This document discusses multimedia networks and properties of video and audio. It covers three types of multimedia network applications: 1) Streaming stored audio/video such as YouTube where content is pre-recorded. 2) Conversational voice/video over IP like Skype which is delay sensitive but loss tolerant. 3) Streaming live audio/video similar to broadcasting where a delay of up to 10 seconds is acceptable. It also describes techniques for compressing, encoding, and transmitting video and audio over networks, including spatial/temporal redundancy, PCM encoding, MP3/AAC compression, and client-side buffering used in UDP, HTTP, and adaptive HTTP streaming.

1. Information and Communication
Technologies(ICT)
Principles and Perspectives
Lecture : Multimedia Networks
Study Guide(17) essential reading (7)

2. Multimedia Applications
• A multimedia network application as any network application that
employs audio or video
• Can be classified in
1. streaming stored audio/video
2. conversational voice/video-over-IP
3. streaming live audio/video.

3. Video Properties
• High bit rate (100Kbps to 3Mbps)
• Bit rate comparison for different apps
• 90% internet traffic is consumed by videos
Kbytes Seconds Number of photos
Avg Photo size Total time 10 secs per image Kbps kilo bit size
Photos 200 4000 400 80 80MB 160 1600000
audio 4000 64 64MB 128
Video 4000 1000 1GB 2000

4. Video properties: Compression
• Videos can be compressed, trading off quality with bit rate
• Video
• Sequence of Images (24-30 per second)
• Image is array of pixels, each pixel encoded represent luminance and color
• For compression we can exploit redundancies
1. Spatial Redundancy
2. Temporal Redundancy

5. Spatial redundancy
• Repetition within an image
• More white space, can be compressed more

6. Temporal redundancy
• Repetition from image to subsequent image
• If subsequent image is same then no need to re-encode it

7. Compression continues
• Today multiple versions of same videos are created with different bit rate
• User can choose according to their network speed
• Video in video conference application is compressed on the fly

8. Audio Properties
• Analog audio signal is sampled at some fixed rate (e.g. 8000 per second)
• Each sample is rounded to one finite number of values. Operation known as
quantization. These values quantization values
• Each quantization value is represented by fix number of bits
• For example of 8000 samples per second is quantized and represented by
8bits. Then for one second digital signal rate would be
8000 x 8 = 64000 bps

9. Pulse code modulation
• PCM is basic encoding scheme
• 8000 samples with 8 bit per sample results in 64kbps
• CD use PCM with 44100 samples and 16 bit per sample (705.6kbps) Mono
• 1.411 Mbps for stereo
• Compression techniques
• Mp3 can compress to many different bit rates(128kbps is most popular)
• Advanced audio coding(AAC)

10. Type of multimedia network apps
Streaming stored audio/video
• Content is pre-recorded video or audio (youtube)
• 50% downstream traffic is of streaming stored video
• Features
• Streaming: play out starts as soon as some portion is received. Playing as well as
receiving later parts of video
• Interactivity: play, pause, forward
• Continuous playout: video should play according to original timing else frame freezing
or frame skipping experience

11. Streaming stored audio/video
• Most important performance measure is throughput
• It should be minimum of bit rate of video
• Techniques like prefetching, buffering

12. Conversational voice/video-over-IP
• Internet telephony (Voice over IP)
• Skype, Google talk
• Properties
• Timing consideration
• Tolerance of data loss

13. Conversational voice/video-over-IP
• Delay sensitive (Timing consideration)
• 150 millisec not perceived by human
• 150-400 millisec acceptable
• Loss tolerant
• Annoying but not particularly harmful if we compare with web, email

14. Streaming Live Audio & Video
• Similar to broadcasting
• Delay of up to 10 seconds is acceptable

15. Streaming stored video
1. UDP streaming
2. HTTP streaming
3. adaptive HTTP streaming
• A common characteristics of all 3 forms is client side buffering

16. Client side buffering

17. UDP Streaming
• Server tries to match client’s consumption rate
• Drawbacks
1. If available bandwidth drops client will face freezing or skipping
2. Requires media control server to process client to server interactivity requests. Play,
pause , managing client state becomes complex
3. Firewalls sometimes block UDP ports

18. HTTP Streaming
• video is simply stored in an HTTP server as an ordinary file with a specific
URL
• TCP congestion control and flow control
• Prefetching:
• Client can attempt to download video at higher rate than the consumption rate

19. Http Streaming TCP buffers

20. Analysis of client side buffering
• B is size in
bits
• Q number of
bits must be
buffered
before playout

21. Early Termination and Repositioning the
Video
• Requesting future point in time in the video
• Server can forget earlier requests and can start sending according to new
• Http byte range header is used for this
• Some pre-fetched video will go unwatched
• To avoid this client uses moderate size buffers
• Repositioning and early termination are analogous to cooking a large meal, eating only a
portion of it, and throwing the rest away, thereby wasting food

22. Adaptive Streaming and DASH
• On http choose video from different available bit rates
• With DASH, each video version is stored in the HTTP server, each with a
different URL