Speech Denoising and Compression Methods in Mobile Satellite Communication

notownbuffAI and Robotics

Nov 17, 2013 (3 years and 9 months ago)

51 views

Speech Denoising and Compression Methods in Mobile Satellite Communication

ABSTRACT


In this Dissertation of mine, my aim is to provide a clear understanding of how a speech
is de Noised for a mobile device in order for enhancing an effective

way of communicating data
over long distances without any loss of information and clarity of the speech signal. I shall
describe in this paper the technique of de noising a Speech signal produced by the Humming
sound of the GSM mobile phones using the met
hod of Notch filtering and adaptive noise
cancellation using the Wiener filter method.


The motivation for me to undertake this topic for my Research and Development module
is to directly work upon enhancing the Signal to Noise ratio and to decre
ase the Noise level in
the transmitted signals to the Mobile devices and the develop a speech compression technique
which will compress large information data into smaller packets of intelligent signal so as to
allow the user to speak uninterrupted without

the voice been broken at the receiver. My
foundation will be to De noise the speech signals, to improve the Code, Transmission, improve
the Quality of Sound, to decrease the noise presence in the background and then to compress
the resulting data into sma
ller packets of signals which can be then across the receiver for the
user to communicate the information


Speech transmitting devices such as Cell phones are most often face the problem of noisy
environments and Speech signals in the mobile phone
s are mostly subjected to reduced quality
of strength because of the Noise in the surroundings. The underlying principal is that to find a
way to reduce the Noise interference levels in the signals by means of De noising the noisy data
and to allocate more

space for the speech data to be compressed so that more number of users
can talk on the mobile phones clearly without any problem of hearing what the other person
says by Speech compression technique.


In this Research topic, I shall pr
ovide a justification of my work on generating a de
noised signal by using a Notch Filter which involves sampling of signal to compress the sound
waves and cancel the noise by using the Speech cancellation frame and then deploying back to
the mobile receiv
er which will be provided with the Active Noise control to eliminate any
leftover noise which is there in the receiver
.


In this project I shall emphasize upon a fast way of de noising a robust speech signal by
using a linear time invariant No
tch filter model of speech recognition. In this speech model, we
implement the concept of Speech Frame Cancellation to convert a noise signal into samples
providing less Mean square error when compared to an original clean signal data and de noising
the s
peech signal producing low error rates. [1] And then later on simulating the speech signal
with the help of MATLAB
-
6 by using a single channel adaptive filter and testing the signal using
an Artificial GSM mobile.