CN-122002087-A - Quality-aware profile filtering in content delivery

CN122002087ACN 122002087 ACN122002087 ACN 122002087ACN-122002087-A

Abstract

In some embodiments, a method determines a content instance to transmit to a client device. The content includes segments associated with a plurality of profiles. A target index and a bandwidth distribution are determined. The method determines a plurality of top level profile decisions for a plurality of segments, wherein a top level profile decision lists top level profiles selectable for corresponding ones of the plurality of segments. Based on the bandwidth distribution, performance of the target metrics is determined for the plurality of top level profile decisions. The method selects a top level profile decision from the top level profile decisions based on the performance of the target metrics of the top level profile decisions. The top level profile decision is used to determine a target quality. During content playback, the target quality is used to select a top level profile of a clip in the requested clip.

Inventors

HUANG LEMEI
Dai tongyu
ZHANG WENHAO
CHEN SI
Thomas. haughty

Assignees

迪士尼企业公司
北京优佳佳软件技术开发有限公司

Dates

Publication Date: 20260508
Application Date: 20250715
Priority Date: 20241105

Claims (20)

1. A method, comprising: determining a content instance to transmit to a client device, wherein the content instance includes a plurality of segments associated with a plurality of profiles; Determining a target index and bandwidth distribution; Determining a plurality of top level profile decisions for the plurality of segments, wherein a top level profile decision lists top level profiles that can be selected for respective ones of the plurality of segments; determining, based on the bandwidth distribution, a performance of the target metrics for the plurality of top-level profile decisions; Selecting a top-level profile decision from the plurality of top-level profile decisions based on performance of the target metrics of the plurality of top-level profile decisions, and A target quality is determined using the top level profile decision, wherein the target quality is used to select a top level profile for a segment of the plurality of segments requested during playback of the content instance.
2. The method of claim 1, wherein the target indicator is selected from a first indicator based on data savings and a second indicator based on quality loss.
3. The method of claim 1, wherein the bandwidth distribution is based on a download bandwidth probability.
4. A method as recited in claim 3, wherein the bandwidth distribution is based on information from the client device.
5. The method of claim 1, wherein determining the target indicator comprises: and receiving a target value of the target index.
6. The method of claim 5, wherein the target value of the target indicator is adjustable for each content instance or client device.
7. The method of claim 1, wherein determining the plurality of top-level profile decisions comprises: determining an index value of a quality index of a profile of the plurality of profiles, and A section partition is determined based on the index values of the quality indicators of a segment, wherein the section partition lists the index values of a set of quality indicators to form sections associated with respective profiles.
8. The method of claim 7, wherein determining the plurality of top-level profile decisions comprises: determining a section division of a plurality of segments of the plurality of segments, and A representation is determined that includes a single interval partition representing the interval partitions of the plurality of segments.
9. The method of claim 8, wherein: the representation comprises a plurality of intervals and, Each interval listing a top level profile of a corresponding one of the plurality of segments, an A top level profile decision is determined for each interval.
10. The method of claim 1, wherein determining the performance of the target metrics for the plurality of top-level profile decisions comprises: Evaluating a probability of selecting a profile of the plurality of profiles, and Predicting a first value of the target metric based on performance of the metric using a constraint from the top-level profile decision, and predicting a second value of the target metric without using the constraint, wherein the first value and the second value are used to determine a third value of performance of the target metric for selecting the top-level profile decision.
11. The method of claim 10, wherein determining the performance of the target metrics for the plurality of top-level profile decisions comprises: Predicting performance of a second indicator using constraints from the top level profile of the top level profile decision and not using the constraints, and Predicting a fourth value of the second metric based on performance of the metric using a constraint from the top-level profile decision, and predicting a fifth value of the second metric without using the constraint, wherein the fourth value and the fifth value are used to determine a sixth value of performance of the second metric for selecting the top-level profile decision.
12. The method of claim 11, wherein the top level profile decision is selected using the target metrics based on the second metrics.
13. The method of claim 1, wherein selecting the top level profile decision comprises: And comparing target values of the target indexes of corresponding top-level profile decisions in the plurality of top-level profile decisions to select the top-level profile decision.
14. The method of claim 13, wherein the top-level profile decision comprises a target value of the target metric that is closer to the target metric than another top-level profile decision of the plurality of top-level profile decisions.
15. The method of claim 13, wherein comparing the target value of the target indicator comprises: the top level profile decision is selected using the target value of the second indicator.
16. The method of claim 1, wherein determining the target quality using the top-level profile decision comprises: A target quality value is selected based on an interval of quality values associated with the top level profile decision.
17. The method of claim 1, wherein a top-level profile decision of the plurality of top-level profile decisions is associated with an interval of different quality values.
18. A non-transitory computer-readable storage medium having stored thereon computer-executable instructions that, when executed by a computing device, enable the computing device to: determining a content instance to transmit to a client device, wherein the content instance includes a plurality of segments associated with a plurality of profiles; Determining a target index and bandwidth distribution; Determining a plurality of top level profile decisions for the plurality of segments, wherein a top level profile decision lists top level profiles that can be selected for respective ones of the plurality of segments; determining, based on the bandwidth distribution, a performance of the target metrics for the plurality of top-level profile decisions; Selecting a top-level profile decision from the plurality of top-level profile decisions based on performance of the target metrics of the plurality of top-level profile decisions, and A target quality is determined using the top level profile decision, wherein the target quality is used to select a top level profile for a segment of the plurality of segments requested during playback of the content instance.
19. The non-transitory computer-readable storage medium of claim 18, wherein determining the performance of the target metrics for the plurality of top-level profile decisions comprises: Evaluating a probability of selecting a profile of the plurality of profiles, and Predicting a first value of the target metric based on performance of the metric using a constraint from the top-level profile decision, and predicting a second value of the target metric without using the constraint, wherein the first value and the second value are used to determine a third value of performance of the target metric for selecting the top-level profile decision.
20. An apparatus, comprising: one or more computer processors, and A computer-readable storage medium comprising instructions for controlling the one or more computer processors to: determining a content instance to transmit to a client device, wherein the content instance includes a plurality of segments associated with a plurality of profiles; Determining a target index and bandwidth distribution; Determining a plurality of top level profile decisions for the plurality of segments, wherein a top level profile decision lists top level profiles that can be selected for respective ones of the plurality of segments; determining, based on the bandwidth distribution, a performance of the target metrics for the plurality of top-level profile decisions; Selecting a top-level profile decision from the plurality of top-level profile decisions based on performance of the target metrics of the plurality of top-level profile decisions, and A target quality is determined using the top level profile decision, wherein the target quality is used to select a top level profile for a segment of the plurality of segments requested during playback of the content instance.

Description

Quality-aware profile filtering in content delivery Technical Field Embodiments relate generally to content delivery. Background In adaptive bitrate streaming, content is encoded into multiple profiles with different bitrates or qualities or with different bitrates and qualities. The media player may then adaptively switch between different profiles during playback according to certain criteria, such as available bandwidth. However, not all configuration files are necessary for all types of devices. For example, on some devices, such as smartphones with limited computing resources or smaller screens, the user may not notice the quality improvement resulting from selecting a very high bit rate profile. However, providing a very high bit rate may lead to a higher risk of rebuffering, i.e. the available content is insufficient to meet the playing rate of the media player, and may also increase the user cost due to transmitting cellular data, or increase the company cost of transmitting data. Disclosure of Invention In some embodiments, a method determines a content instance to transmit to a client device. The content includes segments associated with a plurality of profiles. A target index and a bandwidth distribution are determined. The method determines a plurality of top level profile decisions for a plurality of segments, wherein a top level profile decision lists top level profiles selectable for corresponding ones of the plurality of segments. Based on the bandwidth distribution, performance of the target metrics is determined for the plurality of top level profile decisions. The method selects a top level profile decision from the top level profile decisions based on the performance of the target metrics of the top level profile decisions. The top level profile decision is used to determine a target quality. During content playback, the target quality is used to select a top level profile of a clip in the requested clip. Drawings The drawings are included for illustrative purposes only and are provided to provide examples of the possible structure and operation of the disclosed inventive systems, apparatus, methods, and computer program products. The figures in no way limit any changes in form and detail that may be made by those skilled in the art without departing from the spirit and scope of the disclosed embodiments. FIG. 1 depicts a simplified system for performing profile filtering in accordance with certain embodiments. FIG. 2 depicts a simplified flow diagram of a method of determining top level profile decisions, according to some embodiments. Fig. 3A depicts an example of a video quality interval in accordance with some embodiments. FIG. 3B depicts an example of a simplified representation according to some embodiments. FIG. 4 depicts a simplified flow diagram of a method of predicting performance of top level profile decisions, according to some embodiments. Fig. 5 depicts a simplified flow diagram of a method of selecting a target video quality, in accordance with some embodiments. FIG. 6 depicts one example of a computing device according to some embodiments. Detailed Description Techniques for content delivery systems are described herein. In the following description, for purposes of explanation, numerous examples and specific details are set forth in order to provide a thorough understanding of some embodiments. Some embodiments defined by the claims may include only some or all of the features in these examples, or be combined with other features described below, and may further include modifications and equivalents of the features and concepts described herein. Overview of the System In some embodiments, the system uses a profile filtering process to exclude some of the profiles (profiles) available to the client device. Content, such as video, audio, video and audio, or other multimedia content, may be encoded in a plurality of profiles that correspond to different levels, which may be different levels of bit rate, quality, or both. These profiles may be included in a profile ladder (profileladder). The system may receive adjustable parameter settings that meet specific objectives, such as quality of service objectives. In some embodiments, the system may use the goal of maximum tolerable quality loss or minimum required data saving ratio, but other goals may be used depending on the performance metrics used. The value of the target may also vary, for example, absolute values or percentages may be used. The maximum tolerable quality loss may be the maximum amount of quality lost by removing the profile from the profile ladder. The minimum required data saving ratio may be a minimum data saving ratio based on data saved by removing one of the configuration files. These parameters may be adjusted and based on these values, the process may analyze the configuration files of the plurality of clips to select a target video quality for the content. The target video quality may be sent to th