f
play

f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha - PowerPoint PPT Presentation

Nov 06, 2023 •600 likes •1.21k views

Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 1 Towards Understanding

  1. Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 1

  2. Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 2

  3. 3

  4. 4

  5. 5

  6. Hoppe’s Progressive Mesh Edge Collapse Vertex Split 6

  7. At the sender = ... + v k v 4 v 3 v 2 v 1 base mesh 7

  8. Transmission UDP TCP v 4 ... base v 1 v 2 v 3 v k mesh 8

  9. At the receiver v 4 ... base v 1 v 2 v 3 v k mesh ... ... 9

  10. Vertex Split what to split v how to split v1 v2 10

  11. base mesh 11

  12. 12

  13. 13

  14. 14

  15. complete mesh 15

  16. Receiver-Driven Protocol sender viewer what to split how to split 16

  17. View-Dependent Streaming: only request what is visible, in decreasing order of importance 17

  18. Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 18

  19. “quality” time request 19

  20. “quality” data rate delay time request 20

  21. 6 seconds 120 KBps 21

  22. 22

  23. 400 ms 120 KBps 23

  24. 24

  25. 400 ms 20 KBps 25

  26. 26

  27. Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 27

  28. what are the acceptable delay and data rate when streaming progressive meshes? 28

  29. why? 29

  30. design of peer discovery protocol 30

  31. provisioning sender’s bandwidth 31

  32. designing error control protocol 32

  33. what affects user tolerance level? 33

  34. shape and size of mesh 34

  35. user task 35

  36. eye-mesh distance 36

  37. shape and size of mesh user task eye-mesh distance 37

  38. how? 38

  39. let users interact with meshes transmitted with specific (delay, data rate) parameters 39

  40. ask: is it acceptable? 40

  41. The Details 41

  42. Thai Statue 253 KB + 16.8 MB 42

  43. Happy Buddha 221 KB + 2 MB 43

  44. Dragon 247 KB + 12.1 MB 44

  45. 20 40 60 80 100 KBps KBps KBps KBps KBps 0.4 s 1 s 2 s 3 s 4 s 5 s 6 s 45

  46. 20 40 60 80 100 KBps KBps KBps KBps KBps 0.4 s 1 s 2 s 3 s 4 s 5 s 6 s 46

  47. show to user in random order user unaware of parameters user can rotate/translate user indicates whether quality is satisfiable 47

  48. 38 participants 48

  49. 20 40 60 80 100 KBps KBps KBps KBps KBps ✓ ✓ ✓ ✗ ✗ 0.4 s 1 s 2 s Inconsistent results are filtered out 3 s 4 s 5 s 6 s 49

  50. Thai Statue Happy Buddha Dragon 100 75 % Satisfied Users 50 25 0 1 2 3 4 5 6 RTT (seconds) 50

  51. Thai Statue Happy Buddha Dragon 100 75 % Satisfied Users 50 25 0 20 40 60 80 100 Data Rate (KBps) 51

  52. 20 40 60 80 100 KBps KBps KBps KBps KBps ✓ ✗ ✗ 0.4 s ? ✓ 1 s 2 s ? ✗ 3 s ✗ 4 s ✗ 5 s ✗ 6 s 52

  53. what does it mean? 53

  54. 1 sec users have higher delay tolerance compared to other interactive media applications 54

  55. progressiveness helps? 55

  56. 480 kbps 16 - 18 K vertices / seconds average US upstream BW: 1.1 Mbps average YouTube video: 328 kbps 56

  57. what’s next? 57

  58. shape and size of mesh user task eye-mesh distance 58

  59. prefer higher rate or lower latency? 59

  60. Towards Understanding User Tolerance To Network Latency and Data Rate in f Remote Viewing of Progressive Meshes Ransi De Silva Nilaksha Cheng Wei Wei Tsang Ooi Shengdong Zhao National University of Singapore 60

Recommend

More recommend