Language Identification Using Excitation Source Features by K. Sreenivasa Rao, Dipanjan Nandi

By K. Sreenivasa Rao, Dipanjan Nandi

This publication discusses the contribution of excitation resource details in discriminating language. The authors concentrate on the excitation resource component to speech for enhancement of language id (LID) functionality. Language particular positive aspects are extracted utilizing diverse modes: (i) Implicit processing of linear prediction (LP) residual and (ii) particular parameterization of linear prediction residual. The publication discusses how in implicit processing process, excitation resource positive factors are derived from LP residual, Hilbert envelope (magnitude) of LP residual and section of LP residual; and in particular parameterization technique, LP residual sign is processed in spectral area to extract the appropriate language particular positive aspects. The authors extra extract resource beneficial properties from those modes, that are mixed for reinforcing the functionality of LID structures. The proposed excitation resource positive factors also are investigated for LID in historical past noisy environments. each one bankruptcy of this e-book presents the incentive for exploring the explicit characteristic for LID job, and for that reason speak about the the way to extract these beneficial properties and at last recommend acceptable versions to catch the language particular wisdom from the proposed good points. eventually, the e-book talk about approximately a number of combos of spectral and resource positive factors, and the specified versions to augment the functionality of LID systems.

Show description

Read or Download Language Identification Using Excitation Source Features PDF

Similar ai & machine learning books

Artificial Intelligence Through Prolog

Synthetic Intelligence via Prolog e-book

Language, Cohesion and Form (Studies in Natural Language Processing)

As a pioneer in computational linguistics, operating within the earliest days of language processing through desktop, Margaret Masterman believed that which means, now not grammar, used to be the main to realizing languages, and that machines may possibly make sure the which means of sentences. This quantity brings jointly Masterman's groundbreaking papers for the 1st time, demonstrating the significance of her paintings within the philosophy of technology and the character of iconic languages.

Handbook of Natural Language Processing

This examine explores the layout and alertness of traditional language text-based processing platforms, in line with generative linguistics, empirical copus research, and synthetic neural networks. It emphasizes the sensible instruments to deal with the chosen method

Extra resources for Language Identification Using Excitation Source Features

Example text

Rao, Voice conversion by mapping the speaker-specific features using pitch synchronous approach. Comput. Speech Lang. 24(3), 474–494 (2010) 61. R. Hussain Laskar, K. Banerjee, F. Ahmed Talukdar, K. Sreenivasa Rao, A pitch synchronous approach to design voice conversion system using source-filter correlation. Int. J. Speech Technol. (Springer) 15(3), 419–431 (2012) Chapter 3 Implicit Excitation Source Features for Language Identification Abstract This chapter discusses about the proposed approaches to model the implicit features of excitation source information for language identification.

F. L. Gauvain, Cross lingual experiments with phone recognition. in International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp. 507–510, April 1993 6. F. L. Gauvain, Language identification using phonebased acoustic likelihoods, in International Conference on Acoustics, Speech, and Signal Processing (ICASSP), vol. 1, , pp. I/293–I/296, April 1994 7. Y. Muthusamy, R. Cole, M. Gopalakrishnan, A segment-based approach to automatic language identification, in International Conference on Acoustics, Speech, and Signal Processing (ICASSP), vol.

3 is developed by the above mentioned combination. • However, the combination of HE and RP at each level (HE + RP) represents partial information about excitation source. Hence, to acquire the complete language-specific excitation source information we have combined the evidences of (HE+RP) features from sub, seg and supra levels. The LID system-III at phase-III of Fig. 3 indicates the corresponding combination. 5 Performance Evaluation of LID Systems Developed Using Implicit Excitation Source Features In this work, we have carried out the evaluation of the language models using leavetwo-speaker-out approach.

Download PDF sample

Rated 4.49 of 5 – based on 39 votes