1.4.0
Features:
- #13: add support for microphone as input. Actually, it has been implicitly supported from day one - just never tested nor documented;
- #14: add support for
--action
option allowing to select TV action betweenmute
andlower_volume
(new feature). Mute action remains a default; - #14: add support for
--volume
option allowing to specify target volume for thelower_volume
action. See documentation for volume specifier formats; - #10: add support for
--mute_timeout
option allowing to specify in seconds a TV action timeout, after which TV volume is restored to its original value. The default is 600 s (10 mins);0
disables timeout; - #37: incorporate results of Dejavu tuning, providing for about two times better recognition in expense of increased CPU load. This feature requires a new Dejavu release 0.1.3-ds1.3.0 and regeneration of jingles database (v20220909 or later). As a consequence, the default for hit confidence
--rec_confidence
is restored from 5% to 10%; - #40: treat jingles without entry/exit flags as non-actionable instead of always actionable;
- various internal refactoring.