Method for generating video- and audience-specific encoding ladders

A method comprising:

• extracting a set of video features representing properties of a video segment;
• generating a set of bitrate-resolution pairs based on the set of video features, each bitrate-resolution pair in the set of bitrate-resolution pairs:
• defining a bitrate;
• and
• defining a resolution predicted to yield a quality score, approximating a maximum quality score, characterizing the video segment encoded at the bitrate and the resolution;
• accessing a distribution of audience bandwidths representative of likely viewers of the video segment;
• and
• for each bitrate-resolution pair in the set of bitrate-resolution pairs:
• calculating a rendition cost of the bitrate-resolution pair, the rendition cost representing a cost of a rendition of the video segment corresponding to the bitrate-resolution pair;
• calculating a marginal quality score of the bitrate-resolution pair based on the distribution of audience bandwidths and the current set of bitrate-resolution pairs in the encoding ladder, the marginal quality score representing an aggregate improvement in viewing quality according to the audience bandwidth distribution in response to an addition of the bitrate-resolution pair to the encoding ladder;
• and
• calculating a marginal quality-per-cost score of the bitrate-resolution pair based on the marginal quality score of the bitrate-resolution-pair and the rendition cost of the bitrate-resolution pair;
• and
• in response to a quality-per-cost score of at least one bitrate-resolution pair in the set of bitrate-resolution pairs exceeding a threshold quality-per-cost score:
• selecting a first successive bitrate-resolution pair in the set of bitrate-resolution pairs, the first successive bitrate resolution pair characterized by a maximum quality-per-cost score;
• removing the successive bitrate-resolution pair from the set of bitrate-resolution pairs;
• and
• appending the successive bitrate-resolution pair to the encoding ladder.

method of claim 8, wherein generating the set of bitrate-resolution pairs based on the set of video features comprises:

• for each viewing condition in a set of viewing conditions, generating a viewing-condition-specific set of bitrate-resolution pairs based on the set of video features, each bitrate-resolution pair in the viewing-condition-specific set of bitrate-resolution pairs:
• defining a bitrate;
• and
• defining a resolution predicted to yield a quality score, approximating a maximum quality score, characterizing the video segment encoded at the bitrate and the resolution and viewed according to the viewing condition;
• and
• compiling each viewing-condition specific set of bitrate-resolution pairs into the set of bitrate-resolution pairs

The method of claim 11, wherein calculating the marginal quality score of the bitrate-resolution pair for each bitrate-resolution pair in the set of bitrate-resolution pairs further comprises, for each bitrate-resolution pair in the set of bitrate-resolution pairs:

• calculating a second quality score of the lower adjacent bitrate-resolution pair estimating a viewing quality of a second rendition of the video segment corresponding to the lower adjacent bitrate-resolution pair;
• combining the second quality score with the bandwidth segment in the audience bandwidth distribution to generate a status quo aggregate quality score;
• and
• subtracting the status quo aggregate quality score from the improved aggregate quality score to calculate the marginal quality score of the bitrate-resolution pair.

A method comprising:

• extracting a set of video features representing properties of a video segment;
• generating a set of bitrate-resolution pairs based on the set of video features;
• accessing a distribution of audience bandwidths representative of likely viewers of the video segment;
• and
• for each bitrate-resolution pair in the set of bitrate-resolution pairs:
• calculating an encoding cost of a rendition corresponding to the bitrate-resolution pair;
• calculating a storage cost of the rendition corresponding to the bitrate-resolution pair;
• calculating a distribution cost of the rendition corresponding to the bitrate-resolution pair;
• and
• combining the encoding cost, the storage costs, and the distribution cost to calculate the rendition cost.