US 12,294,771 B1
Generating and evaluating insertion markers in media
Kalpesh Bhagwandas Thavrani, Seattle, WA (US); Bharanidharan Arul Janakiammal, Seattle, WA (US); Pranav Kapoor, Lynnwood, WA (US); Caileigh Ann Robertson Halbach, Berlin (DE); Abhinaw Dixit, Redmond, WA (US); and Durga Prasad Guntoju, Snohomish, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Jun. 30, 2022, as Appl. No. 17/854,394.
Int. Cl. H04N 21/845 (2011.01); H04N 21/234 (2011.01); H04N 21/431 (2011.01)
CPC H04N 21/8455 (2013.01) [H04N 21/234 (2013.01); H04N 21/4316 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
identifying at least a first portion of first audio data having an audio amplitude below a predetermined level for a predetermined duration;
identifying at least a second portion of the first audio data in which an energy content in at least one frequency band of the first audio data changes by more than a predetermined amount in less than a predetermined amount of time;
determining, based on identifying at least the first portion and the second portion, a first plurality of insertion markers representing points in the first audio data for potentially inserting supplemental audio data for playback in coordination with the first audio data;
selecting a first insertion marker of the first plurality of insertion markers using one or more of a random or pseudorandom selection;
determining second audio data for insertion at the first insertion marker;
causing a first user device to output the first audio data;
causing the first user device to output the second audio data when playback of the first audio data reaches the first insertion marker;
receiving first data representing user behavior associated with the first insertion marker;
determining, using at least the first data, that a second insertion marker of the first plurality of insertion markers has an improved efficacy compared to the first insertion marker;
causing a second user device to output the first audio data; and
causing the second user device to output third audio data when playback of the first audio data reaches the second insertion marker.