slides - network systems lab @ sfu

Report
Flexible Transport of 3-D Videos
Over Networks
Ahmed Hamza
Network Systems Lab
Simon Fraser University
July 15, 2013
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Introduction
Introduction
 In the near term, popular 3-D media will most likely be in the
form of stereoscopic and multi-view video.
 Transmission of 3-D media, via broadcast or on-demand, to
end users with varying 3-D display terminals (e.g., TV, laptop,
and mobile devices) and bandwidths is one of the biggest
challenges to bring 3-D media to the home and mobile
devices.
 Two main platforms for 3-D video delivery:
 digital television (DTV) platforms
 Internet Protocol (IP) platforms
Platform for 3D Media Transport
IP-based Delivery Platforms
 IPTV
 multimedia services delivered over IP-based managed
networks that provide the required level of quality of
service (QoS) and experience, security, interactivity, and
reliability
 WebTV
 services offered over Internet connections that support
best effort delivery with no QoS guarantees, making them
accessible anytime, anywhere as opposed to IPTV
Hybrid DTV-IP Approach
 The DVB channel is constrained by the physical channel
bandwidth to allow transmitting multi-view video (MVV).
 The IP platform is more flexible in terms of bandwidth but is
not reliable.
 A more recent research direction is to consider a
combination of DVB and IP platforms to deliver MVV to
provide free-view TV/video experience.
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Stereoscopic Video
 The most simple 3D video data representation
 Each of the two captured views is presented to one of
the eyes
 Can be multiplexed either spatially (passive) or
temporally (active)
 Temporal multiplexing has the advantage of maintaining the full
resolution of each view
 Disadvantage:
 hardware representation dependency (acquisition process is
tailored to a specific type of displays, baseline distance between
the two cameras is fixed)
Multiplexing Stereo Video
Spatial Multiplexing
(half the resolution)
Time Multiplexing
(double the frame rate)
Video Plus Depth
 2D video signal along with geometry information of the scene
texture
depth map
Multi-view Plus Depth (MVD)
Cam-6
Cam-3
Cam-0
3D Image Warping
Example
Ismaël Daribo and Hideo Saito, “A Novel Inpainting-Based Layered Depth Video for 3DTV,” IEEE
Transactions on Broadcasting, vol. 57, no. 2, June 2011
Layered Depth Video (LDV)
Main Layer
(central color view and depth map)
Enhancement Layer
(color and depth occlusions)
projected on central viewpoint
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Three-Dimensional Video Coding
 3-D video encoding depends on the transport option and raw
video format.
 Simulcast encoding:
 encode each view and/or depth map independently using a
scalable or non-scalable monocular video codec
 enables streaming each view over separate channels
 clients can request as many views as their 3-D displays require
 Dependent encoding:
 encode views using MVC to decrease the overall bit rate by
exploiting the inter-view redundancies
 a special inter-view prediction structure must be employed to
enable view-scalable and view-selective adaptive streaming
Multi-view Video Coding (MVC)
 Multi-view extension of H.264/AVC
 Enables inter-view prediction
 Prediction structure is simplified by restricting interview prediction to anchor pictures only
 Large disparity or different camera calibration affects
coding efficiency
 Reference MVC software (JMVC)
 temporal and view scalability
Multi-view Video Coding (MVC)
Multi-view Plus Depth Coding
 Independently code views and depth maps
 Dependent encoding is also possible
 Exploit correlation between texture and depth map
 Examples:
 sharing the texture video MVs with the depth map
 utilizing inter-layer motion prediction tool in SVC
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Transport Protocols
 Transmission Control Protocol (TCP)
 may not be suitable for streaming live video with a strict
end-to-end delay constraint
 lack of control on delay (retransmissions)
 rapidly changing transmission rate (congestion control)
 provides good performance when available network
bandwidth is about twice the maximum video rate (few
seconds pre-roll delay)
Transport Protocols
 Datagram congestion control protocol (DCCP)
 implements bidirectional unicast connections
 both data and acknowledgements can flow in both directions
 congestion-controlled, unreliable datagrams
 congestion control mechanism selected at connection
startup
 outperforms TCP under congestion when a video
streaming scenario is considered
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
P2P Streaming
 Traditional client-server unicast streaming model is not
scalable by nature.
 Advantage of P2P solutions
 scalable media distribution (reduce the bandwidth requirement
of the server by utilizing the network capacity of the
clients/peers)
 P2P solutions use overlay networks (data are redirected to
another peer by the application)
Tree-Based Approach
 Efficient for delivering content from the server that is at
the top of the tree to peers that are connected to each
other in parent–child fashion.
 Shortcomings:
 ungraceful peer exit leads its descendants to starvation
 replicating the content for feeding multiple trees leads to
redundancy within the network
Tree-Based Approach
Mesh-Based Approach
 Data are distributed over an unstructured network in
which each peer can connect to multiple peers.
 Increased connectivity alleviates the problem of
ungraceful peer exit.
 building multiple connections dynamically requires a
certain amount of time (initiation interval)
 More suitable for applications that may tolerate some
initiation interval.
 Example: BitTorrent
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Adaptive Streaming
 A mechanism should exist to estimate the network
conditions so as to adapt the video rate accordingly, in
order to optimize the received video quality.
 Estimation can be performed by
 requesting receiver buffer occupancy status (to prevent
buffer underflow/overflow)
 combining receiver buffer status with bandwidth
estimation
Adaptive Streaming
 DCCP + TCP-friendly rate control (TFRC)
 TFRC rate calculated by DCCP can be utilized by the sender
to estimate the available network rate
 When the video is streamed over TCP, an average of the
transmission rate can be used to determine the
available network bandwidth
 Basic method in DASH
Video Rate Adaptation Methods
 Adapting video rate to available bandwidth depends on
the encoding characteristics of the views.
 One or more views can be encoded multiple times with
varying bit rates, sender can switch between these
streams according to the network conditions
 Similar to HTTP live streaming
 Encoding views once with multiple layers using SVC and
switching between these layers
 Real-time encoding with source rate control
 Difficult with MVV
Adaptive Stereoscopic
Video Streaming
 The behavior of the human visual system is another
paradigm for QoE-aware rate adaptation.
 Exploit the suppression theory
 human visual system (HVS) tolerates lack of highfrequency components in one of the views
 One of the views may be presented at a lower quality
without degrading the 3-D video perception.
 Asymmetric quality allocation
Just Noticeable Distortion for
Asymmetric Stereo Coding
 Asymmetry can be achieved by scaling the quality in
one of the views (secondary view)
 in spatial, signal-to-noise ratio (SNR) or temporal
dimensions
 Questions
 Which method should be used?
 What is the level of asymmetry before observers start
noticing visible degradations?
Just Noticeable Distortion for
Asymmetric Stereo Coding
Video Sequence
Threshold PSNR (dB)
Parallax Barrier
Polarized Projector
Adile
31.9
33.07
Iceberg
31.64
33.05
Flower Pot
31.19
33.2
Train Tunnel
31.74
32.88
 Results show that the “just noticeable” threshold PSNR is
 33 dB for the polarized projection display
 31.5 dB for the parallax barrier display
Asymmetric Encoding
for Adaptive Streaming
 Asymmetric Coding at a Fixed Rate Using MVC
 Spatial asymmetry
 using additional down-sampling steps in the encoding
process
 Temporal asymmetry
 skipping frames skipping from secondary view
 SNR (quality) asymmetry
 straightforward compared to other types of asymmetry
(encoding quality of a view depends on the quantization
parameter used)
Asymmetric MVC Coding
 Alternating views are coded at high and low quality.
 Inter-view dependencies should be carefully
constructed (predict only from high-quality views).
Asymmetric MVC Coding
Asymmetric MVC Coding
Asymmetric Encoding
for Adaptive Streaming
 Scalable Asymmetric Coding Using SVC
 It is possible to obtain spatial and/or quality scalable right
and left views if they are simulcast coded using the SVC
standard.
 Two encoding options for achieving scalable asymmetric
stereoscopic video bitstreams when simulcast coding is
used:
 encoding both views using SVC
 encoding one view with SVC and the other with H.264/AVC
Scalable Video Coding (SVC)
Asymmetric Encoding
for Stereoscopic 3D Video
 Can be done in two ways:
 encode both views using SVC
 base layer of each view is encoded with a quality ~32 dB
 enhancement layers are encoded at the maximum quality
according to channel capacity
 only one view (the first) is scalably encoded
 second view is encoded using non-scalable H.264/AVC
 When the available link capacity is high, the scalable coded
view (with the enhancement layer) becomes the high-quality
view.
Asymmetric Encoding
for Stereoscopic 3D Video
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Adaptive Multi-view
Video Streaming
 Straightforward approach:
 extend the concept of asymmetric coding to MVV streaming (for
relatively small number of views)
 A more efficient (in terms of bandwidth consumption) and
flexible (in terms of number of views) approach:
 streaming the MVD representation (includes view scalability)
 View-selective encoding and interactive streaming of multiview video
 requires computer vision methods for real-time head/gaze
tracking, can be used to limit the number of views transmitted
View Scaling
 Discarding one view entirely and falling back to 2D video is
not a good choice.
 switching from 3D to 2D results in significant viewing discomfort
 With multi-view video (MVV) format, view scaling is a
possible option
 missing view(s) may be outside of the user’s field of view or can
be replaced by an artificial view generated at the client side
 Challenge
 How to determine which view should be discarded for minimum
degradation in perceived quality?
QoE-based Adaptation Policy
 Subjective tests to evaluate the performance of scaling
methods in terms of delivered QoE under different
network conditions.
 5-view 3D display at 1920x1200 screen resolution
 12 male and 4 female assessors (7 experts)
Description
Method #
Detail
Symmetric Quality
Scaling
Asymmetric Quality
Scaling
View Scaling
1
2
3
4
5
6
SNR
Spatial
SNR
Spatial
3c+3d
2c+2d
QoE-based Adaptation Policy
 Recommended adaptation policy:
State
Method
1
All views transmitted at max quality
2
Asymmetric SNR scaling of intermediate views
3
Keep only edge views (+ depth) and use DIBR
Adaptation-ready Encoding
 Introduce quality difference between adjacent views.
 View that are either transmitted or not are encoded with
H.264/AVC for high coding efficiency.
 Views that may have different qualities to achieve
asymmetry are encoded using SVC.
 Example:
 For a five-view display, can perform this efficiently using SVC for
views 2 and 4.
MVV Adaptation Example
a)
b)
c)
High link capacity (4.5 Mbps)
Low link capacity (3.3 Mbps)
Very Low capacity (2.1 Mbps)
Outline
 Introduction
 State of the Art




3D Video Representation
3D Video Coding
Transport Protocols
P2P Streaming
 Adaptive 3D Video Streaming
 Stereo Video
 Multi-view Video
 Case Study: DIOMEDES
Case Study: Project DIOMEDES
 European project
 Peer-assisted multi-view video broadcast
 Scalable architecture that utilizes the upload capacity of
peers to assist distribution of up to 200 views and
associated 3-D audio
 Main Idea:
 DVB-T signal provides stereoscopic 3-D media as a
baseline
 P2P distribution of remaining MVV views over IP to enable
immersive free-view TV experience
DIOMEDES Architecture
 Three modules:
 3-D content server
 master peers
 3-D media streaming server
 Peers that use both DVB
and IP channels synchronize
the received signals.
DIOMEDES Client
BitTorrent Protocol
 Adopts a mesh-based topology
 Flat connections with no hierarchy
 Adopts a divide-and-conquer approach and splits
content into equally sized chunks
 Peers are of two types:
 Seeders: have the whole content and upload chunks to
other peers
 Leechers: have some missing chunks
BitTorrent Protocol
 Chunk exchange is managed by two governing policies:
 Rarest-first chunk scheduling
 Determines the chunks to be requested
 Favors chunks that least distributed
 Tit-for-Tat




Determines which chunk requests are to be accepted
Sort neighbours based on their level of contribution
May deny requests from neighbours at lower ranks
Optimistic unchoking
Modifications to BitTorrent for
3D Video Streaming
 Chunk Mapping
 Variable-size layered chunks
 All chunks are self-decodable
 Each chunk contains multiple GoPs
 Adaptive video streaming
 in 3D video streaming, rate adaptation is not
straightforward and may depend on external information
such as the user’s field of view, the encoding scheme, and
the display properties
Modifications to BitTorrent for
3D Video Streaming
 P2P engine determines when to perform adaptation (discard/add a
stream).
 Adaptation module determines which streams should be affected first.
Modifications to BitTorrent for
3D Video Streaming
 Chunk Downloading
 ready-to-play buffer
 buffer duration is a variable that provides feedback on the
overall content retrieval rate
Modifications to BitTorrent for
3D Video Streaming
 Chunk uploading
 Request prioritization
 Favor requests that belong to streams of high priority
 Depth map streams should have the highest priority
because they are used to generate multiple views at the
client side.
 Base and enhancement layers may be prioritized similar to
the case of 2D video streaming
Other Use Cases
 Full-resolution stereoscopic 3D video delivery
 Full resolution source video is encoded as an
enhancement layer to the base stream in a framecompatible format that is transmitted over the DTV
channel.
 The enhancement layer is transmitted to enable full
resolution 3D video for users with Internet access
Other Use Cases
Other Use Cases
 Head tracking system for multi-view video delivery
 head tracking system coupled with a stereoscopic display
 View pairs change according to a user’s viewing position
 if the available link capacity is low, only required video
streams are received, based on the feedback from the
head tracking device
 increase efficiency of rapid view selection by using a
sparse camera arrangement and transmitting
corresponding depth maps
Conclusions
 Digital TV platforms are not flexible to support multi-view
video (cannot provide sufficient bandwidth).
 Three adaptive streaming solutions:
 Asymmetric streaming
 Streaming using MVD
 Selective streaming
 Combining adaptation methods with adaptive P2P video
streaming will provide a successful 3D video services solution
in the near future.
 Streaming holographic 3D video over IP might be possible on
the long term.
References
 Flexible Transport of 3-D Video Over Networks, Proceedings of the
IEEE, 2011
 Peer-to-peer system design for adaptive 3D video streaming, IEEE
Communications Magazine, 2013
 DIOMEDES: Content Aware and Adaptive Delivery of 3D Media
over P2P/IP and DVB-T2, Networked & Electronic Media (NEM)
Summit, 2011
 Evaluation of Asymmetric Stereo Video Coding and Rate Scaling for
Adaptive 3D Video Streaming, IEEE Transactions on Broadcasting,
2011
Thank You!

similar documents