PRESENTED BY,
V.V.NAGA VARUN
J.SHRAVAN
K.B.GOWTHAM

---

[attachment=9576]
INTRODUCTION
 Speech coding is defined as a digital representation of the speech signal .
 Speech compression is achieved by representing each sample of digitized data by lesser number of bits.
 In compression technique the 13 to 16 bits data is represented by an equivalent of 4 or 8 bits.
 In decompression technique the compressed data is converted back to original data.
BLOCK DIAGRAM
HARDWARE
 Efficient hardware is developed by this project to implement the powerful and versatile ADPCM speech compression and decompression algorithm.
 The hardware has been developed using powerful DSP processor ADSP-2105.
 The speech processing system consists of ADSP-2105,CODEC,EPROM,RAM,Amplifier sections,mike and speakers.
FUNCTIONING
 The speech segment to be compressed is fed as input through MICRO PHONE.
 The analog speech signal is digitized by CODEC, compressed by ADSP-2105 and stored in RAM.
 The decompression process can be started by interrupting the processor.
 When the processor is interrupted, the information in memory is fetched by the processor and decompressed and is then fed to CODEC.
 The CODEC converts the digital signal to analog signal. The analog signal is amplified and output through the speaker
FEATURES
The speech compression and decompression technique is used in:
 Cellular Phones.
 Voice Mail Transmission systems.
 Speech recognition systems.

[attachment=11790]
SPEECH COMPRESSION AND DECOMPRESSION USING ADPCM TECHNIQUE
BASICS
Speech coding may be defined as a digital representation of the speech signal that provides efficient storage, transmission, recovery and faithful reconstruction of the original speech.
The speech compression is achieved by representing each sample of digitized data by lesser number of bits. For maintaining the voice quality the speech sample has to be represented by 13 to 16 bits.
In compression techniques the 13 to 16 bits data is represented by an equivalent 4 or 8 bit data. In decompression techniques the compressed data is converted back to original data. One of the important requirement in these techniques is that the voice quality has to be maintained.
HARDWARE
This project is proposed to develop an efficient hardware to implement the powerful and versatile ADPCM speech compression and decompression algorithm.
The hardware required for speech compression and decompression have been developed using powerful DSP processor ADSP-2105. The speech processing system consists of ADSP - 2105, CODEC, EPROM, RAM, Amplifier sections, mike and speakers.
FUNCTIONING
The system is ready to accept speech signal once it is reset. The speech segment to be compressed is fed as input through microphone. The analog speech signal is digitized by CODEC, compressed by ADSP-2105 and stored in RAM.
The decompression process can be started by interrupting the processor.
When the processor is interrupted, the information in memory is fetched by the processor and decompressed and is then fed to CODEC. The CODEC converts the digital signal to analog signal. The analog signal is amplified and output through the speaker.
FEATURES
The speech compression and decompression techniques are implemented in cellular phones, voice mail transmission systems and speech recognition system.
What is ADPCM
Adaptive delta pulse code modulation. An audio compression algorithm for digital audio based on describing level differences between adjacent samples
Adaptive Differential Pulse Code Modulation is an audio encoding compression technique that encodes the difference between the predicted value of the signal instead of the absolute value of the original waveform so that the compression efficiency is improved.
This difference is usually small and can thus be encoded in fewer bits than the sample itself. ADPCM is used in PC sound cards, on CD-i and CD-ROM XA discs
CODEC
Codec is an abbreviation for Coder-Decoder. It's an analog-to-digital (A/D) and digital-to-analog (D/A) converter for translating the signals from the outside world to digital, and back again
ADSP
Adaptive Digital Signal Processing; exclusive to SGC, a DSP technology which processes the signal to eliminate unwanted noise and improve incoming signal
DSP
DSP stands for Digital Signal Processor. DSP chips are widely used in sound cards, fax machines, modems, cellular phones, high-capacity hard disks and digital TVs
DSP chips are used in sound cards for recording and playback and speech synthesis. Other audio uses are amplifiers that simulate concert halls and surround-sound effects for music and home theater
Digital Signal Processing uses mathmatics to operate on a digital signal (such as a digital audio stream) to generate some type of altered output. DSP is used heavily in software and hardware effects processing.
DSP chips are found on an increasing number of sound cards to provide extra audio processing power and help relieve the computers CPU of this type of work, much like a 3D graphics accelerator would for rendering 3D graphics