Lesson 11: Applications (Video)

Question 1

What can you tell about bit rate for video and it’s comparison to the bit rate of audio or browsing through an online photo gallery?

Answer 1

Video browsing requires more bandwidth than photo browsing which is about twice that of music.
(video 2Mbps > photo 320 bps > music 128 bps)

Question 2

What are the characteristics of streaming stored video?

Answer 2

video is interactive 
continuous playback (no freezing)
normally stored on a CDN

Question 3

What are the characteristics of streaming live audio and video?

Answer 3

many simultaneous users in different locations

delay-sensitive (10 sec max)

Question 4

What are the characteristics of conversational voice and video over IP?

Answer 4

3+ users
high delay-sensitive (0.4 sec max)
loss-tolerant

Question 5

How does the encoding of analog audio work (in simple terms)?

Answer 5

Audio is encoded by taking many (thousands) of samples per second, and then rounding each sample’s value to a discrete number within a particular range

Pulse Code Modulation (PCM) may record speech at 8,000 Hz at 8 bits and audio CD at 44,100 Hz @ 16bits

Question 6

What are the three major categories of VoIP encoding schemes?

Answer 6

narrowband, broadband, and multimode

Question 7

With regards to telephony and VoIP, what are the functions that signaling protocols are responsible for?

Answer 7

1) User location
2) Session establishment - handling the callee accepting, rejecting, or redirecting a call.
3) Session negotiation - the endpoints synchronizing with each other on a set of properties for the session.
4) Call participation management - handling endpoints joining or leaving an existing session.

Question 8

What are 3 metrics used to measure the quality of VoIP?

Answer 8

• end-to-end delay
• jitter
• packet loss

Question 9

One of the QoS VoIP metrics is “delay jitter”. How does this phenomenon occur?

Answer 9

data packets experience different delays and are received at different times. The result can be more delayed packets and gaps in audio recieved.

Question 10

How does FEC (Forward Error Correction) deal with the packet loss in VoIP? What are the tradeoffs of FEC?

Answer 10

transmits redundant data alongside the main transmission.

tradeoffs:
redundant data may be of a lower quality
transmission requires more bandwidth
receiver may need to receive more before playing

Question 11

How does interleaving deal with the packet loss in VoIP/streaming stored audio? What are the tradeoffs of interleaving?

Answer 11

packets mix chunks of audio together. If a packet is lost, small gaps are preferred to one large one.

tradeoffs:
wait longer before starting

Question 12

How does error concealment technique deal with the packet loss in VoIP?

Answer 12

when a packet is missing, it repeats last packet or interpolates between the surrounding packets to guess at the sound.

Question 13

Provide a high-level overview of adaptive video streaming.

Answer 13

video is:

1) recorded
2) compressed
3) secured using DRM
4) replicated amongst geographic servers within CDN
5) downloaded by user
6) decoded and rendered on the screen

Question 14

An efficient video compression can be achieved in two-ways. What are these ways?

Answer 14

1) spacial redundancy - pixels within an image can be similar.
2) temporal redundancy - consecutive frames can be similar

Question 15

What are the 4 steps of JPEG compression?

Answer 15

1. Transform RGB image into color (Cr, Cb) and brightness ( Y ) components
2. Obtain frequency domain matrix by applying the Discrete Cosine Transformation to 8x8 blocks.
3. Compress the resulting matrix using a pre-defined Quantization table.
4. Perform a lossless encoding.

Question 16

Explain video compression.

Answer 16

In video compression the initial or i-frame is encoded as a jpeg. Next frames in the same scene are encoded and called predicted frames (or p-frames).
Additionally, a Bi-frame (or b-frame may be encoded between the i & p-frames to improve coding efficiency.

Question 17

What is the difference between constant bitrate encoding and variable bitrate encoding (CBR vs VBR)?

Answer 17

CBR - output size of the video is fixed

VBR - output size remains the same on an average, but varies based on the underlying scene complexity.

Question 18

Which protocol is preferred for video content delivery - UDP or TCP? Why?

Answer 18

TCP

• if an i or p-frame is lost decoding is impossible (TCP ensures delivery)
• TCP provides congestion control.

Question 19

What was the original vision of the application-level protocol for video content delivery and why was HTTP chosen eventually?

Answer 19

Original vision was to have specialized video servers that remembered the state of the clients. Intelligence is on the server and clients do minimal work.

HTTP was in existence, providers can use existing CDNs, bypassing middleware with HTTP was possible.

Question 20

When streaming video, what is a “byte-range request” and why is it useful?

Answer 20

A request for a small sample of the video file.

This is useful to avoid downloading the entire video file such as when browsing.

Question 21

What is bitrate adaptation

Answer 21

The ability of a server to adapt to a bitrate specified by the client during a video stream.

For example, A server will host video in short segments at various bitrates (250kbps, 500 kbps, 3Mbps, etc.). Depending on the bandwidth available, the client will request the segment and quality. Throughout a video bitrate may vary or adapt.

Question 22

How does the bitrate adaptation work in DASH?

Answer 22

Video in DASH is divided into chunks and each chunk is encoded in multiple bitrates. The bitrate adaptation algorithm at the client adapts the video bitrate to request based on network conditions.

Question 23

What are the goals of the bitrate adaptation algorithm?

Answer 23

optimize the user’s viewing quality of experience

this is done by:

• low or zero re-buffering
• high quality video
• low quality variations
• low start latency

Question 24

What are the different signals that can serve as an input to a bitrate adaptation algorithm?

Answer 24

• network throughput
• video buffer: full buffer can allow time to download high quality vid, low buffer can be filled quickly with low quality.

Question 25

Define the buffer-filling rate.

Answer 25

The amount of video in seconds received per second.

the network bandwidth divided by the chunk bitrate.

If the bandwidth is 10 Mbps, and the bitrate of the chunk is 1 Mbps. Then, in 1 second we can download 10s of video

Question 26

What steps does a simple rate-based adaptation algorithm perform?

Answer 26

1. estimate bandwidth

2. select max bitrate for bandwidth

Question 27

Explain the problem of bandwidth over-estimation with rate-based adaptation.

Answer 27

If the bandwidth drops suddenly, a high bit rate may be requested before, the video player’s calculation of the average bandwidth has converged causing the player to rebuffer.

Question 28

Explain the problem of bandwidth under-estimation with rate-based adaptation.

Answer 28

Due to the nature of DASH’s on-off request behavior to fill a buffer, video streaming can lose its share of the TCP bandwidth. It can lose so much share that it selects a lower bit-rate which only gives up more bandwidth to competing flows.

Question 29

End-to-end delay is the accumulation of all of these sources of delay.

Answer 29

• the time it takes to encode the audio,
• the time it takes to put it in packets,
• all the normal sources of network delay that network traffic encounters such as queueing delays,
• “playback delay,” which comes from the receiver’s playback buffer
• decoding delay, (reconstruct the signal.)
  (see QoS for VoIP)

Question 30

What range of end-to-end delay for VoIP is noticeable but acceptable?

Answer 30

150-400 ms

Question 31

What are the mitigation techniques for delay jitter?

Answer 31

maintaining a buffer a “jitter buffer” or “play-out buffer.”

Question 32

What are VoIP protocols three major methods of dealing with packet loss?

Answer 32

1. FEC (Forward Error Correction),
2. interleaving,
3. error concealment.

Question 33

Define temporal redundancy.

Answer 33

Temporal redundancy occurs in a continuous scene when consecutive pictures are similar.

Question 34

In terms of video streaming what is a playout buffer?

Answer 34

A bucket (specified in seconds) used to store video to account for variations in network download.

Question 35

When streaming what are the two states of the playout buffer?

Answer 35

Filling state - occurs when buffer is empty

Steady state - client request more when buffer falls below threshold.