Abstract: The technology described herein relates to implementing an adaptive bitrate (ABR) algorithm at edge nodes. A method for implementing an ABR algorithm at an edge node may include receiving at the edge node a request for a video segment from a client according to the client's ABR algorithm, the request indicating a quality. A weighted sum score for each of a set of qualities may be computed based on a quality score and a fairness score using the ABR algorithm at the edge node, the qualities including at least the requested quality and another quality. A modified request may be generated in response to the weighted sum score for the other quality being better than the weighted sum score for the requested quality. The modified request may be sent to a server. The video segment in the other quality may be received from the server and provided to a client.

Abstract: According to embodiments of the disclosure, fast multi-rate encoding may be performed using machine learning by encoding a lowest quality representation to determine encoding parameters, processing raw data of the video using a neural network to obtain an intermediate output comprising encoding features, augmenting the intermediate output with additional encoding features to form a final tensor, and processing the final tensor with another neural network to obtain a classification output comprising a split or not split decision for an image data block. The classification output may be used to encode a highest quality representation, and then other representations of the video.

Abstract: The technology described herein relates to online per-title encoding. A method for online per-title encoding includes receiving a video input, generating segments of the video input, extracting a spatial feature and a temporal feature, predicting bitrate-resolution pairs based on the spatial feature and the temporal feature, using a discrete cosine transform (DCT)-based energy function, and per-title encoding segments of the video input for the predicted bitrate-resolution pairs. A system for online per-title encoding may include memory for storing a set of bitrates, a set of resolutions, and a machine learning module configured to predict bitrate resolution pairs based on low-complexity spatial and temporal features.

Abstract: A computer-implemented method and system for transcoding input video content is provided. The method includes decoding the input video content from a first format to a first set of raw video data. Encoding the first set of raw video data into an intermediate format and storing the video data in the second intermediate format. Also encoding the first set of raw video data into a third desired output format to extract video parameters and determining optimized encoding parameters for encoding the video content into the final output video. The method includes decoding the stored video data encoded into the intermediate format into a second set of raw video data and encoding the second set of raw video data into the third desired output format using the optimized encoding parameters to generate the final output video.

Abstract: The technology described herein relates to online per-title encoding. A method for online per-title encoding includes receiving a video input, generating segments of the video input, extracting a spatial feature and a temporal feature, predicting bitrate-resolution pairs based on the spatial feature and the temporal feature, using a discrete cosine transform (DCT)-based energy function, and per-title encoding segments of the video input for the predicted bitrate-resolution pairs. A system for online per-title encoding may include memory for storing a set of bitrates, a set of resolutions, and a machine learning module configured to predict bitrate resolution pairs based on low-complexity spatial and temporal features.

Abstract: A scalable per-title encoding technique may include detecting scene cuts in an input video received by an encoding network or system, generating segments of the input video, performing per-title encoding of a segment of the input video, training a deep neural network (DNN) for each representation of the segment, thereby generating a trained DNN, compressing the trained DNN, thereby generating a compressed trained DNN, and generating an enhanced bitrate ladder including metadata comprising the compressed trained DNN. In some embodiments, the method also may include generating a base layer bitrate ladder for CPU devices, and providing the enhanced bitrate ladder for GPU-available devices.

Abstract: A multimedia player downloads chunks (parts of the segment file) during the download of a segment of a stream of segments of a low-latency stream. The first chunks of a segment may be downloaded while the segment is still being written to the CDN server. A chunk-based prediction adaptation logic uses throughput measurements on a chunk instead of a segment and specifically looks at bursts in a sliding window. This data is used to build a prediction of future throughput by applying linear adaptive filter algorithms that may rely on recursive least squares. This adaptation logic leads to very accurate bandwidth predictions and as a consequence, better user experience, compared to existing adaptation algorithms.

Abstract: The technology described herein relates to implementing an adaptive bitrate (ABR) algorithm at edge nodes. A method for implementing an ABR algorithm at an edge node may include receiving at the edge node a request for a video segment from a client according to the client’s ABR algorithm, the request indicating a quality. A weighted sum score for each of a set of qualities may be computed based on a quality score and a fairness score using the ABR algorithm at the edge node, the qualities including at least the requested quality and another quality. A modified request may be generated in response to the weighted sum score for the other quality being better than the weighted sum score for the requested quality. The modified request may be sent to a server. The video segment in the other quality may be received from the server and provided to a client.

Abstract: An original input content is subjected to multiple constant quality probe encodes for a defined set of resolutions. In one embodiment, probe encodes encode a few parts of the original source video, for example, 30 seconds from 5 different positions. Each probe encode delivers an average bitrate that is required to achieve the configured constant quality. The mean value of the average bitrate is taken per resolution. This results in a list of bitrates that map to a resolution that would achieve the best quality, a custom bitrate table. Based on the custom bitrate table, an optimized bitrate ladder is computed. The process starts with a configurable minimum bitrate and steps up by a bitrate step size that is between a configurable min and max bitrate step size until the bitrate of the highest resolution is reached.

Abstract: Techniques relating to per-title encoding using spatial and temporal resolution downscaling is disclosed. A method for per-title encoding includes receiving a video input comprised of video segments, spatially downscaling the video input, temporally downscaling the video input, encoding the video input to generate an encoded video, then temporally and spatially upscaling the encoded video. Spatially downscaling may include reducing a resolution of the video input, and temporally downscaling may include reducing a framerate of the video input. Objective metrics for the upscaled encoded video show improved quality over conventional methods.

Abstract: According to embodiments of the disclosure, information of higher and lower quality encoded video segments is used to limit Rate-Distortion Optimization (RDO) for each Coding Unit Tree (CTU). A method first encodes the highest bit-rate segment and consequently uses it to encode the lowest bit-rate video segment. Block structure and selected reference frame of both highest and lowest bit-rate video segments are used to predict and shorten RDO process for each CTU in middle bit-rates. The method delays just one frame using parallel processing. This approach provides time-complexity reduction compared to the reference software for middle bit-rates while degradation is negligible.

Abstract: An original input video file is encoded using a machine learning approach. The encoder performs a detailed video analysis and selection of encoding parameters that using a machine learning algorithm improves over time. The encoding process is done using a multi-pass approach. During a first pass, the entire video file is scanned to extract video property information that does not require in-depth analyses. The extracted data is then entered into an encoding engine, which uses artificial intelligence to produce optimized encoder settings. The video file is into a set of time-based chunks and, in a second pass, the encoding parameters for each chunk are set and distributed to encoding nodes for parallel processing. These encoder instances probe-encode each chunk determine the level of complexity for the chunk and to derive chunk-specific encoding parameters.

Abstract: Disclosed are systems and methods for lightweight transcoding of video. A distributed computing system for lightweight transcoding includes an origin server and an edge node, the origin server having a memory and a processor and configured to receive an input video comprising a bitstream, encode the bitstream into a set of representations corresponding to a full bitrate ladder, generate encoding metadata for the set of representations, and provide a representation and encoding metadata for the set of representations to an edge node, the edge node having a memory and a processor and configured to transcode the bitstream, or segments thereof, into the set of representations, and to serve one or more of the representations to a client.

Abstract: According to embodiments of the disclosure, fast multi-rate encoding may be performed using machine learning by encoding a lowest quality representation to determine encoding parameters, processing raw data of the video using a neural network to obtain an intermediate output comprising encoding features, augmenting the intermediate output with additional encoding features to form a final tensor, and processing the final tensor with another neural network to obtain a classification output comprising a split or not split decision for an image data block. The classification output may be used to encode a highest quality representation, and then other representations of the video.

Abstract: An original input content is subjected to multiple constant quality probe encodes for a defined set of resolutions. In one embodiment, probe encodes encode a few parts of the original source video, for example, 30 seconds from 5 different positions. Each probe encode delivers an average bitrate that is required to achieve the configured constant quality. The mean value of the average bitrate is taken per resolution. This results in a list of bitrates that map to a resolution that would achieve the best quality, a custom bitrate: table. Based on the custom bitrate table, an optimized bitrate ladder is computed. The process starts with a configurable minimum bitrate and steps up by a bitrate step size that is between a configurable min and max bitrate step size until the bitrate of the highest resolution is reached.

Abstract: An original input content is subjected to multiple constant quality probe encodes for a defined set of resolutions. In one embodiment, probe encodes encode a few parts of the original source video, for example, 30 seconds from 5 different positions. Each probe encode delivers an average bitrate that is required to achieve the configured constant quality. The mean value of the average bitrate is taken per resolution. This results in a list of bitrates that map to a resolution that would achieve the best quality, a custom bitrate table. Based on the custom bitrate table, an optimized bitrate ladder is computed. The process starts with a configurable minimum bitrate and steps up by a bitrate step size that is between a configurable min and max bitrate step size until the bitrate of the highest resolution is reached.

Abstract: A multimedia player downloads chunks (parts of the segment file) during the download of a segment of a stream of segments of a low-latency stream. The first chunks of a segment may be downloaded while the segment is still being written to the CDN server. A chunk-based prediction adaptation logic uses throughput measurements on a chunk instead of a segment and specifically looks at bursts in a sliding window. This data is used to build a prediction of future throughput by applying linear adaptive filter algorithms that may rely on recursive least squares. This adaptation logic leads to very accurate bandwidth predictions and as a consequence, better user experience, compared to existing adaptation algorithms.

Abstract: An original input video file is encoded using a machine learning approach. The encoder performs a detailed video analysis and selection of encoding parameters that using a machine learning algorithm improves over time. The encoding process is done using a multi-pass approach. During a first pass, the entire video file is scanned to extract video property information that does not require in-depth analyses. The extracted data is then entered into an encoding engine, which uses artificial intelligence to produce optimized encoder settings. The video file is into a set of time-based chunks and, in a second pass, the encoding parameters for each chunk are set and distributed to encoding nodes for parallel processing. These encoder instances probe-encode each chunk determine the level of complexity for the chunk and to derive chunk-specific encoding parameters.

Abstract: According to embodiments of the disclosure, information of higher and lower quality encoded video segments is used to limit Rate-Distortion Optimization (RDO) for each Coding Unit Tree (CTU). A method first encodes the highest bit-rate segment and consequently uses it to encode the lowest bit-rate video segment. Block structure and selected reference frame of both highest and lowest bit-rate video segments are used to predict and shorten RDO process for each CTU in middle bit-rates. The method delays just one frame using parallel processing. This approach provides time-complexity reduction compared to the reference software for middle bit-rates while degradation is negligible.

Abstract: An original input video file is encoded using a machine learning approach. The encoder performs a detailed video analysis and selection of encoding parameters that using a machine learning algorithm improves over time. The encoding process is done using a multi-pass approach. During a first pass, the entire video file is scanned to extract video property information that does not require in-depth analyses. The extracted data is then entered into an encoding engine, which uses artificial intelligence to produce optimized encoder settings. The video file is into a set of time-based chunks and, in a second pass, the encoding parameters for each chunk are set and distributed to encoding nodes for parallel processing. These encoder instances probe-encode each chunk determine the level of complexity for the chunk and to derive chunk-specific encoding parameters.