Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

Voice prediction and compression using digital signal proc

VIEWS: 16 PAGES: 10

Voice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compression using digital signal procVoice prediction and compres

More Info
									  VOICE PREDICTION AND COMPRESSION
    USING DIGITAL SIGNAL PROCESSING

ABSTRACT :
                 In the fast growing communication world digital processing plays a vital
role. This paper deals with compression of the voice signal. The voice signal is amplified
by preamplifier .And fed to CODEC. The CODEC has an ADC (analog to digital
converter) which receives. The signal from the preamplifier and converts it into the
digital data .the converted digital data is encoded in serial format and transmitted to DSP
processor CODEC .The processor has two memory blocks externally, Data memory and
Boot memory. The voice signal is compressed in the ratio of 1:20 and stored in the data
memory. The required way to process the signals is stored in the program memory.
The program for the digital signal processor has been developed to receive the data from
CODEC (code decoder) and to compress it using linear predictive coding (LPC).the
compressed speech data and decodes it .the coded data is fed to DAC(built in ODEC)and
DAC reproduces exact analog signal and the power amplifier amplifies the output. the
main application of this paper is secured transmission of voice records

LIST OF CONTENTS :
1. INTRODUCTION
2. VOICE COMPRESSION
        2.1 SAMPLING
        2.2 QUANTIZATION
        2.3 CODING
3. VOICE PREDICTION
        3.1 LINEAR PREDICTIVE CODING
4. HARDWARE DESCRIPTION
        4.1 BLOCK DIAGRAM
        4.2 EXTERNAL MEMORY
        4.3 ADSP-2105
         4.4 PREAMPLIFIER
         4.5 CODEC
         4.6 POWER AMPLIFIER
5. VOICE DESCRIPTION
6. FUTURE DEVELOPMENTS


            6.1 SECURED TRANSMISSION OF VOICE RECORDS
            6.2 BLOCK DIAGRAM
7. CONCLUSION
8. BIBILIOGRAPHY


INTRODUCTION:
                     The digital signal processors are widely used for analog signal
processing such as compression of a voice signal, filtering and image processing .one of
the main features of such high –end digital signal processor is their ability to execute
every instruction in single clock cycle due to parallelism introduced in processing.
                    In this paper, we are going to see about the application of voice
prediction and compression using digital signal processing .the voice signal is amplified
by preamplifier and fed to CODEC .the CODEC has an ADC (analog to digital
converter) which receives the signal from preamplifier and converts it into digital data
.the converted digital data is encoded in serial format and transmitted to DSP processor
by CODEC.
                         The processor has two memory blocks externally, Data memory and
Boot memory. The voice signal is compressed in the ratio of 1:20 and stored in the data
memory. The required way to process the signals is stored in the program memory.
                         The program for the digital signal processor has been developed to
receive the data from CODEC (code decoder) and to compress it using linear predictive
coding (LPC).the compressed speech data and decodes it .the coded data is fed to
DAC(built in CODEC)and DAC reproduces exact analog signal and the power amplifier
amplifies the output .
 VOICE COMPRESSION:
           The input voice signal is picked by the microphone, which converts original
voice signal into electrical signal .the preamplifier amplifies the converted electrical
signal. Input power level of 0-5v is maintained at constant level by power supply circuit.
The power supply circuit has voltage regulator to maintain the dc voltage level at
constant rate.
             The amplified analog signal is fed up through CODEC, which process analog
voice signal in to digital signal .at first, it converts them into a digital form that is a
sequence numbers having finite decision .conceptually, and we view A/D conversion as a
three step process.
Sampling
             This is the conversion of continuous time signal into discrete time signal
obtained by taking samples of continuous time signal at discrete time instants.thus,if
Xa(t) is input to the sampler ,the output is Xa(nt)=X(n),where T is the sampling interval .
QUANTIZATION :
                 This is the conversion of discrete time continuous valued signal .the value
of each signal sample is represented by a value selected form a finite set of possible
values. Then difference between quantized sample Ex(n) and the quantized output Xq(n)
is called quantization error.
CODING :
                 In this step the discrete signal is coded. That is, it is converted into machine
language .in CODEC both A/D and D/A converters are used

VOICE PREDICTION :
                 The voice is predicted by linear predictive coding .the speech signal is
estimated from the past output as a linear function of the digital quantizing systems ,
which is called linear predictive coding (LPC).calculate the first three format center
frequencies and bandwidths of each member’s speech sample the algorithm divides each
member’s represented speech .samples into a number of 15ms frames ,and then
calculated the coefficient of each speech frame that where determined by taking the root
of each polynomial or all-pole model of the vocal track derived from the coefficients .
                         LPC processor



                    N        M               W(n)              P




                                 X(n)
Pre-              Frame                 windowing         Auto
emphasis          blocking                                correlation
                                                          analysis




Temporal          Parameter             LPC                 LPC
derivative        Weighting             parameter           analysis
                                        conversion


C^m(t)




              Finally the format frequencies and bandwidth calculated from all the
frames where time average to obtain a single format version for each class member .the
LPC approach is employed to estimate the format frequencies and bandwidth on the basis
of its proven utility in speech and its computational efficiency. the processor processes
the coded digital signal and this processes is know as compression .the compressed
digital signals are stored in the external data memory of 8KB capacity .here we have to
choose the ratio 20:1 to compressed the signals .that is 20 seconds speech is compressed
into one second data. The compressed data is stored in the data memory .the compressed
signal is fed through the power amplifier.
HARDWARE DESCRIPTION:
              The block diagram of dsp based voice prediction and compression is
shown in figure.


EXTERNAL MEMORY:
              There are two types of memory .they are: data memory and boot memory.
each processor contains on-chip RAM or ROM, so that the portion of the program
memory and portion of data memory are spaced in on chip. The boot memory space and
data memory space can be used to load on-chip program from external EPROM during
the reset. Each memory has a capacity of 8KB.




                                                                  PRE                SECRET
                                                                                     SREECH
                                                                  AMP                RECORD
 DATA


MEMORY                                ADSP
                                      2105

 BOOT
MEMORY                                                           CODEC




                                                                 PWR
                                                                 AMP
DATA MEMORY:
The processor has 14bit address bus and data memory address bus (DMA), which is,
multiplexed off-chip. The data memory pin indicates that the address bus is being driven
with data memory address and memory can be selected. The data memory contains data
in digital form.


BOOT MEMORY:
               The entire program memory or any portion of it can be loaded from an
external source using a boot sequence. To interface with inexpensive EPROM, the
processor loads one byte instruction at a time .the boot memory initialize the ADSP-2105
after reset.




ADSP-2105 PROCESSOR:
   The heart of the paper is ADSP -2105.it is a 16 bit fixed point DSP microprocessor
with on –chip memory. It has enhanced hardware architecture for three buses most of the
instructions set are single cycle into multifunction. It has on-chip program memory ROM
or RAM and data memory RAM.
          The Main feature of ADSP – 2105 are


       25 MIPS, 40ns maximum instruction rate.
       separate on-chip buses for program and data memory
       program memory stores both instruction sets and data
       Dual data addressing generators with modulo and bit-reverse addressing.
       Efficient program sequencing with zero overhead.
       Automatic booting of on-chip program memory from external memory EPROM.
       Host interface provides easy interface to 68000,80C51,ADSP-21XX etc.,
The ADSP -2105 processor is a single chip microcomputer optimized for DSP and other
high speed numeric data processing application. The ADSP -2105 processors are all built
on a common core .each processor combines the core DSP architecture computation
units , data address generators ,program sequencer with different features such as on-
chip program and data memory RAM , a programmable timer , 1or2 serial ports and on
the ADSP – 2105,a host interface port
CODEC:
               CODEC: coder and decoder. CODEC converts the received voice signal
into digital signal and it is given to the processor. The processor controls the CODEC
operation. While retriving, it converts the digital signal into the analog signal. It has built
–in ADC and DAC.
PREAMPLIFIER:
      The voice signal from the microphone is given to the pre-amplifier. It is used to
amplify the voice signal that has high input impedance, low output admittance, large
bandwidth and high voltage gain.

AUDIO POWER AMPLIFIER:
      The audio power amplifier receives the data from CODEC. The gain of audio
amplifier is high and is given to speaker. The output impedance of audio amplifier and
speaker should match, in order to provide no noise.
NAND GATE :
      The NAND gate function is the complement of AND gate. The NAND gate is used
to reset the circuit. It provides quadruple 2-inputs. If the two inputs are high, the
processor will reset.
POWER SUPPLY :
        It consisting of a step-down transformer to produce an output of 9volts (P-P),
500ma.then it is given to a full wave rectifier and regulator to produce constant DC
power supply

DECOMPRESSION OF THE SIGNAL:
            To decompress the output signal, the compressed signal it is fed through the
CODEC. it converts the compressed digital signal into the decompressed digital signal.
The CODEC converts the digital signal into the analog signal that is amplified by the
power amplifier .the power supply unit activates the power amplifier .the output of the
power amplifier is shown on the speaker. The original signal is received at the speaker.
     FUTURE DEVELOPMENTS:
                                                                           PRE                  SECRET
                                                                           AMP                  SREECH
                                                                                                RECORD
      DATA


     MEMORY                                   ADSP
                                              2105

      BOOT
     MEMORY                                                               CODEC




                                                                           PWR                   SPK
                                                                           AMP




                                           TELEPHONE                                        DECOMPRESSION
DECODER            PWR AMP                                         TELEPHONE
                                           TRANSMISSION                                         KIT
                                                                  RECEIVER
                                           LINE




     Voice prediction and compression will activate various future developments in different
     electronic environment .the various application of voice prediction and compression are
     telenetworking , security, net accessing etc.,let us see about the model of security
     transmission of voice records in this paper.
SECURED TRANSMISSION OF VOICE RECORDS:
            In future, compression of voice signal provides high security to the
transmission of secret speech records from one place to another place. Consider an
example, in defense arm stations we need to transmit orders, control data from one station
to another. Also the original secret voice records are transmitted from the control room to
various stations. At that time the transmitting data should be kept secret. The only way to
keep our data secret we should encode the data in the transmitting end and decode it in
the receiving end. In this case there is a possibility of data being decoded by unauthorized
person they should not retrieved the original signal. Hence in order to avoid this we
should compress the original signal and then encode it. So that no unauthorized person
can decode it .This gives reliable security transmission of the voice records. This is the
main advantage and also greatest application of voice prediction and compression using
digital signal processing.


CONCLUTION:
Thus we conclude that the compression of voice signal plays a vital roll in the
telenetworking, security and net accessing by the voice signal section this application is
widely applicable. Using the digital signal processing high speed numeric processing can
be achieved. In day today life secured transmission is very much necessary not only to
military people but to every individual person. More security is provided to the
transmission of voice records by using digital signal processing in voice prediction and
compression. From this it is possible to obtain secured transmission of data.

BIBILIOGRAPHY:
1. ADSP2100 family user manual third edition, analog devices.
2. DSP by Allan. V.Oppenheim&Ronald W.Schafer, PHI.
3. DSP-Principle, algorithm&application by John.G.Proakis.
4. Introduction to DSP by Johny R.Johnson.

								
To top