Commit Graph

28 Commits

Author SHA1 Message Date
Daniel Wolf c821788569 Discard very short segments of voice activity
This prevents short flickers from false VAD positives.
This fixes a regression recently introduced in 2bbad258c0.
2019-01-23 21:43:17 +01:00
Daniel Wolf 21392d32cc Improve read performance for WAVE files
Fixes #58
2019-01-23 21:43:16 +01:00
Daniel Wolf f3d4cfbb31 Fix gaps in phonetic recognition
Randomly, entire utterances yielded no phones with the phonetic recognizer.
The cause was a check for empty utterances that made sense for word
recognition, but not for phonetic recognition.
2019-01-21 22:22:28 +01:00
Daniel Wolf 9ba5fc9653 Fix special characters in output file path
Issue #54
2019-01-04 21:29:03 +01:00
Daniel Wolf 2bbad258c0 Do not use multithreading for VAD
WebRTC adapts to the audio signal. If we slice the audio clip into multiple
shorter clips, then perform VAD on them in parallel, the result may not be as
good.
2019-01-04 21:02:41 +01:00
Daniel Wolf a723942f22 Convert audio to 8kHz before feeding it to WebRTC for VAC
This prevents false positives at the beginning of the audio stream. Fixes issue #52.
2019-01-04 21:02:41 +01:00
Daniel Wolf 1393534624 Prevent utteranceToPhones from reporting incomplete progress 2019-01-03 13:10:05 +01:00
Daniel Wolf 494b1ebebd Log incomplete tasks 2019-01-03 13:10:05 +01:00
Daniel Wolf 367c645bb3 Code cleanup
* Fix linter warnings
* Unify code formatting
* Fix typos
2019-01-02 22:27:45 +01:00
Daniel Wolf cdc9c5b1c4 Upgrade to utf8proc 2.2.0
Fixes issue #38
2019-01-02 15:18:00 +01:00
Daniel Wolf e87ccc8816 Hide progress spinner when ProgressBar gets destructed 2019-01-02 14:38:53 +01:00
Daniel Wolf 3189fc8976 Don't clear away last progress bar 2019-01-02 14:30:03 +01:00
Daniel Wolf e325917abe Preserve progress when continuing console output 2019-01-02 14:30:02 +01:00
Daniel Wolf d0c9a294e9 Correctly print final progress when ProgressBar is destructed
Fixes issue #46
2019-01-02 14:30:02 +01:00
Daniel Wolf 11a29ddef6 Split progress code 2019-01-02 13:22:03 +01:00
Daniel Wolf 610f490046 Implement generic concept of recognizers with options pocketSphinx and phonetic 2019-01-01 22:51:01 +01:00
Daniel Wolf 3ed38ada2f Fix path separator 2018-10-08 20:31:13 +02:00
Daniel Wolf 8648ed5640
Merge pull request #44 from DanielSWolf/feature/#40-ogg-vorbis
Support for Ogg Vorbis file format
2018-09-01 10:34:54 +02:00
Daniel Wolf c22550f7f8 Support Unicode paths when opening Ogg Vorbis files 2018-09-01 10:24:05 +02:00
Daniel Wolf 5e7e6f5f87 Increase Ogg Vorbis read performance through buffering
Reduces runtime by factor 10
2018-09-01 10:24:05 +02:00
Daniel Wolf e13c222e28 Add support for Ogg Vorbis file format
#40
2018-09-01 10:24:05 +02:00
Daniel Wolf 1625de64e2 Move file utilities into their own file 2018-09-01 10:24:04 +02:00
Daniel Wolf d077aae74c Add libvorbis 2018-09-01 10:23:58 +02:00
Daniel Wolf b4a2a295d1 Patch libogg for generic Unix compilers 2018-09-01 10:23:53 +02:00
argent0 e02de085bb Fix build error resulting from assigning boost::uuids::random_generator (#41)
Using a static generator to generate uuids

This time explicitly writing the type.
2018-07-19 20:19:38 +02:00
Daniel Wolf a446209c31 Add libogg 2018-07-13 22:59:47 +02:00
Daniel Wolf 1c85dbfc3a Added ReSharper C# settings 2018-02-02 21:33:18 +01:00
Daniel Wolf e5c39efeeb Moved main executable into its own directory along with its build logic
This will make it easier to add other artifacts that require build steps.
2018-01-24 18:38:09 +01:00