Releases: zh-plus/openlrc
Releases · zh-plus/openlrc
v1.0.1
v1.0.0
v0.2.3
v0.2.2
v0.2.1
This update adds a preprocessor to enhance input audio (loudness normalization & noise suppression).
New features:
- Loudness Normalization from ffmpeg-normalize
- Noise Suppression from DeepFilterNet
Other Changes:
- Now all the intermediate files are saved in
./path/to/audio/preprocess
.
v0.2.0
This update switch the underlying transcription model from whisperx
to faster-whisper
, which enable VAD parameter tuning.
New features:
- Switch whisperx back to faster-whisper for VAD parameter tuning.
Other Changes:
- Update translation prompt
from machinewrapped/gpt-subtrans@82bd2ca. - Change the default parameters of
faster-whisper
for consistent transcription.
v0.1.4
This update fixes minor bugs and introduces sentence_split
to reduce subtitle length.
New features:
- Add
word_align
andsentence_split
for non-word-boundary languages to split long text into sentences. - Add text-normalization for help matching sentences.
- Add skip-translation support.
Other Changes:
- Use
pathlib
to handle paths. - Improve timeline accuracy.
v0.1.3
This update add input video support and introduce context configuration.
New features:
- Add input video support.
- Add context configuration for inputs.
Other Changes:
- Add test suites to CI.
- Add language detection for translated content.
- Improve prompt by adding background info.
- Update punctuator model.
- Replace
opencc
with more light-weightzhconv
.
v0.1.2
This update improves the timeline consistency of translated subtitles. Thanks gpt-subtrans!
New features:
- Fix misaligned timeline issue by improving translation prompt.
- Add output srt format support.
- Add changeable temperature and top_p parameter for GPTBot.
Other Changes:
- Report total OpenAI translation fee for multiple audios.
- Improve repeat-checking algorithm.
v0.1.1
This update enhances the efficiency of processing multiple audio files.
New features:
- Implementation of a producer-consumer model to process multiple audio files.
Other Changes:
- Update logger with colored format.
- Minor parameter modification that makes the timeline of translation more intuitive.