3a. Speech Coders

WLL M I CRO-CELLULAR WILL M A CRO-CELLULAR WILL 32 kb/s 13.6 kb/s DECT GSM , CDMA, D-AMPS

BIT RATES AUDIBLE RANGE TELEPHONIC PCM VSAT WLL MOBILE 64 kb/s 20 Hz TO 20000Hz 16kb/s , 14.4kb/s , 9.6kb/s , 2.4 kb/s 32 kb/s , 13.6 kb/s 13.6 kb/s 300Hz to 3400Hz

WHY TO REDUCE THE BIT RATE ? TO ACCOMMODATE MORE INDIVIDUAL SPEECH CHANNELS WITHIN THE SAME TRANSMISSION CHANNEL (MEDIA BW ) THE COST TO BE PAID ? SYSTEM CAPACITY COST REVENUE LIMITED (TRANSMISSION) CHANNEL BW

HOW COME IS IT POSSIBLE? IT IS POSSIBLE TO REDUCE THE BW WITHOUT MUCH COMPROMISE ON QUALITY

PROPERTIES OF HUMAN SPEECH & EAR REDUNDANCIES IN THE SPEECH SIGNALS INTERMITTENT NATURE OF SPEECH PERCEPTUAL LIMITS OF HUMAN EAR

SPEECH REDUNDANCIES SPEECH ENERGY IS CONCENTRATED AT THE LOWER FREQUENCIES SPEECH SPECTRUM CHANGES RELATIVELY SLOWLY (AS THE VIBRATION RATE OF THE VOCAL CORDS ALSO CHANGES RELATIVELY SLOWLY) SPEECH SOUNDS CAN BE MODELLED (AS PERIODIC & NOISY EXCITATIONS PASSING THROUGH THE VOCAL TRACT / FILTER)

INTERMITTENT NATURE OF SPEECH SPEECH TRANSMISSION CHANNEL LIES INACTIVE (3) END PARTY DELAYS 25% FACTORS CONTRIBUTING INACTIVENESS (1) TALK / LISTEN NATURE OF A NORMAL TWO WAY TELEPHONIC CONVERSATION TRANSMISSION IN ONE DIRECTION IS FOR < 50% (2) PAUSES 33% SPEECH SPURTS 1.5 seconds

TLAF AVERAGE FRACTION OF THE TOTAL CONNECT TIME DROPS TO AS LOW AS 25 % = T elephone L oad A ctivity F actor = TLAF > 0.25 ITU THE CICUIT UTILISATION CAN BE IMPROVED BY DEMAND ASSIGNMENT OF THE CHANNEL

EAR STRONGER SENSITIVITY TO A SMALLER PORTION OF THE AUDIBLE FREQUENCY SPECTRUM LESS SENSITIVITY TO THE PHASE OF THE SIGNAL

THE DEVICE WHAT IS THE DEVICE…..? SPEECH CODECS

CLASSIFICATION / CATEGORIES WAVEFORM CODECS PCM DPCM ADPCM CODE EXCITED LINEAR PREDICTIVE VOCODERS LPC HYBRID CELP

PERFORMANCE PARAMETERS COMPLEXITY MOPS DELAY 30 m sec BIT RATE QUALITY COST

BIT RATE DGREE OF COMPRESSION 8 BIT PCM 64 kb/s 11 BIT PCM 88 kb/s ADPCM 32 kb/s LPC < 4.8 kb/s

SPEECH - QUALITY OBJECTIVE MEASURES SUBJECTIVE MEASURES SNR HARMONIC DISTORTION BW M.O.S MEAN OPINION SCORE

OBJECTIVE MEASURES 4 QUALITY CLASSES BROADCAST TOLL C OMMUNICATION SYNTHATIC >7 kHz >30 3.1 kHz >30 2-3 2-3 HIGH QUALITY RADIO/TV PSTN POORER BUT STILL INTELLIGIBLE WLL , MOBILE, MILITARY BW SNR HD % < 2 kb/s RECOGNIZABILITY LOST MACHINELIKE ROBOTICS

SUBJECTIVE MEASURES M.O.S. QUALITY LISTENING EFFORTS 5 EXCELLENT NO EFFORT REQUIRED 4 GOOD ATTENTION NECESSARY BUT NO APPRICIABLE EFFORT REQUAD 3 FAIR MODERATE EFFORT REQUAD 2 POOR CONSIDERABLE EFFORT REQUAD 1 BAD NO MEANING UNDERSTOOD WITH REASONABLE EFFORTS

M.O.S. 11 BIT PCM 4.9 08 BIT PCM (64 kb/s) 4.3 ERROR FREE SPEECH ADPCM (32 kb/s) 4.1 MINIMUM 3.5

PARAMTERS AT A GIVEN BIT RATE , SPEECH QUALITY CAN BE IMPROVED BY ------------- THE COMPLEXITY OF THE ALGORITHM & -------------THE DELAY. INCREASE IN INCREASING

CLASSIFICATION / CATEGORIES WAVEFORM CODECS SOURCE INDEPENDENT SUPERIOR SPEECH QUALITY HR BIT RATES VOCODERS SOURCE CODER VERY HIGH ECONOMY IN BIT RATE COMPLEX

VOCODERS CODER DECODER SPEECH DIGITIZED (CODED SPEECH) SIGNALS CODED SPEECH SPEECH OUTPUT

CODER ADC ADC ADC SPECTRUM ANALIZER EXCITATION ANALIZER M U X SPEECH VOCAL TRACT RESPONSE(FILTER ) VOICE DECISION PITCH DIGITIZED SPEECH

DECODER NOISE SOURCE FILTER SPEECH O/P SYNTHESIZED PITCH GEN DEMUX PITCH VOICING DECISION FILTER SHAPING DATA CODED SPEECH I/P

EXAMPLES BIT COMPLEXITY DELAY QUALITY M.O.S. RATE kb/s MOPS ms PCM 64 0.01 0 TOLL (High) 4.3 ADPCM 32 0.1 0 TOLL (High) 4.1 LPC 2.4 1.0 35 SYNTHETIC 2.5 MULTI PULSE 8.0 10 35 COMMUN 3.7 LPC

EXAMPLES RELP VSAT 16kb/s RESIDUAL EXCITED LINEAR PREDICTIVE RPE-LTP GSM 13.6 kb/s REGULAR PULSE EXCITED LONG TERM PREDICTIVE

TUTORIAL (1) WILL - IS IT A FACILITY TO THE SUBSCRIBER JUST LIKE CELLPHONE? OR A TOOL FOR THE OPERATORS ? (2)DIFFERENCE BETWEEN WLL & WILL IS-- (3)GIVE THREE MAJOR DIFFERENCES BETWEEN MICRO & MACRO CELL WILL (4) GIVE THREE EXAMPLES OF EACH

TUTORIAL (5) WHAT DO YOU MEAN BY “ CELL ” IN WILL? (6) WHY CELLS ARE REQUIRED IN A WILL SYSTEM ? (7) ON WHAT FACTORS DOES THE CELL SIZE DEPEND? HOW TO DOUBLE THE CELL SIZE ? (8) WHY TO FURTHER DIVIDE THE CELL INTO SECTORS ?

TUTORIAL (9) CALCULATE HOW MANY SUBSCRIBERS CAN BE SERVED BY ONE BS/ CELL ASSUMING FOLLOWING DATA SIX SECTORS PER CELL WITH 10 RF 1.25 MHz BW CHANNELS IN EACH SECTOR 0.1 Erlg TRAFFIC PER SUBSCRIBER & 20 TO 40 ACTIVE LINES PER RF CHANNEL

TUTORIAL (10) DRAW FUNTIONAL SCHEMATIC BLOCK DIGRAM OF - DECT SYSTEM ARCHITECTURE - CDMA ””””””””””””””””” (11) IN DECT HOW DIFFERENT TONES (eg dial tone , NU tone) ARE SENT FROM CBS TO SE

TUTORIAL (12) WHAT IS THE BANDWIDTH / BIT RATE (kbps) PER SUBSCRIBER IN CDMA AND DECT and HOW IS IT ACHIEVED ? (13)WHAT ROLE THE SPEECH CODECS PLAY IN INCREASING THE SYSTEM CAPACITY IN A WLL SYSTEM? (14) SPEECH CODECS CAN BE CLASSIFIED AS ------

TUTORIAL (15) DIFFERENTIATE BETWEEN 1G AND 2G SYSTEMS OF MOBILE SERVICES . (16) HOW MANY BS AND SUBSCRIBERS CAN BE SERVED BY ONE DIU ?

TUTORIAL FILL IN THE BLANKS (1).MAXIMUM TRAFFIC HANDLED BY ONE CBS (DECT) IS --- AT 1%BLOCKING AND THUS 1 CBS CAN SERVE --------SUBSCRIBERS IF 0.05 Erlg / SUBS. THE INFORMATION RATE IS ------kbps ie THE RATE OF ------SPEECH CODEC

TUTORIAL (2) DECT WLL TENDER COVERS ----CITIES WITH ----LINES IN TOTAL -----SUPPLIED BY --------------(HFCL/SHYAM/LG/LUCENT) & -----------BY ------------------(HFCL/SHYAM/LG/LUCENT).

TUTORIAL (3) IN VPT ------(CDMA/DECT) BASED WLL IS BEING INSTALLED WHICH IS OF ------(HFCL/SHYAM/LG/LUCENT) MAKE WITH ----(0/25/75/100 %)MOBILE AND ----(0/25/75/100 %)FIXED LINES ----------(68000 / 5 LAKH)LINES IN ALL

TUTORIAL (4) FWT (LG’S CDMA FOR VPT) HAS ------------(INBUILT / SEPARATE) BATTERY WITH-----(2 / 20 HOURS) TALK TIME--------(10/20 HOURS) CHARGE TIME USES -------(SOLAR / MAINS / EITHER) CHARGING . (5)THE DIFFERENCE BETWEEN WLL AND WILL IS THAT

TUTORIAL (1)Vocoder with lower bit rate & the same Quality speech will be MORE / LESS Complex , MORE /LESS expensive (cost) , will add LESS/ MORE Delay (2)M.O.S. should not be less than------------

TUTORIAL TRUE / FALSE (1)IN CDMA THE CELL SIZE CAN BE INCREASE TO ANY LEVEL JUST BY INCREASING THE BS POWER (2)GSM , CDMA & DAMS ARE EUROPIAN STANDARDS WHEREAS DECT , CT-2 & PHS ARE AMERICAN STANDARDS.

TUTORIAL TRUE / FALSE (3)GSM & DECT USE FDD WHEREAS DAMPS & CDMA USE TDD. (4)DAMPS IS CALLED DUAL MODE BECAUSE IT CAN WORK BOTH AS A MICRO-CELLULAR AS WELL AS A MACRO-CELLULAR WLL.

3a. Speech Coders

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to 3a. Speech Coders (20)

Recently uploaded (20)

3a. Speech Coders