It should be noted that a value for a threshold may be independent of the value of encoding parameters. For example, all trial encoding outputs with an encoding runtime per frame exceeding a certain threshold are pruned, regardless of the value of the maximum bit rate encoding parameter, or the value of any other encoding parameter in the encoding profile.

Changing an encoding parameter value can simultaneously affect multiple data space dimensions, such as bitrate, encoding speed, and video quality. Thus, following the data space pruning process, the optimal selection of encoding parameter profiles, which may be part of analyzing the results of the trial encodings (308), for an event-adaptive encoding ladder is a multi-dimensional optimization problem. The optimization involves accounting for tradeoffs along multiple dimensions (e.g., bitrate, encoding speed, and video quality) based on a cost function.

A cost function can be tailored to particular streaming video service system design goals. Cost function choices include but are not limited to (1) minimizing bit rate for a given quality constraint, (2) maximizing quality for a given bit rate constraint, (3) minimizing bit rate given quality and encoding runtime constraints, or (4) maximizing delivered quality (measured at the playback devices) given, for example, bit rate, network bandwidth, and playback device constraints.