Chapter 12 Multimedia Information

1. Chapter 12Multimedia Information Chapter Figures

2. Lossless data compression Data expansion ASDF9H... ASDF9H... 11010101... Communication Networks Figure 12.1

3. Assume • 5 symbol information source: {a,b,c,d,e} • symbol probabilities: {1/4, 1/4,1/4,1/8,1/8} 0 1 Symbol Codeword a 00 b 01 c 10 d 110 e 111 0 1 0 1 00 01 1 10 0 c a b 110 111 e d 17 bits aedbbad.... mapped into 00 111 110 01 01 00 110 ... Note: decoding done without commas or spaces Communication Networks Figure 12.2

4. (a) Building the tree code by Huffman algorithm e d c b a .15 .10 .05 .50 .20 1 .15 2 .30 3 .50 4 1.00 (b) The final tree code 1 0 a a 0 b 10 c 110 d 1110 e 1111 1 0 b 1 0 c 1 0 e d Communication Networks Figure 12.3

5. “Blank” in strings of alphanumeric information ------$5----3-------------$2--------$3------ • “0” (white) and “1” (black) in fax documents Communication Networks Figure 12.4

6. Inputs: Run Length Codeword Codeword (m = 4) 1 0 00..00 0000 01 1 00..01 0001 001 2 00..10 0010 0001 3 00..11 0011 00001 4 . . 000001 5 . . 0000001 6 . . . . . . . . . . 000...01 2m – 2 11..10 1110 000...00 run >2m – 2 11..11 1111 m Communication Networks Figure 12.5

7. Example: Code 1, m = 4 000…001000…01000…0100…001… 137 bits Runs 25 15 57 36 Symbols >14 10>14>14>14 12 >14>14 6 >14 0 1111 1010 1111 1111 1111 1100 1111 1111 0110 1111 0000 44 bits 15w 10wb 15w 15w 15w 12wb 15w 15w 6wb 15w b Communication Networks Figure 12.6

8. Inputs: Run Length Codeword Codeword (m = 4) 1 0 10..00 10000 01 1 10..01 10001 001 2 10..10 10010 0001 3 10..11 10011 00001 4 . . 000001 5 . . 0000001 6 . . . . . . . . . . 000... 01 2m – 1 11..11 11111 000... 00 run >2m – 1 0 0 m + 1 Communication Networks Figure 12.7

9. Example: Code 2, m = 4 000…001000…01000…0100…001… 137 bits Runs 25 15 57 36 Symbols >15 9 >15>15>15 9 >15>15 4 15 0 11001 0 0 0 11001 0 0 10100 11111 26 bits Encoded Stream Decoded Stream 16w 9wb 16w 16w 16w 9wb 16w 16w 4wb 15wb Communication Networks Figure 12.8

10. (a) Huffman code applied to white runs and black runs (b) Encode differences between consecutive lines Communication Networks Figure 12.9

11. “All tall We all are tall. All small We all are small.” Can be mapped into: “All_ta[2,3]We_[6,4]are[4,5]._[1,4]sm[6,15][31,5].” Communication Networks Figure 12.10

12. Th e s p ee ch s i g n al l e v el v a r ie s w i th t i m(e) Communication Networks Figure 12.11

13. Smooth signal Successive differences Communication Networks Figure 12.12

14. Communication Networks Figure 12.13

15. Communication Networks Figure 12.14

16. Quantize the difference between prediction and actual signal: Encoder + d ( n ) to channel QUANTIZER + -  + LINEAR PREDICTOR h x ( n ) + +  x ( n ) Decoder ~ n d y ( n ) ( ) + + + LINEAR PREDICTOR h  x ( n ) ˜ ˜ ˆ y ( n )  - x ( n ) = x ( n ) +  d ( n )  - x ( n ) = d ( n ) -  d ( n ) = e ( n ) The end-to-end error is only the error introduced by the quantizer! Communication Networks Figure 12.15

17. (a) X(f) Q(f) f W -W 0 X(f) (b) Q(f) f W 0 -W Communication Networks Figure 12.16

18. 1-D DCT (a) X(f) x(t) (time) (frequency) (b) 2-D DCT 100 95 85 .. 102 99 70.. 101 80 70.. 95 77 65.. 80 10 3 2 0 8 2 1 0 .. 2 0 0 .. 0 0 ... 0 ... (n,m) Space (u,v) Frequency Communication Networks Figure 12.17

19. m x n block red component . . . . . . m x n block green component . . . . . . . . . 8m x 8n pixel color picture . . . m x n block component signal Communication Networks Figure 12.18

20. 111 22 15 5 1 0 0 0 14 17 10 4 1 0 0 0 2 2 1 0 0 0 0 0 -4 -4 -2 -1 0 0 0 0 -3 -3 -1 0 0 0 0 0 -1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 180 150 115 100 100 100 100 100 250 180 128 100 100 100 100 100 190 170 120 100 100 100 100 100 160 130 110 100 100 100 100 100 110 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 DCT 8x8 block of 8-bit pixel values Quantized DCT Coefficients Communication Networks Figure 12.19

21. 8x8 block Quantization VLC coding DCT Symmetric DCT/I-DCT transfer Huffman Tab DC: DPCM VLI AC: O-run/VLI Quantization Matrices Communication Networks Figure 12.20

22. Info = M bits/pixel x (WxH) pixels/frame x F frames/second 2+ 2- 1+ 1- . . . 30 fps Communication Networks Figure 12.21

23. Fn Fn-1 (+1,+2) Hybrid Encoder: (x,y) Frame buffer Fn-1 Motion Vector Intra- /inter- frame processor Fn Encoder - Motion Vector - Error Block - Intra Block Communication Networks Figure 12.22

24. Intra Huffman VLC Q 8x8 block DCT - Inter Q-1 Filter CRC error and Fixed-length control p x 64 Frame Memory I-DCT Motion Estimation Motion Vector Communication Networks Figure 12.23

25. 1-D examples: x x Quantize individual samples Interpolation Linear prediction Fn+1 Fn+1 Bidirectional MC Fn Fn-1 - Intra - Forward - Backward - Bidirectional Fn Fn-1 Communication Networks Figure 12.24

26. Pred 12 Bidirectional Motion Compensation Intra 9 B11 Pred 6 B10 B8 B7 Pred 3 B5 16 x 16 bidirectional macroblocks Intra 0 B4 B2 B1 - Intra - Forward - Reverse - Bidirectional Communication Networks Figure 12.25

27. SNR scalability Enhancement stream High quality Decoder Base stream Low quality Decoder Spatial Scalability Enhancement stream Large image Decoder Base stream Small image Decoder Communication Networks Figure 12.26

28. Audio encoder Packetizer Packet stream multiplexer Program stream PES streams Video encoder Packetizer Transport stream multiplexer Transport stream MPEG systems layer Communication Networks Figure 12.27