Leveraging media quality scores with AWS Elemental MediaPackage

Streaming media quality is influenced by many factors. To ensure the highest quality viewing experience, services in the video streaming workflow must have awareness of, and ability to act on, indicators of quality down to the segment level in a stream. MediaPackage uses media quality confidence score (MQCS) functionality to receive, act on, calculate, and communicate quality scores of live streaming content. When adopted into your entire workflow, MQCS provides an automated operation to ensure that you provide the optimum viewing experience to your customers.

The following sections describe how MediaPackage leverages media quality scores from stream inputs and through origin outputs to continuously monitor video quality, correct for errors in the live streaming workflow, and help faciliate debugging through increased visibility.

To enable MQCS with MediaPackage, select the MQCS settings on your CMAF channels. For more information about channel settings, see Creating a channel in AWS Elemental MediaPackage.

How MQCS works

The following is a high-level explanation of how MediaPackage uses quality scores from its inputs and in its outputs. For a more detailed explanation about quality scores and how to set up your quality-aware workflow, see the blog Improve your viewers' live streaming experience with Media Quality-Aware Resiliency.

MQCS with inputs to MediaPackage

MQCS in input streams to MediaPackage builds on live input redundancy functionality. With input redundancy, MediaPackage receives two input streams to a channel. If the active stream is missing any segments, MediaPackage automatically switches to the other stream. MediaPackage continually monitors the streams and switches between them as needed based on segment availability.

With MQCS, the upstream encoder (MediaLive) calculates a quality score for the stream at the ABR-level, based on the health of the source bitstream or elementary, error recovery, and segment errors. The encoder communicates this score to MediaPackage. When the input streams have identical health (segment availability and latency), MediaPackage determines which stream to use for source content based on which has the highest quality score. MediaPackage can effectively provide dynamic segment replacement as it receives content from the encoder.

You can also use MQCS to aid in Cross-region failover when you enable MediaPackage to include quality scores in responses to requests from downstream systems, such as CDNs (CloudFront). The next section describes how MQCS works with outputs from MediaPackage.

MQCS with outputs from MediaPackage

As MediaPackage receives streams from the encoder, it re-calculates the quality score for each segment and track type. It also calculates an average quaity score for all segments in the media sequence. MediaPackage communicates this score through CMSD headers in its responses to downstream systems, such as CDNs (CloudFront) and other quality-monitoring systems.

Downstream systems can use the quality scores from MediaPackage to determine from which origin to serve content, failing over between origins to provide the highest quality viewing experience. CloudFront uses the quality scores as an input to its media quality-aware resiliency functionality (MQAR). For information about MQAR, see Media quality-aware resiliency in the Amazon CloudFront Developer Guide.

Requirements for using MQCS

Your workflow must meet these requirements to use MQCS: