B2P1 - Channel coding

Question 1

automatic repeat request (ARQ)

Answer 1

IF (there are errors OR no acknowledgement) THEN (data is resent);

Question 2

block (stateless) code

Answer 2

a channel code with no memory in encoder;

Question 3

chase combining

Answer 3

where decoder uses both sent_data and resent_data;

Question 4

check digits

Answer 4

the parity bits added to data for error-control coding;

Question 5

code rate

Answer 5

= k/n;

Question 6

code redundancy

Answer 6

parityBits / codewordBits = (n - k)/n;

Question 7

coding gain (dB)

Answer 7

how much power is saved with the S/N ratio due to using error-control coding;

Question 8

concatenation

Answer 8

output_code1 = input_code2;

Question 9

convolutional / tree / trellis (stateful) code

Answer 9

a channel code with memory in encoder;

Question 10

cyclic redundancy check (CRC)

Answer 10

error-detection using mod-2_arithmetic and applied to binary_sequences;

Question 11

d.c. coefficient

Answer 11

reps the average value of all pixels within an 8 x8 macroblock in JPEG, and appears as top-left entry in DCT matrix repping the macroblock;

Question 12

error-control coding

Answer 12

adding check bits to message data to detect and / or correct errors;

Question 13

error-detection coding

Answer 13

where errors can be detected but not corrected;

Question 14

forward error correction (FEC)

Answer 14

any method of error-correction where data does not need to be resent;

Question 15

Golay code

Answer 15

a (23, 12)-blockCode with minHamDistance = 7;

Question 16

discrete cosine transformation (DCT)

Answer 16

for JPEG and MPEG:

DCT(spatialRep) = frequencyRep;

Question 17

hard-decision decoding

Answer 17

where demodulator only passes binary values {0,1} to decoder;

Question 18

hybrid automatic repeat request (HARQ)

Answer 18

IF (errors > FEC_capability) THEN (resend);

Question 19

incremental redundancy

Answer 19

where resend = send + extraCheckBits;

Question 20

interleaving

Answer 20

matrixCodewords ==> transpose(matrixCodewords);

Question 21

frequency masking

Answer 21

where loud sounds cover up the quiet sounds of neighbouring frequencies;

Question 22

latency

Answer 22

a temporal quantity repping coding-overhead or full bit-transmission period;

Question 23

low-density parity check (LDPC) codes

Answer 23

advanced error-correction codes that can approach the Shannon limit;

Question 24

parity-check matrix

Answer 24

a matrix of 1s and 0s whose rows map each codeword-bit to parity-check-bits;

Question 25

perfect code

Answer 25

where all information from parity bits is used;

Question 26

puncturing (stateful)

Answer 26

for stateful codes, where optimally selected output-bits are deleted to increase the code rate;

Question 27

inverse DCT

Answer 27

for JPEG and MPEG: | DCT_inv(frequencyRep) = spatialRep

Question 28

Reed-Solomon (RS) codes

Answer 28

allow for codeRate-errorCorrection trade-offs and can handle errorBursts;

Question 29

shortened code

Answer 29

where an (n,k) code block is shrunk by padding it with m*0s so that it becomes an (n-m,k-m) code;

Question 30

soft-decision decoding

Answer 30

where demodulator passes real-number values [0,1] to decoder;

Question 31

symbol

Answer 31

an element from an alphabet;

Question 32

symbol error

Answer 32

for RS coding, where an error is detected but its location is unknown;

Question 33

syndrome

Answer 33

for error-control code: | parityCheck(codewordBlock) = syndrome;

Question 34

trellis-coded modulation (TCM)

Answer 34

[stateful (de)coding + (de)modulation] for digitalTransmission;

Question 35

turbo codes

Answer 35

softDecisions + concatenation + iterativeDecoding ==> approaches channelCapacity;

Question 36

Type I HARQ

Answer 36

where resend = send;

Question 37

Type II HARQ

Answer 37

where resend != send;

Question 38

Viterbi decoding

Answer 38

used for stateful codes;