Patent Grant
10567780
The present disclosure generally relates to communication networks. More particularly, and not by way of any limitation, the present disclosure is directed to a system and method for providing 360° immersive video in a streaming network environment.
The introduction of virtual reality has brought new applications to the forefront in addition to improving several existing technologies. One improvement over existing technologies can be seen in the case of 360° immersive video, also variously referred to as panoramic video, 360-degree video or 360 video, and the like.
360-degree video offers the user with an immersive “being there” experience. The increased immersion of virtual reality can easily be applied to video, providing superior user experience over the traditional video that is projected on flat surfaces. The popularity of navigable 360-degree video systems has also grown with the advent of omnidirectional capturing systems and interactive displaying systems, such as head-mounted displays (HMDs) or headsets. However, content providers have been contending with bandwidth constrained network environments to deliver 360-degree video content in an efficient way in order to insure a satisfactory viewing experience because 360-degree video assets are ultra high resolution spherical videos, which contain an omnidirectional view of the scenes requiring enormous amounts of data.
Current 360 video headsets are 2K-resolution display devices, covering 1K per eye. In order to achieve the best quality in the headset, a typical network requires sending an 8K 360 video stream to the device. It is known that video compression allows efficient utilization of bandwidth in a media streaming network by reducing the number of bits to represent a picture. Whereas advances in video compression technologies continue to grow apace, several lacunae remain in the field of 360 video delivery and display with respect to efficiently managing bandwidth in today's network environments, thereby requiring further innovation as will be set forth hereinbelow.
The present patent disclosure is broadly directed to systems, methods, apparatuses, devices, and associated non-transitory computer-readable media and network architecture for effectuating optimized 360° immersive video viewing experiences. In one aspect, certain embodiments are directed to optimized encoding schemes that may be implemented in an arrangement involving encoding of source video streams into tile-encoded streams having different qualities. In a further aspect, certain embodiments are directed to utilizing user gaze vector information in determining tile weights based on the location of tiles with respect to a user's viewport. In still further aspects, example embodiments are directed to tile selection and bandwidth annealing schemes where bandwidth-optimized tiles are selected responsive to gaze vector information and bandwidth allocation for stitching into a multiplexed coded video sequence for enhanced viewing experience.
In one embodiment directed to a media preparation system and method, a media input stream is coded into a first set of bitstreams having different qualities, e.g., related to QPs (quantization parameters), wherein each Group-of-Pictures (GOP) starts with an intra-coded (I) frame followed by a set of frames including at least one predictive-coded (P) frame. The media input stream is also coded into a second set of bitstreams having different qualities, wherein each GOP starts with an I-frame followed by a plurality of X-frames (also referred to as Block-Intra Encoded frames or All-Intra Encoded frames). In one variation, each X-frame is encoded having a slice header of a P-frame and comprising blocks of only intra-coded data (e.g., I-blocks). In another variation, an X-frame is encoded having a slice header of a bi-predictive coded frame (B-frame), with I-blocks of image data only. During playout of the 360° immersive video content, when a viewport changes during mid-GOP and/or if bandwidth allocation changes, higher quality data in an X-frame is stitched into the coded video sequence with respect to the new viewport for enhanced viewing experience with minimum latency.
In another embodiment directed to a media preparation system and method, a media input stream is coded to generate a plurality of bitrate representations of the media input stream, each bitrate representation having a separate video quality related to a QP value used for the bitrate representation. Each bitrate representation is encoded into a plurality of phase-encoded bitstreams, each phase-encoded bitstream that belongs to a particular bitrate representation comprising a number (N) of frames with a specific GOP structure having a GOP size (p), wherein the plurality of phase-encoded bitstreams equals the GOP size. For each pth phase-encoded bitstream, N frames are encoded as follows: at least a first frame is encoded as an intra-coded (I) frame; and a frame at a frame position i, for 2≤i≤N, that satisfies {i Mod (GOP size)}=p is encoded as an X-frame that has a slice header of a predictive-coded (P) frame and comprises blocks of only intra-coded data (I-blocks), otherwise, that frame is encoded as a P-frame having blocks of predictive-coded data (P-blocks) with a P-slice header. In example embodiments herein, “Mod” is the modulo operator returning the remainder after division. Other variations may employ X-frames with B-slice headers. During playback of the 360° immersive video content, when a viewport changes during mid-GOP and/or if bandwidth allocation changes, higher quality data in an X-frame from a phase-encoded bitstream having a suitable phase is muxed into the coded video sequence for enhanced viewing experience with minimum latency. In example embodiments herein, a bitstream having a suitable phase is a bitstream for which an X-frame is located at a desired frame location, e.g., at the earliest frame in decoding order that can be muxed into the coded video sequence.
In an example implementation, various types of tiled bitstreams, BIE bitstreams and/or PE bitstreams having different qualities may be generated based on at least one of High Efficiency Video Coding (HEVC) H.265 compression, Alliance for Open Media (AOMedia) Video 1 (AV1) compression and H.266 compression, also referred to as Versatile Video Coding (VVC) or Future Video Codec (FVC) compression.
In another aspect, an embodiment directed to a method operating at an immersive video optimization system comprises, inter alia, receiving a gaze vector from a client device operating to display a 360° immersive video asset to a user, wherein each video frame comprises an array of tiles projected on a 3-dimensional (3D) display environment viewed by the user in which the user is immersed, the gaze vector defining a gaze direction in the 3D display environment where the user is viewing at any particular time. An angular separation is determined between the gaze vector and a directional vector associated with each tile location respectively corresponding to the array of tiles in the 3D display environment. Responsive to the angular separations, a plurality of tile weights corresponding to the array of tiles are determined for use in selecting tiles of different bitrate qualities of the 360° immersive video asset in order to facilitate the assembly of a video frame having different qualities across the 360-degree view to be delivered to the client device. In one variation, the gaze direction is obtained by tracking an orientation of the user's headset associated with the client device for displaying the 360° immersive video asset. In another variation, the gaze direction is obtained by tracking a movement of the user's eyeballs with respect to different portions of the 3D display environment while the 360° immersive video asset is being displayed. In still further variations, a combination of ocular and head movements associated with the user may be implemented for determining gaze vector information.
In yet another aspect, an embodiment directed to a tile selection and bandwidth annealing method operating at an immersive video optimization system comprises, inter alia, retrieving one or more stream manifest files describing characteristics of tile groupings per frame for each tile-encoded bitstream corresponding to a particular one of a plurality of bitrate representations of a media input stream. Each bitrate representation or stream is encoded having a separate video quality that is related to a corresponding QP value used for the bitrate representation, wherein each frame of a tile-encoded bitstream comprises an array of tiles organized into at least one slice per frame, with a plurality of frames forming a GOP structure of the tile-encoded bitstream. Tile weights corresponding to an array of tiles forming a frame based on respective tile positions and variant weights corresponding to the plurality of tile-encoded bitstreams of the media input stream are obtained. An adequacy metric or value may be determined as a function of a variant weight and a tile weight for each tile/GOP combination across a GOP structure for each of the tile-encoded bitstreams. Responsive to the adequacy metric values, tiles having different bitrate qualities or tiles from bitstreams having different bitrate qualities may be selected from corresponding tile-coded bitstreams in an iterative process for effectuating knapsack optimization with respect to an overall bandwidth allocation, wherein the bitrate qualities of the selected tiles are optimized to satisfy a transmit buffer model for transmitting a multiplexed video output stream. The selected tiles are provided to a multiplexer or tile stitcher for generating a frame containing the selected tiles as part of the multiplexed video output stream, wherein the tiles in a viewport are selected from higher quality bitstreams. In one variation, the tile stitcher may be provided in a network-centric configuration at a video optimization node. In another variation, the tile stitcher may be provided in a device-centric configuration at a client device operative to consume 360° immersive video content.
In a still further aspect. the present disclosure is directed to a method of presenting video tiles of a 360° immersive video stream on a virtual reality (VR) device, wherein video quality is varied across a 360° panoramic view. The method includes, inter alia, determining a bitrate for each individual video tile based on an overall allocated bitrate of the video stream and the location of the individual tile relative to the user's direct field of vision. The method may include determining a user's current direct field of vision (e.g., Region of Interest or ROI); presenting within the user's current direct field of vision, video tiles having a bitrate higher than the overall bitrate of the video stream (or video tiles comprising blocks that are coded using relatively low QP values); and presenting outside the user's current direct field of vision, video tiles having a bitrate lower than the overall bitrate of the 360° video stream (or video tiles comprising blocks coded using relatively higher QP values), wherein the total bitrate of the video tiles within the user's field of view and the video tiles outside the user's field of view does not exceed the overall allocated bitrate of the 360° video stream. Skilled artisans will recognize upon reference hereto that the relative QP values and relative bitrate values are generally in an inverse relationship in an example implementation and may be interchangeably used in a particular embodiment, mutatis mutandis.
In another embodiment, the present disclosure is directed to a 360° video optimization system configured to present video tiles of a video stream on a VR device, wherein video quality is varied across a 360° panoramic view. The system determines a bitrate for each individual video tile based on an overall allocated bitrate of the 360 video stream and the location of the individual tile relative to the user's direct field of vision. The system accesses a video manifest (or obtains a link to the manifest) and receives an indication of the overall bitrate of the video stream from a video back office. The system receives or otherwise determines the user's current direct field of vision with respect to the VR device. The system presents within the user's current direct field of vision, video tiles having a bitrate higher than the overall bitrate of the 360° video stream; and presents outside the user's current direct field of vision, video tiles having a bitrate lower than the overall bitrate of the 360° video stream, wherein the total bitrate of the video tiles within the user's field of view and the video tiles outside the user's field of view does not exceed the overall allocated bitrate of the 360° video stream.
In still further aspects, an embodiment of a system, apparatus, or network element is disclosed which comprises, inter alia, suitable hardware such as processors and persistent memory having program instructions for executing an embodiment of the methods set forth herein.
In still further aspects, one or more embodiments of a non-transitory computer-readable medium or distributed media containing computer-executable program instructions or code portions stored thereon are disclosed for performing one or more embodiments of the methods of the present invention when executed by a processor entity of a network node, apparatus, system, network element, subscriber device, and the like, mutatis mutandis. Further features of the various embodiments are as claimed in the dependent claims.
Example embodiments disclosed herein provide several benefits in an immersive media consumption environment. Based on the encoding embodiments of the present invention and selecting tiles of different qualities based on an input gaze vector, bandwidth allocation and a transmit buffer model, frames may be advantageously assembled with the tiles optimized for higher quality viewports. Because the frames are selectively viewport-optimized, transport of high quality multiplexed streams is possible even in bandwidth-constrained environments without sacrificing the viewing experience. Example embodiments may be advantageously configured such that the highest quality tiles will always be delivered in the direct view, with controlled degrading qualities across multiple areas farther from the direct field of vision with the lowest quality in the area that is in the diametrically opposite direction of where the user is looking. Accordingly, when a stream is delivered to the device, the user always gets the highest video QoE in the area that they are directly looking at. Further, when the user moves their head, the mid-GOP switching of example embodiments allows receiving high quality tiles as quickly as possible with minimal latency. With the tiles encoded for gradual refresh, when a user changes their field of vision, example embodiments can further reduce the latency of the video as the size of the video buffer may be minimized by sending several high quality tiles in the initial upgrade of the next frame to deliver. Over the course of the next several frames, an example embodiment gradually increases the quality of the remaining tiles until the quality of tiles is reached based on the current field of vision and allowed bandwidth.
When considering a managed network environment, still further benefits and advantages may be realized in the practice of an example embodiment of the present invention. Because an example embodiment may be configured to utilize managed networks and optimize the delivery within a dedicated amount of network bandwidth, clients do not have to estimate the amount of bandwidth available and make adjustments for the bitrate of tiles being requested. This allows for the client to operate with a minimal buffer of 1-3 frames and avoid getting into paused buffering situations. Accordingly, a much higher reliability for the delivery of the stream may be achieved while providing a minimal latency for quality upgrades.
Additional benefits and advantages of the embodiments will be apparent in view of the following description and accompanying Figures.
Embodiments of the present disclosure are illustrated by way of example, and not by way of limitation, in the Figures of the accompanying drawings in which like references indicate similar elements. It should be noted that different references to “an” or “one” embodiment in this disclosure are not necessarily to the same embodiment, and such references may mean at least one. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to effect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.
The accompanying drawings are incorporated into and form a part of the specification to illustrate one or more exemplary embodiments of the present disclosure. Various advantages and features of the disclosure will be understood from the following Detailed Description taken in connection with the appended claims and with reference to the attached drawing Figures in which:
In the description herein for embodiments of the present invention, numerous specific details are provided, such as examples of components and/or methods, to provide a thorough understanding of embodiments of the present invention. One skilled in the relevant art will recognize, however, that an embodiment of the invention can be practiced without one or more of the specific details, or with other apparatus, systems, assemblies, methods, components, materials, parts, and/or the like. In other instances, well-known structures, materials, or operations are not specifically shown or described in detail to avoid obscuring aspects of embodiments of the present invention. Accordingly, it will be appreciated by one skilled in the art that the embodiments of the present disclosure may be practiced without such specific components. It should be further recognized that those of ordinary skill in the art, with the aid of the Detailed Description set forth herein and taking reference to the accompanying drawings, will be able to make and use one or more embodiments without undue experimentation.
Additionally, terms such as “coupled” and “connected,” along with their derivatives, may be used in the following description, claims, or both. It should be understood that these terms are not necessarily intended as synonyms for each other. “Coupled” may be used to indicate that two or more elements, which may or may not be in direct physical or electrical contact with each other, co-operate or interact with each other. “Connected” may be used to indicate the establishment of communication, i.e., a communicative relationship, between two or more elements that are coupled with each other. Further, in one or more example embodiments set forth herein, generally speaking, an element, component or module may be configured to perform a function if the element may be programmed for performing or otherwise structurally arranged to perform that function.
As used herein, a network element, node or subsystem may be comprised of one or more pieces of service network equipment, including hardware and software that communicatively interconnects other equipment on a network (e.g., other network elements, end stations, etc.), and is adapted to host one or more applications or services, either in a virtualized/non-virtualized environment, with respect to a plurality of subscribers and associated user equipment (UE) nodes that are operative to receive/consume content in a media distribution network where media content assets may be distributed and delivered using stream-based or file-based mechanisms. As such, some network elements may be disposed in a wireless radio network environment whereas other network elements may be disposed in a public packet-switched network infrastructure, including or otherwise involving suitable content delivery network (CDN) infrastructure that may comprise public, private, or mixed CDNs. Further, suitable network elements including one or more embodiments set forth herein may involve terrestrial and/or satellite broadband delivery infrastructures, e.g., a Digital Subscriber Line (DSL) network architecture, a Data Over Cable Service Interface Specification (DOCSIS)-compliant Cable Modem Termination System (CMTS) architecture, switched digital video (SDV) network architecture, a Hybrid Fiber-Coaxial (HFC) network architecture, a suitable satellite access network architecture or a broadband wireless access network architecture over cellular and/or WiFi connectivity. Accordingly, some network elements may comprise “multiple services network elements” that provide support for multiple network-based functions (e.g., 360° immersive A/V media preparation delivery policy management, session control, QoS policy enforcement, bandwidth scheduling management, content provider priority policy management, streaming policy management, and the like), in addition to providing support for multiple application services (e.g., data and multimedia applications including 360° immersive video assets (also referred to as 360-degree video assets or simply 360 video assets) in varying qualities or definitions). Example subscriber end stations or client devices may comprise various devices, tethered or untethered, that may consume or deliver media content assets using streaming and/or file-based downloading technologies, which may involve some type of rate adaptation in certain embodiments. Illustrative client devices or UE devices may therefore include any device configured to execute, inter alia, one or more client applications for receiving, recording, storing, and/or decoding/rendering 360 video content, live media and/or static/on-demand media, which may comprise Virtual Reality (VR) media, Augmented Reality (AR) media, Mixed Reality (MR) media, from one or more content providers, e.g., via a broadband access network, using HTTP, HTTPS, RTP, and the like. Accordingly, such client devices may include Next Generation IP-based STBs, networked TVs, personal/digital video recorders (PVR/DVRs), networked media projectors, portable laptops, netbooks, palm tops, tablets, smartphones, multimedia/video phones, mobile/wireless user equipment, portable media players, portable gaming systems or consoles (such as the WHO, Play Station 3®, etc.) operating in concert with 3D display devices and the like, which may access or consume 360-degree content/services provided via a suitable media distribution network wherein a bandwidth and Quality of Experience (QoE) scheme may be provided in accordance with to one or more embodiments set forth herein.
One or more embodiments of the present patent disclosure may be implemented using different combinations of software, firmware, and/or hardware. Thus, one or more of the techniques shown in the Figures (e.g., flowcharts) may be implemented using code and data stored and executed on one or more electronic devices or nodes (e.g., a subscriber client device or end station, a network element, etc.). Such electronic devices may store and communicate (internally and/or with other electronic devices over a network) code and data using computer-readable media, such as non-transitory computer-readable storage media (e.g., magnetic disks, optical disks, random access memory, read-only memory, flash memory devices, phase-change memory, etc.), transitory computer-readable transmission media (e.g., electrical, optical, acoustical or other form of propagated signals—such as carrier waves, infrared signals, digital signals), etc. In addition, such network elements may typically include a set of one or more processors coupled to one or more other components, such as one or more storage devices (e.g., non-transitory machine-readable storage media) as well as storage database(s), user input/output devices (e.g., a keyboard, a touch screen, a pointing device, and/or a display), and network connections for effectuating signaling and/or bearer media transmission. The coupling of the set of processors and other components may be typically through one or more buses and bridges (also termed as bus controllers), arranged in any known (e.g., symmetric/shared multiprocessing) or heretofore unknown architectures. Thus, the storage device or component of a given electronic device or network element may be configured to store code and/or data for execution on one or more processors of that element, node or electronic device for purposes of implementing one or more techniques of the present disclosure.
Referring now to the drawings and more particularly to
Skilled artisans will recognize that the foregoing generalized example network environment 100 may be implemented in a hierarchical network architecture, with various aspects of media capture and preparation, including, e.g., source stream stitching, projection mapping, source media compression, tiled/ABR encoding/transcoding, packaging, etc., as well as distributing/uploading and edge node processes taking place in different network portions disposed at different hierarchical levels, involving one or more operators, content delivery networks (CDNs), edge networks, and the like. Further, in some implementations, at least some of the foregoing apparatuses and processes may be cloud-based. In some arrangements, a CDN can be a large distributed system of servers deployed in multiple data centers connected to the Internet or other public/private communications network. A CDN can be a managed or unmanaged network, and can also be a federation of managed or unmanaged networks.
An example embodiment of a media server/source system operatively associated within the foregoing example network environment may therefore be configured, e.g., as a global headend, to accept media content from live sources and/or static file sources, e.g., online content providers such as Hulu®, Netflix®, YouTube®, or Amazon® Prime, as well as VOD catalog or content providers or studios such as, e.g., Disney, Warner, Sony, etc. Media content from live sources may comprise live programming captured relative to any type of event, e.g., sporting/entertainment/gaming events, concerts, live TV shows, live news broadcasting sources, such as, for instance, national broadcasters (e.g., NBC, ABC, etc.) as well as cable broadcaster channels like Time Warner channels of CNN, ESPN, CNBC, etc., and local broadcasters, etc., including any secondary media insertions such as advertisement media channels.
Without limitation, an example network architecture 200 (which may form a portion of the environment shown in
In one arrangement, tiled encoder 210 may be configured to generate tiled-encoded bitstreams as a plurality of phase-encoded streams for each bitrate representation of the media input stream 202, wherein each phase-encoded stream for a particular bitrate representation is provided with a specialized frame at a particular location in the Group-of-Pictures (GOP) structure of the stream depending on the phase as will be set forth in additional detail further below. This scheme of encoding may be referred to as Phased Encoding (PE) scheme with respect to certain embodiments of present invention. In another arrangement, tiled encoder 210 may be configured to generate a pair of tiled-encoded bitstreams, e.g., a first and a second tile-encoded bitstream, for each bitrate representation of the media input stream 202, wherein a first encoded bitstream may comprise a regular or standard tile-coded bitstream generated according to a known or heretofore unknown coding scheme and a second encoded bitstream may be coded such that a specialized frame is provided at each location in a GOP structure, as will be set forth in additional further below. This scheme of encoding may be referred to as Block-Intra Encoding (BIE) or All-Intra Encoding (AIE) scheme with respect to certain embodiments of the present invention.
Regardless of whether PE-coding scheme or BIE-coding scheme is used, a packager 214 is operative to package the tile-encoded bitstreams 212 and generate suitable manifest files describing characteristics of tile groupings per frame for each tile-encoded bitstream, e.g., tile location, slice header information, various types of metadata including picture timing, color space information, video parametric information, etc., which may be stored at a suitable packaged media storage facility 240, along with suitable stream manifests 241. A network edge node 216 including a video optimization system 215 comprising a plurality of modules or subsystems is operative in association with a video back office system 238 for effectuating a 360° immersive video session with a premises device 236 of subscriber premises 234 that is served by a managed bandwidth pipe 232 effectuated via a suitable access network (e.g., a DSL/DOCSIS network portion having suitable infrastructure that may include, e.g., routers, DSLAM/CMTS elements, etc., or suitable 3G/4G/5G radio access network elements, including fixed wireless infrastructure in certain implementations, and the like), generally represented by node or element 230.
In one arrangement, video optimization system 215 may comprise a tile selection subsystem 218 that is operative responsive to bandwidth annealing and QoE management policies, as well as user gaze vector information, inter alia, to provide tiles 220 selected from different video quality bitstreams to a tile combining and stream generation subsystem 222. Multiplexed video frames with tiles from different bitstreams 224 may be provided to a delivery service 226 for facilitating the transmission of muxed tile stream 228 to the downstream infrastructure 230. Broadly, when a user request 250 for a 360° immersive video session is generated, it is processed by the video back office system 238 and forwarded to the video optimization system 215 via a message 252 for obtaining a session ID and associated location information for the requested 360° media. Responsive to a response message 251 from the video optimization system 215, the video back office system 238 is operative to provide a response 248 including appropriate URL information for the media and a session ID to the requesting device 236. User gaze information (which may be a default setting initially) and associated session ID information may be provided to the infrastructure element 230 as a message 246, which may be propagated to the video optimization system 215 as message 254. Also, the infrastructure element 230 is operative to provide a dynamic bandwidth allocation message 254 that includes the session ID information to the video optimization system 215 in a related or separate process. As noted previously, tile selection subsystem 218 may be configured to operate in response to control messages relative to bandwidth allocation, user gaze vector information, or both, for selecting tiles having different video qualities, which may be combined or stitched into frames in order to generate a muxed tile-encoded video output stream. In one arrangement, the tile combining and stream generation subsystem 222 may be provided as part of the video optimization system 215 during video stream delivery. In another arrangement, the tile stitching may be effectuated during playout on the client side (e.g., at the client device 236 or some other premises equipment associated therewith) rather than on the server side. In this arrangement, a client-side stitching functionality is operative to receive the selected tiles and perform the necessary stitching in order to generate a stitched stream to be decoded and rendered. Various embodiments relative to the foregoing processes, subsystems and components will be set forth in further detail in the following sections.
For purposes of example embodiments herein, a GOP structure is a group of successive pictures in a coded video stream, which specifies the order in which intra- and inter-frames are arranged. Each coded video stream comprises successive GOPs, from which the visible frames may be generated. Generally, a GOP structure may contain the following picture types: (1) I-picture or I-frame (intra coded picture)—a picture that is coded independently of all other pictures. Each GOP begins (in decoding order) with this type of picture. (2) P-picture or P-frame (predictive coded picture)—contains motion-compensated difference information relative to previously decoded pictures. In older designs such as MEPG-1, H.262/MPEG-2 and H.263, each P picture can only reference one picture, and that picture must precede the P picture in display order as well as in decoding order and must be an I or P picture. These constraints do not apply in the newer standards such as, e.g., H.264/MPEG-4 AVC, H.265/HEVC, etc. (3) B-picture or B-frame (bi-predictive coded picture or bidirectionally predictive coded picture)—which contains difference information from the preceding and following I- or P-frame within a GOP, and contains motion-compensated difference information relative to previously decoded pictures. In older designs such as MPEG-1 and H.262/MPEG-2, each B-picture can only reference two pictures, the one which precedes the B picture in display order and the one which follows, and all referenced pictures must be I or P pictures. These constraints do not apply in the newer standards such as, e.g., H.264/MPEG-4 AVC, H.265/HEVC, etc. (4) D-picture or D-frame (DC direct coded picture)—serves as a fast-access representation of a picture for loss robustness or fast-forward in certain types of video (e.g., MPEG-1 video).
In general, an I-frame indicates the beginning of a GOP. Afterwards several P and B frames may follow. The I-frames contain the full image and do not require any additional information to reconstruct it. Typically, encoders use GOP structures that cause each I-frame to be a “clean random access point,” such that decoding can start cleanly on an I-frame and any errors within the GOP structure are corrected after processing a correct I-frame. The GOP structure is often referred by two numbers, for example, M=3, N=12. The first number tells the distance between two anchor frames (I or P). The second one tells the distance between two full images (I-frames), which is the GOP size. For the example M=3, N=12, the GOP structure is {IBBPBBPBBPBBI}. Instead of the M parameter the maximal count of B-frames between two consecutive anchor frames can be used. For example, in a sequence with pattern {IBBBBPBBBBPBBBBI}, the GOP size is equal to 15 (length between two I frames) and distance between two anchor frames (M value) is 5 (length between I and P frames or length between two consecutive P Frames).
While a typical GOP starts with an I-frame, some embodiments herein provide a structure where a GOP may commence with an X-frame instead, in addition to placing the X-frames at specific locations or replacing the P- and/or B-frames in the GOP structure as will be set forth in additional detail further below.
Skilled artisans will recognize that depending on codec implementation, a picture or frame may be partitioned into a number of ways at different levels of granularity, for example, to facilitate, inter alia, coding efficiency, parallel processing, etc. In one arrangement, a frame may be partitioned into a number of coding tree units (CTUs), each containing certain number of luma coding tree blocks (CTBs) and chroma CTBs, which in turn may comprise multiple coding blocks (CBs). A frame may be split into one or more slices, each being a spatially distinct region of a frame that may be encoded separately from any other region in the same frame and identified with a slice header. In general, slices are self-contained and contain a sequence of CTUs that are processed in the order of a raster scan, wherein slices can be coded as I-slices, P-slices, or B-slices similar to I-frames, P-frames, or B-frames, respectively. In one arrangement, slices may be used to effectuate resynchronization to minimize data losses, and may contain a varying number of CTUs per slice depending on the activity in a video scene.
In addition to slices, an encoding scheme may also define a number of tiles per frame, which may also be configured to be self-contained and independently decodable rectangular or square regions of a picture, based on vertical and horizontal partitioning to form a grid, in order to facilitate parallel processing at the encode and decode stages. In one variant, the self-contained and independently decodable tiles may use temporal prediction from the co-located tiles of previously encoded pictures or frames. Multiple tiles may share header information by being contained in the same slice, where a tile may comprise a certain number of CTUs. It is not a requirement that each tile include the same number of CTUs. Accordingly, in one arrangement, the tiles of a frame may have different sizes. If a frame contains a single slice, the tiles of the frame will therefore have the same slice header and picture header information. In another arrangement, a frame may include one or more slices, each slice containing one or more tiles, and each tile in turn containing one or more CTUs.
For purposes of the present patent application, because video frames may be partitioned in numerous ways and at different levels, the terms “coding tree unit”, “coding tree block”, “coding unit”, “macro block”, or “block” or terms of similar import will be generally treated as an abstract unit of coding that may applied with respect to a tile, slice and/or frame without limitation to any particular video compression standard or technology.
Returning to
If BIE (also referred to as All-Intra Encoding, as noted elsewhere in the present patent application) is selected, the video source stream 602 may be encoded/transcoded into a plurality of streams with varying qualities and/or bitrates (block 616). In one example embodiment, each of the streams may be tile-encoded using a standard coding scheme (e.g., HEVC, AV1, etc.) to generate normal or regular tile-encoded streams 618. Similar to the discussion above with respect to the phased-tiled streams 610, reference numeral 622-1 refers by way of illustration to quality information relating to a regular tile-encoded stream 620-1 having a QP setting of 30 and/or a bitrate of around 7.0 Mbits/s, which may be indicative of a lower end of quality. Likewise, reference numeral 622-N refers to quality information relating to a regular tile-encoded steam 620-N having a QP setting value of 16 and/or a bitrate of around 105.6 Mbits/s, which may be indicative of a higher quality stream.
Additionally, the video source stream 602 is also encoded/transcoded into a plurality of streams with corresponding qualities and/or bitrates (block 617) where each stream is tile-encoded such that all frames of it's GOP structure are provided as X frames. By way of illustration, reference numeral 632 refers to a plurality of BIE-coded and tiled streams, wherein quality information 636-1 having a QP setting of 30 and/or a bitrate of around 7.0 Mbits/s (also sometimes abbreviated as Mbs or Mb/s) relates to a lower quality BIE-coded tiled stream 634-1 while quality information 636-N of QP setting of 16 and/or a bitrate of around 105.6 Mbits/s relates to a higher quality BIE-coded tiled stream 634-N.
Skilled artisans will recognize upon reference hereto that when an encoder is configured with a target QP, the bitrate of an encoded bitstream is somewhat averaged over the course of the bitstream. For instance, if a QP of 10 is targeted in a source encoding scheme, it is possible that a low bitrate may be seen in areas of no motion (e.g., resulting in 4 Mbs). In areas of high motion, it is possible that the bitrate could shoot up to 200 Mbs. Thus, in an example encoding scheme that targets specific QPs as set forth in the foregoing, the bitrates of the output steams could be variable over a range. Accordingly, it should be appreciated that the bitrates shown in association with the QPs of PE or BIE streams in
Accordingly, it should be understood by the reader that although certain examples and portions of the description within this disclosure are provided assuming the use of a fixed quantization (QP) value per stream, streams in practice may contain QP values that vary between pictures and within a picture as noted above. An encoder according to an embodiment of the present invention may control its output bitrate by the means of a rate-control or the like, and thereby change the QP value between pictures. An encoder may also encode pictures within one stream using varying QP values to optimize the visual quality of the stream. Within one picture, the QP value may change between blocks using e.g., adaptive quantization mechanisms to optimize the visual quality as known in the art. The use of “QP” in phrases within this disclosure such as e.g., but not limited to, “encoded with that QP”, “video of different QP values”, “generated videos with different QP values”, “stream having a QP value of N”, “QP value of the video stream” should be understood as a way of characterizing streams such that a stream associated with a lower QP value is of higher bitrate and higher quality than one associated with a higher QP value, and not that the QP is kept static for each block in a stream.
It should be further appreciated that adaptive bitrate encoding and tile encoding of media assets may be integrated within an apparatus as part of a content preparation system in one example embodiment such that various types of encoding and/or transcoding may take place in different sequences and/or in parallel processes. Further, additional functionalities such as projection-mapping, source stream stitching, packaging, etc., may also be combined or otherwise integrated with the tile-coding/transcoding schemes of the present patent application depending on implementation.
Skilled artisans will recognize that at least a portion of the foregoing steps, acts or operations may comprise media preparation and (pre)provisioning with respect to one or more 360° immersive video assets distributed in a network environment or architecture illustrated in
In general, an embodiment of a BIE configuration method may be configured to receive or obtain as input a source video stream for 360° immersive video and a list of output video qualities (e.g., a list of QP values, such as {QP1=16, QP2=18, QP3=20, QP4=22, QP5=24, QP6=26, QP7=28, QP8=30, or other indicia based on targeted bitrates}). Accordingly, without limitation, for every output video quality (e.g., every QP value), two video streams may be encoded—a regular/standard HEVC video with that QP or quality, and a Block-Intra HEVC video with that QP/quality—as noted previously. In order to be able at a later time (e.g., shortly before decoding) to stitch tiles from different qualities into the same bitstream, the encoding phase of an embodiment provides that all the video streams have the same base_qp (defined below), while the actual difference between the videos of different QP values may be effectuated by means of qp_delta (defined below) from the base QP. For example, a setting of base_qp=22 may be configured, wherein the parametric values base_qp=22 and qp_delta=−6 may be used to achieve QP=16. In general, these two parameters relate to setting the quality (QP) of a video stream. Recall that all the generated videos with the different qp values need to have the same base_qp, while different QP values may be achieved by using qp_delta from the base_qp. This requirement may be imposed based on one particular time instance. That is, if pictures in a bitstream are numbered, then any two pictures from two bitstreams that are used as input for stitching with the same numbers must use the same base_qp value in one arrangement. For purposes of the present invention, “base_qp” may be described as follows: the ith frame (for every i=1 to N, where N is the total number of frames in a video sequence) in all the encoded versions or bitrate representations of the same video will have the same slice QP value. In other words, slice QP is the base_qp. Although slice QP may be set as the same value in all the generated streams, it can vary over time. For purposes of the present invention, the parameter delta_qp may be described as follows: by assigning a given qp_delta, the first block in each tile that signals QP is configured to signal the delta_qp (that amount of variance from the base QP). It may be noted that there could be a deblocking mismatch after stitching in some embodiments.
Another parameter that may be defined for purposes of the present invention is ROI (Region of Interest), which determines an area of a frame where the tiles can be independently encoded so that the subset of the bitstream corresponding to the ROI can be easily extracted and reconstituted into another bitstream. As noted above, in order to later stitch videos of different QPs, it is desirable to utilize the functionality of base_qp and delta_qp. This is supported for example when using HEVC ROI encoding functionality in one illustrative implementation. Accordingly, when encoding with ROI in an embodiment, the base_qp parameter for the slice QP headers may be defined, in addition to defining an ROI grid (independently defined from the grid/array of the tiles of a frame) such that the area of the grid in the ith row and jth column in the ROI grid gets its own delta_qp. Generally, this allows an embodiment to assign different delta_qp to different areas of the ROI grid, whereby selective delta_qp values may be used for purposes of the present invention. For example, to achieve a given desired QP (say QP=16), the base_qp may be defined (say base_qp=22) using the regular qp parameter, and then by using the ROI grid, all the targeted areas may be assigned a delta_qp of −6, thus effectively achieving a QP of 16 for all the tiles in the ROI grid.
In one embodiment, the content at different qualities may be encoded using the same base_qp (slice QP) for a particular frame. For each quality of that frame, a specific desired QP may be set, wherein the delta_qp syntax elements may be used so that all blocks (or alternatively, as many blocks as possible or desired) of that frame are encoded with that desired QP. Additional aspects of a BIE configuration scheme based on HEVC may be set forth as follows.
The encoder may be set to use tiled encoding. During setup, this may be effectuated by setting an appropriate flag for tiled-encoding, as well as configuring a specific grid structure of the tile (e.g., as shown in
The encoder may be configured to disable temporal motion vectors prediction. Although an example BIE scheme does not use MVs (motion vectors), temporal motion vector prediction (TMVP) settings may need to be identical across streams to enable stitching at a later time. This configuration is optional, in that an embodiment of BIE may be practiced without disabling TMVP.
Also, many other elements of the slice headers may be configured to be identical across streams. For example, elements such as the number of reference pictures to use, the reference picture set, what reference pictures to use for L0, the Picture Parameter Set (PPS) to use, the picture order count, SAO parameters, etc. Further, it is also required that the decoding order is the same for all bitstreams that are to be used as input for bitstream switching. Skilled artisans will recognize upon reference hereto that a variety of slice header elements may be configured accordingly in an example BIE implementation.
Since a slice uses a single PPS id codeword to identify what PPS to use and the PPS references one single Sequence Parameter Set (SPS), all encodings may be done using identical PPS and SPS id values in an example embodiment. Likewise, many syntax elements in the SPSs and PPSs may also be configured to be identical for the multiple encodings. Although not a necessary requirement, an example BIE embodiment may therefore be configured such that the encodings are effectuated using identical SPSs and PPSs. However, it is strictly necessary that some elements in the SPS and PPS are identical in certain arrangements.
Returning to
It should be noted that whereas the foregoing BIE configuration process 800A uses certain parameters, additional or alternative embodiments may be practiced where a BIE scheme may be configured to utilize other parameters in addition to and/or in lieu of the parameters exemplified in the flowchart of
In a further embodiment of the present invention, X-frames may be used once in each GOP (instead of multiple times as in BIE) based on a PE scheme as noted previously. Essentially, PE-based tiled encoding involves a process and apparatus for generating a stream where all the frames have P-slice headers, except for the first frame which is an I-frame, while periodically there is an X-frame (i.e., BIE-frame or AIE-frame), where all blocks are intra-encoded but the slice headers are of P-slices (or B-slices where B-frames are also encoded in a sequence). In general, all slices of any two pictures that are to be potentially used as inputs to stitching need to have the same slice type, slice qp, as well as a number of other settings in the slice header and PPS. In contrast with the BIE scheme set forth above, where all the frames of a GOP are X-frames except for the first one, an embodiment of a PE scheme is configured to provide X-frames only at select frame locations depending on two parameters: period (which is the size of the GOP, i.e., the number of frames in the GOP) and phase (which is an integer in the range {0 to [period-1]}. Frame locations where the X-frames appear in a PE scheme may be determined as follows. Let N be the total number of frames in a stream. The first frame is encoded as an I-frame. For a frame at ith position, 2≤i≤N, it is encoded as a regular P-frame if {i Mod (period)≠phase}; and otherwise (that is, {i Mod (period)=phase}), the frame is encoded as an X-frame (with P-slice headers and all blocks encoded in intra-mode, independent of previous frames). It should be noted that an example PE scheme may provide as many phase-encoded streams for each quality/bitrate representation of the media input as there are frame locations in a GOP (i.e., GOP size).
By using P- or B-slice headers rather than I-slice headers in X-frames for purposes of the present invention, several advantages may be realized in an exemplary embodiment, including but not limited to facilitating mid-GOP switching in a user viewing environment. Assume the user is watching a 360° immersive video program or content where the directly-gazed field of view (FoV) is in high quality (i.e., lower QP) and the user moves his head in the middle of the GOP. The user now sees a low quality video (higher QP) in their new field of view or viewport. The server can send an I-frame of a high quality (low QP) at the beginning of the next GOP, but this introduces a significant latency, as it would take time until the high quality I-frame of the next GOP for the viewport will be presented. It is desirable to receive or obtain an I-frame of the new field of view that is encoded at high quality as soon as possible while in the middle of the GOP. But it is not feasible to just put an I-frame as is in the middle of the GOP in a conventional immersive video viewing environment. By generating an X-frame (i.e., Block-Intra coded frame or All-Intra coded frame) and transmitting it in the middle of the GOP (at any frame location in a GOP structure, for instance), an embodiment of the present invention is thus effectively able to upgrade the quality of the field of view similar to the effect if an I-frame is presented in the middle of the GOP with high quality tiles. By providing P-slice headers in AI- or BI-encoded frames (i.e., AIE/BIE frames or X-frames), an embodiment of the present invention therefore allows a frame having high quality data in a region of interest (ROI) of FoV to be used in the middle of the GOP.
Further, in a tile encoding scheme where a frame is partitioned into tiles and slices, an embodiment of the present invention involving X-frames enables mixing tiles in a single output compressed frame, where some tiles may use spatial or temporal prediction (i.e., inter-picture prediction) and some tiles may use only spatial prediction (e.g., comprising intra-coded blocks only). The tiles consisting of intra-coded blocks only may originate from an X-frame. In the context of the present patent application, the terms “mixing”, “muxing”, “stitching”, “splicing” or terms of similar import with respect to output stream generation may refer to means and methods to concatenate one compressed tile (e.g., tile A) with another compressed tile (e.g., tile B) to form a part of the bitstream representing a single output frame, where tile A and tile B may originate from separate encodings of the content, which will be set forth in additional detail further below.
One of the advantages of a PE scheme relates to overcoming the issue of drift that may be present in a BIE scheme (i.e., drift elimination or reduction). It should be appreciated that while BIE allows replacement a P-frame of the previous viewport with an X-frame of the new viewport, the following frames are regular P-frames of the new viewport that are encoded with predictions made to previous frames. Thus, when a P-frame is replaced with an X-frame and then a following frame uses this X-frame for prediction instead of the original frames of the regular bitstream, there is a potential for drift where prediction errors may accumulate. On the other hand, in phased encoding, the generated stream uses the X-frame at position=<phase>+i*<period> for the prediction of the following P-frames, and thus the situation where a P-frame uses for prediction a different frame than the frame used during encoding is avoided. Hence, there is no prediction error presented due to predicting from a frame that is different than the frame generated during the encoding, and accordingly, any potential drift due to this type of prediction errors is avoided. However, the PE scheme may require a larger amount of storage since storage of the P-frames that follow the X-frames in the GOP is required.
Further, an embodiment of a PE scheme may be advantageously utilized to facilitate gradual refreshing of frames whereby lower latency is achieved during playout by selecting only a subset of the tiles to upgrade their quality and send their appropriate phased-encoded tiles. While in an embodiment of a BIE scheme, a P-frame is replaced with an X-frame, in a gradual refresh frame annealing scheme the PE-coded streams may be used to replace selected tiles with the corresponding tiles taken from the appropriate PE-coded stream. On the other hand, in another embodiment, a BIE scheme may can also advantageously operate on a tile-by-tile basis. With respect to a PE-based embodiment, accordingly, if period is P and frame number is X, one can obtain the corresponding phase by the following equation: Phase={X Mod P}. Thus, during delivery or playout of a coded video sequence, assume that a certain tile T is selected for upgrade to QP quality q in frame X, then one can replace selected tile (in frame X and following frames until the next upgrade/downgrade of T or viewport change) with the tile T from the stream with phase that satisfies the following relationship: Phase={X Mod P} at QP=q. Thereafter, the co-located tiles in the frames following frame X that belong to the same GOP are replaced by the corresponding co-located tiles from the same PE-encoded stream. It should be appreciated that the advantages of concatenating tiles from different streams when a user changes gaze direction are similar to the scenario set forth above where the user changes his gaze during mid-GOP. Identical slice QPs are used for switching/replacing the tiles because if two input tiles are encoded with different actual QP and were encoded with a single slice per picture, then if the slice QP differs, it would not be possible for the QPs of tiles in the output stream to be correct without low-level rewrite of the stream. Additional details regarding gradual refresh frame annealing and tile selection will be set forth further below in reference to additional embodiments of the present patent application.
A potential disadvantage with respect to PE may be that it requires more storage, since an input stream is encoded in many phases, potentially resulting in as many streams as the GOP size (rather than just two streams as in BIE). This disadvantage may be traded off against the advantage of reduced latency without drift in an example implementation. For fastest quality change response, the number of phases may be set equal to the size of the GOP, i.e., the period P, but an example embodiment may provide a trade-off of using fewer phases and consuming less storage while the latency of the quality upgrade may be longer, since tile upgrades will only be done on the next phase.
In an additional or alternative embodiment of a PE scheme, a phase-encoded bitstream may have a frame other than an I-frame as a first frame of the coded video sequence, which may be achieved by appropriate settings in an encoder in accordance with the teachings herein. For example, the first frame can be an X-frame (or some other non-I frame). All other frames of the coded sequence may contain predicted frames (P/B-frames) and X-frames at suitable locations based on phase.
Turning to
It should be noted that an example PE configuration process is roughly similar to a BIE configuration process in one embodiment, which may be performed for every “phased” stream depending on the GOP size. Further, analogous to a BIE configuration process 800A that uses certain parameters, additional or alternative embodiments of a PE configuration process may involve other parameters in addition to and/or in lieu of the parameters exemplified in the flowchart of
As noted previously, a PE-based tiled encoding scheme facilitates a gradual refresh annealing process during 360° video delivery, which will be set forth in further detail below. An embodiment of phased encoding may also be used during the playout where a stitcher executing on the server side or on the client side may be used to combine tiles of different qualities. So, at every frame of the video being played, each tile has a current quality, which may correspond to the QP value, target bitrate or other indicia of the video stream the tile was taken from. When bandwidth is sufficiently large or when the user moves his head and the viewport changes, it would be desirable to upgrade the quality (e.g., lower the QP) of some tiles (the tiles on the new viewport for example). Furthermore, to reduce latency by means of reducing the usage of the buffer on the decoder side, an embodiment of the present invention provides that the entire viewport may not be upgraded at once, but rather upgrade it gradually by means of gradual refresh, only upgrading a few tiles in every frame, keeping the decoder buffer small and thus reducing latency. As will be described in additional detail further below, an example bandwidth annealing apparatus may be configured to effectuate a process for determining which tile to upgrade at every moment based on the bandwidth, the viewport and/or current buffer utilization. Further, such a process may also be configured to determine a quality level (i.e., which QP) to which a tile should be upgraded.
For example, assume that during playout, a tile selection apparatus (described in detail further below) determines to upgrade in the ith frame, tile T to quality QP=q. This determination may be provided as a control input to a tile/frame stitcher module, which retrieves, receives or otherwise obtains tile T from the ith frame of the video stream that was encoded with quality QP=base_qp+delta_qp=q using phased encoding, where the phase is determined by the modular relationship: {phase=i Mod (period)}. Then, until the next time the tile selection process decides to change the quality of this tile, tile T is taken from the same stream (i.e., the phased encoded stream with quality QP=q and with the same phase). Accordingly, it will be appreciated that an additional advantage of the PE scheme beyond the ability to perform a gradual refresh of tiles during upgrades is better video quality. Overall, phased encoding gives a better QoE than a BIE scheme where X-frames are substituted without phases, which can result in drift, and result in lower peak signal-to-noise (PSNR) values, thereby resulting in a lower QoE stream for the remainder of the GOP. As noted previously, a potential drawback of phased encoding is the requirement of multiple streams that can result in significant encode processing overhead and storage space.
Example embodiments regarding how to stitch tile-encoded bitstreams using either PE or BIE schemes will be set forth below. As noted previously, tile-stitching embodiments may be implemented at a server during stream delivery phase or on the client side for playout. In general, example embodiments for stitching tiles involve utilizing bitstreams of different qualities (e.g., based on different QPs, targeted bitrates, or other indicia) as well as ensuring that there is compatibility with respect to various pieces of parametric data relating to video pictures, e.g., Video Parameter Set (VPS), Sequence Parameter Set (SPS), Picture Parameter Set (PPS), Supplemental Enhancement Information (SEI), etc., among the bitstreams from which the tiles may be selected. In general, the tile structure should preferably be constant over time to facilitate stitching, which in turn is related to tile-encoding processes performed by an encoder of the present invention. A bitstream stitcher module is operative in response to an input comprising a list of tiles from different tile-encoded streams, which may be combined to generate a new output bitstream, where tiles closer to the viewport have a higher quality compared with tiles further away from the viewport. Further, an example embodiment to perform the tile combination and stream muxing in accordance with the teachings of the present invention may be configured such that output stream generation still remains compliant within known codec standards such as MPEG HEVC/ITU-T/ISO23008 part 2/H.265 specifications as well as emerging standards such as AV1, H.266, VVC, and the like.
For stitching BIE-coded streams, tiles from the regular streams may be used by default for splicing (e.g., until some control input is provided based on user's gaze or bandwidth allocation). The only instances where tiles from the BIE-coded streams are taken is when either the viewport changed (thereby requiring the X-frames which are frames with P-slice headers that can fit in the middle of the GOP but the tiles are intra encoded so the new viewport can be presented) or when a bandwidth annealing process determines to upgrade the quality of the tile (in which case the Block-Intra frame with the P-slice headers contains the tile with the upgraded higher quality).
By way of illustration, consider a block-intra stream stitching scenario in which there are at least three streams: (1) a regular stream of lower quality (e.g. QP setting of 30); (2) a regular stream of higher quality (e.g. QP setting of 22); and (3) a BIE (all-Intra) stream of higher quality. Broadly, when the viewport changes, the quality of some tiles may be increased. That is done in block 1312, which means that, e.g., a tile at position A that in previous pictures was taken from stream (1) is now taken from stream (3). In the next picture, the tile at position A should be taken from stream (2) if the tile is still within the viewport. If the tile is no longer within the viewport, the position A tile could be taken from stream (1). More particularly, it may be further dependent upon gaze vector information. In other words, it is not just if the tile at position A is in the viewport or not; rather, it is where the tile is located in a gaze-to-weight determination scheme used for tile selection (described in detail further below). Thus, it should be understood that tiles within the viewport depending on where they are located may be upgraded or downgraded based on how far the tile are from the direct line of sight in an example embodiment of the present invention.
In similar fashion, an example tile stitching scheme 1300B involving PE-based tiled streams is illustrated in
Regardless of whether tiles from BIE-coded bitstreams or PE-coded bitstreams are stitched, an example embodiment of stitching may involve taking tiles from different streams having compatible slice headers in addition to other parametric information as set forth previously. In general, slice type (i.e., I/P/B-slice), the slice QP and other fields or parameters that may affect the CABAC decoding process may be monitored to ensure compatibility and compliance. Further, some embodiments, such as example embodiments set forth in
Turning to
Skilled artisans will recognize that in order to splice tiles they need to be retrieved from specific source bitstreams responsive to a tile selection process. To facilitate efficient retrieval, an embodiment of splicing may involve providing a memory-mapped tile pointer cache that allows a quicker referencing of parsed files corresponding to tiles, wherein a file format is optimized to be memory mapped instead of being parsed into RAM. Set forth below is an example file format for purposes of an exemplary splicing embodiment:
file format:
u(16) magiclen;
s(magiclen) magic_string;
loop {// each iteration is a group
u(32) nrec;
for (i=0; i<nrec; i++) {
}
// rec_end_relative[i] is relative to the offset of the file represented by this comment.
// rec_end_abs[i]=this_offset+rec_end_relative[i]
for (i=0; i<nrec; i++) {
// this iteration is per frame
// prefixes are NAL units that appear before the Slice in the access unit
}
}
magic_string is “TPTCACHE1 [ii[ ][il[ ]llil[ ]il[ ]]]”
Referring to
To facilitate gaze-based tile selection control, additional embodiments of the present invention involve monitoring where a user is viewing in a 360° immersive video program (i.e., the user's viewport) and determining appropriate tile weights based on the user's gaze. In general, a gaze vector (GV) may be returned by the user/client device defining a gaze direction in a 3D immersive space displaying 360° video, e.g., where the headset is pointed. In further embodiments, the user's eyeball movement may be tracked for similar purposes. As will be seen below, the tiles of a tiled frame also have direction vectors (which are not dependent on the user's gaze) based on how the frame is mapped in a 3D display environment. A dot product (also referred to as a scalar product or inner product) of tile vector and gaze vector can be calculated to determine the angular separation between the gaze direction and the direction of the middle of any tile of a frame, which may be provided to a weighting function module for determining corresponding tile weights.
Process 1500B sets forth additional details with respect to effectuating gaze-based control in an example embodiment. At block 1522, tile weights may be determined as a function of a cosine of an angular separation between the gaze vector and the directional vector corresponding to a tile location in a suitable 3D spatial projection of a 2D video frame of the 360° immersive video asset. At block 1524, the tile weights may be provided as an input along with a dynamic bandwidth allocation input to a tile selection and bandwidth annealing process, which is further described elsewhere in the present patent application.
In one example embodiment, depending on where the tile is located in relation to the gaze vector, a determination is made how much bandwidth is allocated to that tile location corresponding to the weight. Where the gaze vector a and tile directional vector b are denoted by vectors, their dot product may be determined as follows:
{right arrow over (a)}·{right arrow over (b)}=|a|·|b|cos θ
Upon normalization, i.e., if [â]=gaze/|gaze|, then |a|=1. Likewise, by assigning [{circumflex over (b)}]=tile_direction|tile_direction|, |b|=1. Accordingly, by normalizing, the foregoing relationship simplifies to:
â·{circumflex over (b)}=cos θ
Rather than mapping cos(θ) back to θ to determine a weight, an embodiment of the present invention involves defining a mathematical function to map from cos(θ) to a weight as follows: x=cos(θ), and f(x)={x+1} if x≤0 and f(x)=[α{x+1}] if x<0, where α=a scaling factor, e.g., 0.1. Thus, if the angular separation between a gaze vector and tile directional vector is 0°, cos(θ)=1 and f(x)=2. Likewise, for a tile that is 60° or 300° away from the gaze vector, cos(θ)=0.5 and the corresponding f(x) value is 1.5. In an equirectangular projection of a 3D frame, the angle exactly opposite to where the user is looking is 180°, which yields cos(θ)=−1.0, thereby obtaining a weight f(x) value of 0 regardless of the scaling factor. Accordingly, an example embodiment may provide a suitable scaling factor based on how smoothly or quickly tile qualities may vary in relation to a gaze direction within a frame.
In an example embodiment, instead of actual angular displacements, cosine values corresponding to different locations may be provided in reference to the gaze direction. For instance, if a tile direction vector is 90° or 270° from the gaze vector, x=0.0 may be fed to the weighting function, which yields a weight of 1.0. Likewise, for a tile direction vector is 330° away, x=0.866 is provided to the weighting function, thereby yielding a weight value of 1.866. As a further example, if the tile direction vector is 120° deg away, x=−0.5, is provided to the weighting function, thereby yielding a weight value of 0.05 (assuming α=0.1), which is the same if the tile direction were 240° away from the gaze vector).
Further, both gaze vector information and tile direction vector information may be converted to appropriate tile coordinate information relative to the tile grid used in tile encoding during media preparation for facilitating identification of tiles by rows and columns, which may be input along with the weight information to a tile selection and bandwidth annealing process. One skilled in the art will recognize that the determination of tile coordinate information is dependent on the projection mapping used in an example embodiment.
θ={[tx+½]/nx}2π
φ=└½−{ty+½}/ny┘π
y=sin φ
r=cos φ
z=r*cos θ
x=r*sin θ
Where the encoding has non-uniform tile sizes, the foregoing equations may be modified based on, e.g., pixel areas of individual tiles, etc. By way of illustration, using (i) as the tile index for the left edge of tile column i, (j) as the tile index for the top edge of the tile row j, w is the number of pixel columns, and h is the number of pixel rows, an embodiment of the present invention may be configured to determine the following wherein both xi and yj involve a “floor” operator to round out (i.e., the fractional part is removed) with respect to using an example coding unit or block size (e.g., 64 pixels):
By way of illustration, CPE 1646 may be embodied as a platform 1648 including one or more processors 1656, volatile and nonvolatile/persistent memory 1654, input/output (I/O) interfaces 1660 (e.g., touch screens, gaming controllers, hand-tracking gloves, etc.), as well as one or more 360-degree media/gaming applications 1638 that can effectuate a 3D virtual viewing space or “screen” 1620 for the user 1628 wearing head-mounted display (HMD) 1636. In one example arrangement, HMD 1636 may be wirelessly coupled to CPE 1646 via wireless interface 1642. A plurality of decoder buffers 1645 may be provided as part of an example CPE platform 1646/1648 corresponding to one or more 360° immersive video content channels available to the user 1628.
Additional 3D-media-capable CPE 1634 (e.g., a tablet, phablet or smartphone, etc.) may also be separately or optionally provided. Example CPE apparatus 1646/1634 operating together or separately in conjunction with HMD 1636 may be operative to effectuate 3D virtual viewing space 1620 that is an immersive environment in which the user 1628 can move her point of view in full 360° in one of a vertical plane, a horizontal plane, or both planes, defined in the 3D environment, wherein the viewport 1624 changes accordingly. In an additional or alternative arrangement, CPE apparatus 1646/1634 operating in conjunction with HMD 1636 may be operative to effectuate a 3D virtual viewing space 1620 that may be partially immersive in that it is less than 360° along any one of the axes.
A movement and gaze detection module 1662 is operative to detect a movement in a point of view or gaze direction of the user/subscriber 1628 with respect to the 3D virtual viewing space 1620 and provide a suitable gaze vector output to a serving node as the subscriber 1628 shifts her gaze within the viewing space 1620. In one embodiment, a tile weighting module may be configured to operate at a 360° video optimization node (e.g., node 216 in
Set forth below is a list of gaze vectors provided by a client device in an example implementation over a configurable time window:
In a non-normalized format, example GVs in a Cartesian coordinate system may comprise (x,y,z) values such as [3,5,1]; [10,4,1], etc. In a normalized spherical coordinate system, the GV values may comprise sets of angles such as, e.g., (59.04°,80.27°), where r=radius has been normalized out, θ=polar inclination and φ=azimuth angle. Regardless of the format, whereas the gaze vector information may be provided or otherwise obtained at configurable frequencies, time periods, etc., not all gaze vectors may need to be utilized in a tile weight determination process. For example, tile weights may be determined and utilized only in response to triggering a tile selection and bandwidth annealing process, as noted previously with respect to certain embodiments. Accordingly, unused gaze vector information may be periodically discarded in such embodiments.
Continuing to refer to
In an example embodiment, the bandwidth allocation for an exemplary 360° immersive video session may be 19 Mb/s. The video may be encoded with full 360 video using a 128-tile grid, covering bitrates varying from a high of 105.6 Mb/s with a QP value of 16 to a low of 7 Mb/s with a QP value of 30. The higher quality tiles are targeted at the user's direct field of vision. The quality of tiles degrades (i.e., QP values rise) in proportion to the distance from the user's direct field of vision. The functionality of BWA-TS insures that the overall bandwidth of the 360 video session is not exceeded. The tile selection is based on the bitrate of each tile. In an example when the user is looking up at a cloudy sky in a scene, most of the tiles provided in that viewport are relatively high quality. The content of the tiles when looking up in such a scenario is relatively static (i.e., very little motion) and therefore not as many bits are dedicated to the low motion areas by the encoder. This results in the ability to show tiles from the highest quality video encoding with a QP value of 16. When the bandwidth allocation for the 360 video is reduced (for example from 19 Mb/s to 7 Mb/s), the quality of the tiles is also reduced. In the foregoing example, the highest quality tiles in the direct field of vision may have a bitrate of 22.4 Mb/s with a QP value of 22.
As noted hereinabove, an aspect of the functionality of a BWA-TS module is to insure that the overall bandwidth of an example 360° immersive video session does not exceed a designated bandwidth allocation (e.g., based on network operator policies, content provider policies, subscriber/device policies, or any combination thereof), while still maximizing quality and viewing experience. Optimized tile selection having suitable bitrate qualities may therefore be configured responsive to a user's field of vision, bandwidth allocation/limitation, bitrates per tile as well as a transmit buffer model such that tiles in the direct line of sight have the best quality possible, with decreasing qualities moving farther away from the direct gaze.
An example stream-level manifest for purposes of an embodiment of the present invention is illustrated below:
<?xml version=“1.0”?>
<TiledMediaDefinition>
<TileGroup>
</Representation>
</TileGroup>
</TiledMediaDefinition>
An example lower-level manifest based on DASH-MPD for purposes of an embodiment of the present invention involving multiple phase encoded streams is illustrated below:
<?xml version=“1.0” encoding=“UTF-8” ?>
</Period>
</MPD>
Turning to
Similar to the embodiment of
Example annealing processes set forth herein advantageously facilitate gradual refreshing of frames when a viewport or bandwidth is changed, thereby allowing for the ability to minimize latency in increasing quality based on a user's field of vision and at the same time not overload the bandwidth. Typically, when attempting to perform quality changes on all tiles at the same time, several issues may be encountered due to the result of changing P-tiles for I-tiles at the same time, which are expensive in terms of encoded bitrate. On the other hand, performing this substitution with a minimal client buffer can cause too much delay in delivering the I-slices/frames.
In an example embodiment that employs gradual refreshing, the video streams do not have I-frames (except for the initial I-frame or any other special frames like Instant Decode Refresh or IDR frames). Instead, a video stream has I-blocks or I-tiles that may be distributed throughout a time sequence so that any particular spot on the screen gets an I-block at regular intervals, e.g., by way of phase-encoded streams as described in detail in the earlier sections of the present patent application. Thus, in such a scenario, there is no frame where all the pixels are refreshed by I-blocks. By performing gradual refresh annealing, example embodiments of the present invention can be advantageously configured to level out frame sizes (i.e., in terms of the amount of coded image data) and reduce the bandwidth consequences of injecting an I-frame to upgrade the quality of tiles that enter the FoV or viewport. Whereas a PE scheme may allow selective early refreshes of a tile in a time/frame sequence, it may impose certain bandwidth cost (e.g., due to having multiple I-tiles in a frame, which can cause an increase in the required bandwidth for that time interval corresponding to the transport of that video frame). However, an example embodiment involving PE can be configured such that the advantage of having a steadier level of bytes/frame overweighs such costs.
Over time in a frame sequence, a PE-based embodiment may allow manipulation of the phases of the various tiles around until the I-tiles are roughly evenly distributed in time again. Such a capability can be configured to be user- and/or content-dependent with respect to when this redistribution occurs as it requires the user to keep their field of view steady long enough for it to occur. In order to choose tiles to fill the bandwidth, an example embodiment may involve modeling the byte sizes of frames stretching 3 GOPs into the future (this choice is arbitrary) and performing hypothetical early refreshes (HER) based on the buffer model (e.g., with 3 GOPs in a look-ahead scenario). Based on the embodiments set forth in
It should be appreciated that an example upgrade process may move around in time and in space depending on the look-ahead GOP modeling. In one arrangement, each tile may have a 3-4 GOP horizon, which can each be upgraded as the process is iterated, where future GOP upgrades are for potential future enhancements for early refreshes covering 3-4 GOPs in to the future.
In considering a HER-based implementation, a few potential metrics may be identified and/or employed to obtain a suitable trade-off: (i) dead air, (ii) maximum buffer level, and (iii) end buffer level, among others. In one example implementation, the maximum buffer level may be weighted as a leading criterion for HER upgrades where adequate bandwidth may be freed up to allow tile-GOP quality upgrades.
As set forth in the embodiment of
Skilled artisans will recognize that a heap memory structure employed in the embodiment of
Further, example embodiments also include a penalty factor in scoring the tiles for an upgrade process as noted above. In one arrangement, a penalty may be imposed when an early refresh with an I-tile is required wherein a tile in the current GOP is to be upgraded beyond the quality it had in the previous slice/frame. Such a penalty has the effect of increasing that tile's adequacy which delays the upgrade relative to other tiles in the heap. This allows tile upgrades when the gaze has changed enough but defers early refreshes in marginal cases.
It will be apparent to one skilled in the art that additional/alternative formulations may also be used for scoring tile upgrades in some variations within the scope of the present invention.
bi+1=Max{0,ai−r(ti+1−ti)}
A buffer_size parameter may be defined as follows:
buffer_size=r(latency_frames)Δt
According to the foregoing model, if Max(ai)>buffer_size, it may be indicated as a buffer overflow condition. Thus, as different ni are being added pursuant to a tile upgrade process, the buffer end point level can be checked against a calculated buffer size in order to insure that no buffer violations are engendered in the upgrade process.
Turning to
Client optimization module 2306 preferably includes a 360° immersive video interface module 2321 comprising a manifest parser 2328, a video tile and audio stream downloader 2330, a bandwidth estimation module 2326 and a tile selection module 2324, which may be configured to operate in a manner similar to the embodiments set forth hereinabove with suitable device-centric modifications, mutatis mutandis. An HEVC tile/audio request 2344 may be generated to a network location, e.g., a content provider network or a cloud-based storage, via the Internet 2304, based on a manifest 2340 with respect to a particular content. Requested video tiles and audio data may be received via path 2342. Gaze vector information provided to the immersive video interface module 2321 from the gaze tracking module 2312 (e.g., via path 2322) may be utilized along with bandwidth estimation in selecting tiles per frame, which may be provided via a video signal path 2331 to a dynamically allocated video buffer 2332. Likewise, corresponding audio segments may be provided to an audio buffer 2336 via an audio signal path 2338. Tiles of different qualities may be provided to a tile combiner 2334, which generates a muxed encoded video stream 2346 to the player's video decoder 2314. Encoded audio stream 2348 may be generated from the audio buffer 2336 to the audio decoder 2318. Decoded audio and video data provided to the respective renderers 2320, 2316 of the player 2308 are rendered appropriately for display/presentation in an immersive environment effectuated by the user's display device, essentially similar to the example embodiments set forth previously.
Other I/O or interfaces such as an immersive display interface 2515, touch screen or keypad interface 2520, USB/HDMI ports 2518, Ethernet I/F 2508, and short-range and wide area wireless connectivity interfaces 2512 may also be provided depending on device configuration. Various motion detection and gaze tracking sensors 2516 may also be included, some of which may comprise gyroscopes, accelerometers, position sensors, etc. A hard disk drive (HDD) or local DVR system 2510 may be included in an example implementation for local storage of various program assets. A suitable power supply block 2522 may include AC/DC power conversion to provide power for the device 2500. It should be appreciated that the actual power architecture for the device 2500 may vary by the hardware platform used, e.g., depending upon the core SoC (System-on-Chip), memory, analog front-end, analog signal chain components and interfaces used in the specific platform, and the like.
One skilled in the art will recognize that various apparatuses and systems with respect to the foregoing embodiments, as well as the underlying network infrastructures set forth above may be architected in a virtualized environment according to a network function virtualization (NFV) architecture in additional or alternative embodiments of the present patent disclosure. For instance, various physical resources, databases, services, applications and functions executing within an example streaming network of the present application, including source media processing infrastructure, media containerization, PE/BIE tile encoding and packaging, etc., set forth hereinabove may be provided as virtual appliances, machines or functions, wherein the resources and applications are virtualized into suitable virtual network functions (VNFs) or virtual network elements (VNEs) via a suitable virtualization layer. Resources comprising compute resources, memory resources, and network infrastructure resources are virtualized into corresponding virtual resources wherein virtual compute resources, virtual memory resources and virtual network resources are collectively operative to support a VNF layer, whose overall management and orchestration functionality may be supported by a virtualized infrastructure manager (VIM) in conjunction with a VNF manager and an NFV orchestrator. An Operation Support System (OSS) and/or Business Support System (BSS) component may typically be provided for handling network-level functionalities such as network management, fault management, configuration management, service management, and subscriber management, etc., which may interface with VNF layer and NFV orchestration components via suitable interfaces.
Furthermore, at least a portion of an example network architecture disclosed herein may be virtualized as set forth above and architected in a cloud-computing environment comprising a shared pool of configurable virtual resources. Various pieces of hardware/software associated with PE/BIE tile encoding and packaging, bandwidth annealing and tile selection, tile muxing and containerization, and the like may be implemented in a service-oriented architecture, e.g., Software as a Service (SaaS), Platform as a Service (PaaS), infrastructure as a Service (IaaS) etc., with multiple entities providing different features of an example embodiment of the present invention, wherein one or more layers of virtualized environments may be instantiated on commercial off the shelf (COTS) hardware. Skilled artisans will also appreciate that such a cloud-computing environment may comprise one or more of private clouds, public clouds, hybrid clouds, community clouds, distributed clouds, multiclouds and interclouds (e.g., “cloud of clouds”), and the like.
In the above-description of various embodiments of the present disclosure, it is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of this specification and the relevant art and may not be interpreted in an idealized or overly formal sense expressly so defined herein.
At least some example embodiments are described herein with reference to block diagrams and/or flowchart illustrations of computer-implemented methods, apparatus (systems and/or devices) and/or computer program products. It is understood that a block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by computer program instructions that are performed by one or more computer circuits. Such computer program instructions may be provided to a processor circuit of a general purpose computer circuit, special purpose computer circuit, and/or other programmable data processing circuit to produce a machine, so that the instructions, which execute via the processor of the computer and/or other programmable data processing apparatus, transform and control transistors, values stored in memory locations, and other hardware components within such circuitry to implement the functions/acts specified in the block diagrams and/or flowchart block or blocks, and thereby create means (functionality) and/or structure for implementing the functions/acts specified in the block diagrams and/or flowchart block(s). Additionally, the computer program instructions may also be stored in a tangible computer-readable medium that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable medium produce an article of manufacture including instructions which implement the functions/acts specified in the block diagrams and/or flowchart block or blocks.
As pointed out previously, tangible, non-transitory computer-readable medium may include an electronic, magnetic, optical, electromagnetic, or semiconductor data storage system, apparatus, or device. More specific examples of the computer-readable medium would include the following: a portable computer diskette, a random access memory (RAM) circuit, a read-only memory (ROM) circuit, an erasable programmable read-only memory (EPROM or Flash memory) circuit, a portable compact disc read-only memory (CD-ROM), and a portable digital video disc read-only memory (DVD/Blu-ray). The computer program instructions may also be loaded onto or otherwise downloaded to a computer and/or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer and/or other programmable apparatus to produce a computer-implemented process. Accordingly, embodiments of the present invention may be embodied in hardware and/or in software (including firmware, resident software, micro-code, etc.) that runs on a processor or controller, which may collectively be referred to as “circuitry,” “a module” or variants thereof. Further, an example processing unit may include, by way of illustration, a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor (DSP), a plurality of microprocessors, one or more microprocessors in association with a DSP core, a controller, a microcontroller, Application Specific Integrated Circuits (ASICs), Field Programmable Gate Array (FPGA) circuits, any other type of integrated circuit (IC), and/or a state machine. As can be appreciated, an example processor unit may employ distributed processing in certain embodiments.
Further, in at least some additional or alternative implementations, the functions/acts described in the blocks may occur out of the order shown in the flowcharts. For example, two blocks shown in succession may in fact be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending upon the functionality/acts involved. Moreover, the functionality of a given block of the flowcharts and/or block diagrams may be separated into multiple blocks and/or the functionality of two or more blocks of the flowcharts and/or block diagrams may be at least partially integrated. Furthermore, although some of the diagrams include arrows on communication paths to show a primary direction of communication, it is to be understood that communication may occur in the opposite direction relative to the depicted arrows. Finally, other blocks may be added/inserted between the blocks that are illustrated.
It should therefore be clearly understood that the order or sequence of the acts, steps, functions, components or blocks illustrated in any of the flowcharts depicted in the drawing Figures of the present disclosure may be modified, altered, replaced, customized or otherwise rearranged within a particular flowchart, including deletion or omission of a particular act, step, function, component or block. Moreover, the acts, steps, functions, components or blocks illustrated in a particular flowchart may be inter-mixed or otherwise inter-arranged or rearranged with the acts, steps, functions, components or blocks illustrated in another flowchart in order to effectuate additional variations, modifications and configurations with respect to one or more processes for purposes of practicing the teachings of the present patent disclosure.
Although various embodiments have been shown and described in detail, the claims are not limited to any particular embodiment or example. None of the above Detailed Description should be read as implying that any particular component, element, step, act, or function is essential such that it must be included in the scope of the claims. Reference to an element in the singular is not intended to mean “one and only one” unless explicitly so stated, but rather “one or more.” All structural and functional equivalents to the elements of the above-described embodiments that are known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the present claims. Accordingly, those skilled in the art will recognize that the exemplary embodiments described herein can be practiced with various modifications and alterations within the spirit and scope of the claims appended below.
| Number | Name | Date | Kind |
|---|---|---|---|
| 7975283 | Bedingfield | Jul 2011 | B2 |
| 8767014 | Vaught et al. | Jul 2014 | B2 |
| 9392212 | Ross | Jul 2016 | B1 |
| 9918136 | Cole et al. | Mar 2018 | B2 |
| 20050018911 | Deever | Jan 2005 | A1 |
| 20090175333 | Hsiang | Jul 2009 | A1 |
| 20100046845 | Wedi et al. | Feb 2010 | A1 |
| 20100050221 | McCutchen et al. | Feb 2010 | A1 |
| 20100061461 | Bankoski et al. | Mar 2010 | A1 |
| 20100149073 | Chaum et al. | Jun 2010 | A1 |
| 20120201294 | Segall et al. | Aug 2012 | A1 |
| 20120257677 | Bankoski et al. | Oct 2012 | A1 |
| 20120281928 | Cohen et al. | Nov 2012 | A1 |
| 20130182073 | Yin et al. | Jul 2013 | A1 |
| 20130235152 | Hannuksela et al. | Sep 2013 | A1 |
| 20130294513 | Seregin et al. | Nov 2013 | A1 |
| 20150212576 | Ambrus et al. | Jul 2015 | A1 |
| 20150229948 | Puri et al. | Aug 2015 | A1 |
| 20150288732 | Phillips et al. | Oct 2015 | A1 |
| 20150346812 | Cole et al. | Dec 2015 | A1 |
| 20150382057 | Huang et al. | Dec 2015 | A1 |
| 20160277772 | Campbell et al. | Sep 2016 | A1 |
| 20160353146 | Weaver et al. | Dec 2016 | A1 |
| 20160379606 | Kollin et al. | Dec 2016 | A1 |
| 20170139578 | Dickerson et al. | May 2017 | A1 |
| 20170263046 | Patney et al. | Sep 2017 | A1 |
| 20170316601 | Kakarlapudi et al. | Nov 2017 | A1 |
| 20170339416 | Hendry et al. | Nov 2017 | A1 |
| 20170374375 | Makar et al. | Dec 2017 | A1 |
| 20170374411 | Lederer et al. | Dec 2017 | A1 |
| 20180063512 | Hong et al. | Mar 2018 | A1 |
| 20180103199 | Hendry et al. | Apr 2018 | A1 |
| 20180160160 | Swaminathan et al. | Jun 2018 | A1 |
| 20180164593 | Van Der Auwera et al. | Jun 2018 | A1 |
| 20180220119 | Horvitz et al. | Aug 2018 | A1 |
| 20180249163 | Curcio et al. | Aug 2018 | A1 |
| 20190089643 | Westphal et al. | Mar 2019 | A1 |
| 20190104324 | Han et al. | Apr 2019 | A1 |
| 20190110856 | Barral et al. | Apr 2019 | A1 |
| Number | Date | Country |
|---|---|---|
| 2554877 | Apr 2018 | GB |
| 2557416 | Jun 2018 | GB |
| WO2016129549 | Aug 2016 | JP |
| Entry |
|---|
| ISA/EPO, International Search Report, International Application No. PCT/EP2019/061840, dated Jun. 18, 2019, 16 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/009,003, dated May 16, 2019, 20 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/009,031, dated May 29, 2019, 21 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/046,576, dated Apr. 30, 2019, 17 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/046,269, dated Apr. 16, 2019, 6 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/046,269, dated May 29, 2019, 11 pgs. |
| Sullivan et al., “Overview of the High Efficiency Video Coding (HEVC) Standard”, IEEE Transactions on Circuits and Systems for Video Technology, Dec. 2012, 20 pgs., vol. 22, No. 12. |
| USPTO, Office Action, U.S. Appl. No. 16/046,629, dated Dec. 31, 2018, 9 pgs. |
| Bassbouss et al., “High Quality 360° Video Rendering and Streaming,” Fraunhofer Institute for Open Communication Systems FOKUS, Nov. 2016, 4 pgs. |
| Corbillon et al., “Viewport-Adaptive Navigable 360-Degree Video Delivery,” 2017 IEEE International Conference on Communications (ICC), May 21, 2017, 7 pgs. |
| Kuzyakov et al., “Next-generation video encoding techniques for 360 video and VR,” Engineering Blog: Facebook Code—Facebook, Jan. 21, 2016, 7 pgs. <https://code.facebook.com/posts/1126354007399553/next-generation-video-encoding-techniques-for-360-video-and-vr/>. |
| Misra et al., “An Overview of Tiles in HEVC,” IEEE Journal of Selected Topics in Signal Processing, Dec. 2013, pp. 969-970, vol. 7, No. 6. |
| Sanchez et al., “Compressed domain video processing for tile based panoramic streaming using HEVC,” 2015 IEEE International Conference on Image Processing (ICIP), Sep. 27, 2015, 5 pgs. |
| Sanchez et al., “Low Complexity Cloud-video-Mixing Using HEVC,” 11th Annual IEEE CCNC—Multimedia Networking, Services and Applications, Jan. 10, 2014, pp. 213-218, Las Vegas, Nevada, USA. |
| Skupin et al., “HEVC Tile Based Streaming to Head Mounted Displays,” 2017 14th IEEE Annual Consumer Communications & Networking Conference (CCNC), Jan. 8, 2017, pp. 613-615. |
| Valero et al., “Eye-tracking Based Video Streaming System with Fast Quality Switching,” Master's Thesis at KTH Information and Communication Technology, Jul. 2017, Stockholm, Sweden, 89 pgs. |
| Van Brandenburg, “CDN Optimization for VR Streaming,” Proceedings of the IBC 2017 Conference, Amsterdam, The Netherlands, Sep. 2017, 9 pgs. |
| Advertisement Overlay Insertion for HEVC based services, Fraunhofer Heinrich Hertz Institute, accessed Nov. 10, 2017. <https://www.hhi.fraunhofer.de/en/departments/vca/research-groups/multimedia-commuications/research-topics/ advertisement-overlay-insertion-for-hevc-based-services.html>. |
| ISA/SE, International Search Report, PCT/SE2019/050552, dated Sep. 20, 2019, 5 pgs. |
| ISA/SE, Written Opinion of the International Search Authority, PCT/SE2019/050552, dated Sep. 20, 2019, 9 pgs. |
| ISA/SE, International Search Report, PCT/SE2019/050559, dated Sep. 18, 2019, 4 pgs. |
| ISA/SE, Written Opinion of the International Search Authority, PCT/SE2019/050559, dated Sep. 18, 2019, 7 pgs. |
| Guo et al., Viewpoint Switching in Multiview Video Streaming, 2005 IEEE International Symposium on Circuits and Systems (ISCAS), May 23-26, 2005, IEEE, Japan. |
| Ling et al., Adaptive Multi-Path Video Streaming, Proceedings of the Eighth IEEE International Symposium on Multimedia (ISM'06), Dec. 11-13, 2006, IEEE Computer Society. |
| Lungaro et al., Gaze-aware streaming solutions for the next generation of mobile VR experiences, IEEE Transactions on Visualization and Computer Graphics, Jan. 25, 2018, pp. 1535-1544, vol. 24, IEEE. |
| Ozcinar et al., Dynamic bitrate allocation of interactive real-time streamed multi-view video with view-switch prediction, Signal, Image and Video Processing, Mar. 27, 2017, pp. 1279-1285, vol. 11, Springer-Verlag, United Kingdom. |
| Zhu et al., Inter-view-predicted redundant pictures for viewpoint switching in multiview video streaming, 2010 IEEE International Conference on Acoustics, Speech and Signal Processing, Mar. 14-19, 2010, IEEE, USA. |
| USPTO, Office Action, U.S. Appl. No. 16/008,824, dated Aug. 21, 2019, 8 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/046,629, dated Oct. 3, 2019, 6 pgs. |
| USPTO, Notice of Allowance, U.S. Appl. No. 16/148,001, dated Aug. 8, 2019, 10 pgs. |
| Number | Date | Country | |
|---|---|---|---|
| 20190387237 A1 | Dec 2019 | US |
