| pydub>=0.25.1 | |
| librosa>=0.10.1 | |
| ffmpeg>=1.4 | |
| yt_dlp>=2024.10.22 | |
| wavio>=0.0.8 | |
| # Audio speed-up and slowdown (best quality), if not installed can only speed-up with lower quality | |
| # pyrubberband>=0.3.0 | |
| # stackoverflow.com/questions/75813603/python-working-with-sound-librosa-and-pyrubberband-conflict | |
| # pip uninstall -y pysoundfile soundfile | |
| soundfile==0.12.1 | |
| # Optional: Only for testing for now | |
| # playsound==1.3.0 | |
| # STT from microphone (may not be required if ffmpeg installed above) | |
| # for any TTS: | |
| torchaudio | |
| soundfile>=0.12.1 | |
| # GPU Only: for Coqui XTTS (ensure CUDA_HOME set and consistent with added postfix for extra-index): | |
| # relaxed versions to avoid conflicts | |
| # TTS | |
| #deepspeed | |
| noisereduce | |
| emoji | |
| ffmpeg-python | |
| trainer | |
| pysbd | |
| coqpit | |
| # for Coqui XTTS language helpers (specific versions probably not required) | |
| cutlet>=0.3.0 | |
| langid>=1.1.6 | |
| g2pkk>=0.1.2 | |
| jamo>=0.4.1 | |
| gruut[de,es,fr]>=2.2.3 | |
| jieba>=0.42.1 | |
| # librosa==0.10.1 | |
| # For faster whisper: | |
| # git+https://github.com/SYSTRAN/faster-whisper.git | |