicing information. Science code : 0609 Key Words : Digital speech coding, low bit rate, vector quantization, matrix quantization, residual signal Page number : 200 Adviser : Assoc. Prof. Buyurman BAYKAL
Author
Dr. Selma Özaydın
Institution
How to Cite
Selma Özaydın (Doctorate thesis). Implementation of a 1200 bps low bit rate speech coding algorithm, 2001, Gazi University.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Gazi University
- Occupational accident analysis and modelling in oil and gas drilling sector Turkey(2021)
- Experimental development of the interfacial bond-slip model between textile reinforced mortar strips and masonry walls(2025)
- Deveplopment of semiconductor humidity sensors(2021)
- The effect of computer-assisted and direct strategy teaching on reading comprehension(2021)
- Investigation of engineer applicant students' conceptual and procedural knowledge of integral problem solving process from the point of bloom taxonomy(2021)
- Boric acid doped into sulfonated polystyrene used as polymer electrolytes in a proton exchange membrane fuel cell(2006)
