US 12,469,500 B2
Methods, apparatus and systems for dual-ended media intelligence
Yanning Bai, Beijing (CN); Mark William Gerrard, Rozelle (AU); Richard Han, Beijing (CN); and Martin Wolters, Nuremberg (DE)
Assigned to Dolby Laboratories Licensing Corporation, San Francisco, CA (US); and Dolby International AB, Dublin (IE)
Appl. No. 17/312,011
Filed by DOLBY LABORATORIES LICENSING CORPORATION, San Francisco, CA (US); and DOLBY INTERNATIONAL AB, Amsterdam Zuidoost (NL)
PCT Filed Dec. 10, 2019, PCT No. PCT/US2019/065338
§ 371(c)(1), (2) Date Jun. 9, 2021,
PCT Pub. No. WO2020/123424, PCT Pub. Date Jun. 18, 2020.
Claims priority of provisional application 62/792,997, filed on Jan. 16, 2019.
Claims priority of application No. 19157080 (EP), filed on Feb. 14, 2019.
Prior Publication US 2022/0059102 A1, Feb. 24, 2022
Int. Cl. G10L 19/008 (2013.01); G06F 16/65 (2019.01); G06F 16/68 (2019.01); G10L 19/16 (2013.01)
CPC G10L 19/008 (2013.01) [G06F 16/65 (2019.01); G06F 16/68 (2019.01); G10L 19/167 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A method of decoding audio content from a bitstream comprising:
receiving the bitstream from the encoder, the bitstream including audio content and classification information comprising confidence values, each confidence value being associated with a respective content type and giving an indication of a likelihood range that the audio content is of the respective content type, where the likelihood range of the confidence value includes at least one confidence value between zero likelihood and full likelihood;
decoding the audio content and the classification information;
performing a content analysis on the decoded audio content to determine one or more additional confidence values, where the additional confidence values are different than the confidence values included in the classification information;
selecting, based on the classification information and the one or more additional confidence values, a post processing mode for performing post processing of the decoded audio content; and
calculating control weights for the post processing of the decoded audio content based on the confidence values, and wherein at least one control weight is for a virtualizer that is calculated based on at least two confidence scores or additional confidence values.