Abstract: A method for merging graphics and high dynamic range video data is disclosed. In a video receiver, a display management process uses metadata to map input video data from a first dynamic range into the dynamic range of available graphics data. The remapped video signal is blended with the graphics data to generate a video composite signal. An inverse display management process uses the metadata to map the video composite signal to an output video signal with the first dynamic range. To alleviate perceptual tone-mapping jumps during video scene changes, a metadata transformer transforms the metadata to transformed so that on a television (TV) receiver metadata values transition smoothly between consecutive scenes. The TV receiver receives the output video signal and the transformed metadata to generate video data mapped to the dynamic range of the TV's display.

Abstract: A method and apparatus of frame rate conversion where descriptive information relating to an input video signal is transmitted to the frame rate converter along with the input signal. The descriptive information is used to interpolate frames, allowing the interpolator to make pixel analysis using the descriptive information relating to the original signal. The descriptive information is derived by a compositor/scaler and is transmitted to the frame rate converter with the composited/scaled signal. There may be multiple input signals that are composited to make a final composited video output such as a picture-in-picture display. The information may be transmitted in-band with the video signal received by the frame rate converter. Alternatively, the descriptive information may be transmitted in a separate stream in a side-band manner. In another embodiment, the descriptive information may be transmitted to the frame rate converter separate from the video input source as commands or packets.

Abstract: A method for merging graphics and high dynamic range video data is disclosed. In a video receiver, a display management process uses metadata to map input video data from a first dynamic range into the dynamic range of available graphics data. The remapped video signal is blended with the graphics data to generate a video composite signal. An inverse display management process uses the metadata to map the video composite signal to an output video signal with the first dynamic range. To alleviate perceptual tone-mapping jumps during video scene changes, a metadata transformer transforms the metadata to transformed so that on a television (TV) receiver metadata values transition smoothly between consecutive scenes. The TV receiver receives the output video signal and the transformed metadata to generate video data mapped to the dynamic range of the TV's display.

Abstract: Given an input progressive sequence, a video encoder creates a dual-layer stream that combines a backwards-compatible interlaced video stream layer with an enhancement layer to reconstruct full-resolution progressive video. Given two consecutive frames in the input progressive sequence, vertical processing generates a top field-bottom field (TFBF) frame in a base layer (BL) TFBF sequence, and horizontal processing generates a side-by-side (SBS) frame in an enhancement layer (EL) SBS video sequence. The BL TFBF and the EL SBS sequences are compressed together to create a coded, backwards compatible output stream.

Abstract: One or more derived versions of image content may be obtained by interpolating two or more source versions of the same image content. A derived version may be targeted for a class of displays that differs from classes of displays targeted by the source versions. Source images in a source version may have been color graded in a creative process by a content creator/colorist. Interpolation of the source versions may be performed with interpolation parameters having two or more different values in two or more different clusters in at least one of the source images. A normalized version may be used to allow efficient distribution of multiple versions of the same content to a variety of downstream media processing devices, and to preserve or restore image details otherwise lost in one or more of the source versions.

Abstract: 3D Images may be encoded into reduced resolution image data in a base layer and enhancement layer (EL) image data in one or more enhancement layers. Different types of data compositions may be used in the EL image data. The different types of data compositions may include unfiltered full resolution image data for one or both of left eye and right eye perspectives, or unfiltered full resolution image data for a color channel, e.g., luminance channel, or unfiltered full resolution image data for selected portions of image frames, or fallback data compositions. Based on deciding factors including bitrate requirements and bandwidth constraints, different types of data compositions may be alternatively used by an upstream device to deliver the best possible 3D image data to a wide variety of downstream devices. The upstream device may inform a downstream device of specific types of data compositions with EL image data descriptors.

Abstract: 3D Images may be encoded into reduced resolution image data in a base layer and enhancement layer (EL) image data in one or more enhancement layers. Different types of data compositions may be used in the EL image data. The different types of data compositions may include unfiltered full resolution image data for one or both of left eye and right eye perspectives, or unfiltered full resolution image data for a color channel, e.g., luminance channel, or unfiltered full resolution image data for selected portions of image frames, or fallback data compositions. Based on deciding factors including bitrate requirements and bandwidth constraints, different types of data compositions may be alternatively used by an upstream device to deliver the best possible 3D image data to a wide variety of downstream devices. The upstream device may inform a downstream device of specific types of data compositions with EL image data descriptors.

Abstract: A high resolution 3D Image may be encoded into a reduced resolution image in a base layer and a full resolution unfiltered image in one or more enhancement layers. Encoded asymmetric-resolution image data for the 3D image may be distributed to a wide variety of devices for 3D image processing and rendering. A recipient device may reconstruct the reduced resolution image and the full resolution unfiltered image for 3D image rendering with high subjective perceptual quality due to interocular masking. Full resolution unfiltered images may be alternating between left and right eyes.

Abstract: Methods and systems for image processing and delivery of higher dynamic range cinema content are disclosed. A digital cinema signal with a lower dynamic range is obtained from a digital cinema signal with a higher dynamic range, for example through mapping. The lower dynamic range digital cinema signal is encoded and decoded at the transmitting end. The decoded lower dynamic range digital cinema signal is normalized to produce a set of normalization parameters which enable the mapping process at the receiving end to produce a final image with higher dynamic range that is of a higher quality. Alternatively, the higher dynamic range digital cinema signal is also encoded and decoded at the transmitting end, to produce a set of normalization parameters which enable the mapping process at the receiving end to produce a final image with higher dynamic range that is of a higher quality.

Abstract: Methods and systems for image processing and delivery of higher dynamic range cinema content are disclosed. A digital cinema signal with a lower dynamic range is obtained from a digital cinema signal with a higher dynamic range, for example through mapping. The lower dynamic range digital cinema signal is encoded and decoded at the transmitting end. The decoded lower dynamic range digital cinema signal is normalized to produce a set of normalization parameters which enable the mapping process at the receiving end to produce a final image with higher dynamic range that is of a higher quality. Alternatively, the higher dynamic range digital cinema signal is also encoded and decoded at the transmitting end, to produce a set of normalization parameters which enable the mapping process at the receiving end to produce a final image with higher dynamic range that is of a higher quality.

Abstract: A high resolution 3D image may be encoded into a first multiplexed image frame and a second multiplexed image frame in a base layer (BL) video signal and an enhancement layer (EL) video signal. The first multiplexed image frame may comprise horizontal high resolution image data for both eyes, while the second multiplexed image frame may comprise vertical high resolution image data for both eyes. Encoded symmetric-resolution image data for the 3D image may be distributed to a wide variety of devices for 3D image processing and rendering. A recipient device may reconstruct reduced resolution 3D image from one of the first multiplexed image frame or the second multiplexed image frame. A recipient device may also reconstruct high resolution 3D image by combining high resolution image data from both of the first multiplexed image frame and the second multiplexed image frame.

Abstract: Techniques use multiple lower bit depth codecs to provide higher bit depth, high dynamic range, images from an upstream device to a downstream device. A base layer and one or more enhancement layers may be used to carry video signals, wherein the base layer cannot be decoded and viewed on its own. Lower bit depth input image data to base layer processing may be generated from higher bit depth high dynamic range input image data via advanced quantization to minimize the volume of image data to be carried by enhancement layer video signals. The image data in the enhancement layer video signals may comprise residual values, quantization parameters, and mapping parameters based in part on a prediction method corresponding to a specific method used in the advanced quantization. Adaptive dynamic range adaptation techniques take into consideration special transition effects, such as fade-in and fade-outs, for improved coding performance.

Abstract: An input image is divided into non-overlapping regions. For each of the non-overlapping regions, first output data is predicted with a first prediction function, parameters related thereto and region-specific input image data. For each region with prior-predicted neighbor regions, a pixel border portion, adjacent to the neighbor region, is defined. For the pixels in the defined border portion, second output data is predicted with a second prediction function, parameters related thereto, input image data from the border portion of the current region, and input prediction parameter data from the neighbor region. The first output prediction data is fused with the second output data to predict a final set of output prediction values.

Abstract: One or more derived versions of image content may be obtained by interpolating two or more source versions of the same image content. A derived version may be targeted for a class of displays that differs from classes of displays targeted by the source versions. Source images in a source version may have been color graded in a creative process by a content creator/colorist. Interpolation of the source versions may be performed with interpolation parameters having two or more different values in two or more different clusters in at least one of the source images. A normalized version may be used to allow efficient distribution of multiple versions of the same content to a variety of downstream media processing devices, and to preserve or restore image details otherwise lost in one or more of the source versions.

Abstract: A sequence of enhanced dynamic range (EDR) images and a sequence of standard dynamic range images are encoded using a backwards-compatible SDR high-definition (HD) base layer and one or more enhancement layers. The EDR and SDR video signals may be of the same resolution (e.g., HD) or at different resolutions (e.g., 4K and HD) and are encoded using a dual-view-dual-layer (DVDL) encoder to generate a coded base layer (BL) and a coded enhancement layer (EL). The DVDL encoder includes a reference processing unit (RPU) which is adapted to compute a reference stream based on the coded BL stream. The RPU operations include post-processing, normalization, inverse normalization, and image registration. Decoders for decoding the coded BL and EL streams to generate a backwards compatible 2D SDR stream and additional 2D or 3D SDR or EDR streams, are also described.

Abstract: A method and apparatus for reducing motion judder in a 3D input source are disclosed. The 3D input source is separated into left and right images. Motion vectors for the left and right images are calculated. Frame rate conversion is performed on the left and right images, to produce motion compensated left and right images. The left and right images and the motion compensated left and right images are reordered for display. Alternatively, the motion estimation and motion compensation can be performed on the 3D input source, and the input image and the motion compensated image can then be separated into respective left and right images. The method and apparatus can be adapted to perform 2D to 3D conversion by extracting a 2D input source into left and right 3D images and performing motion estimation and motion compensation.

Abstract: A trellis decoder decodes a stream of encoded symbols, including symbols of a first type (e.g. symbols encoded with a first trellis code) and symbols of a second type (e.g. encoded with a second, more robust, trellis code), without storing path indicators along a trellis for symbols of the first type. In this way, limited memory may be used to store path indicators along the trellis for symbols of the second type. This allows for more accurate decoding of the symbols of the second type. For transitions from symbols of the second type to symbols of the first type, states of the trellis decoder may be stored. In this way, paths may be traced back along the trellis for trellis decoding, without the path indicators for the symbols of the first type.

Abstract: A trellis decoder decodes a stream of encoded symbols, including symbols of a first type (e.g. symbols encoded with a first trellis code) and symbols of a second type (e.g. encoded with a second, more robust, trellis code), without storing path indicators along a trellis for symbols of the first type. In this way, limited memory may be used to store path indicators along the trellis for symbols of the second type. This allows for more accurate decoding of the symbols of the second type. For transitions from symbols of the second type to symbols of the first type, states of the trellis decoder may be stored. In this way, paths may be traced back along the trellis for trellis decoding, without the path indicators for the symbols of the first type.

Abstract: A method and apparatus for reducing motion judder in a 3D input source are disclosed. The 3D input source is separated into left and right images. Motion vectors for the left and right images are calculated. Frame rate conversion is performed on the left and right images, to produce motion compensated left and right images. The left and right images and the motion compensated left and right images are reordered for display. Alternatively, the motion estimation and motion compensation can be performed on the 3D input source, and the input image and the motion compensated image can then be separated into respective left and right images. The method and apparatus can be adapted to perform 2D to 3D conversion by extracting a 2D input source into left and right 3D images and performing motion estimation and motion compensation.

Abstract: A trellis decoder decodes a stream of encoded symbols, including symbols of a first type (e.g. symbols encoded with a first trellis code) and symbols of a second type (e.g. encoded with a second, more robust, trellis code), without storing path indicators along a trellis for symbols of the first type. In this way, limited memory may be used to store path indicators along the trellis for symbols of the second type. This allows for more accurate decoding of the symbols of the second type. For transitions from symbols of the second type to symbols of the first type, states of the trellis decoder may be stored. In this way, paths may be traced back along the trellis for trellis decoding, without the path indicators for the symbols of the first type.