TextToSpeach:TTS

参考文献

OpenAI: speach

• https://platform.openai.com/docs/guides/speech-to-text

AWS: AWS Polly

• https://aws.amazon.com/jp/polly/

Google: gTTS (Google Text-to-Speech)

• https://cloud.google.com/text-to-speech?hl=ja

SpeechT5

• https://huggingface.co/microsoft/speecht5_tts

LISTEN

• https://listen.style/p/listennews/dkcjirkq

環境構築手順

1. リポジトリのクローン

git clone xxx

2. asdfでpythonのバージョン管理

asdf install

3. pythonの仮想環境

python -m venv .venv && source .venv/bin/activate && pip install -r requirements.txt

4. スクリプトの実行

python tts-openai.py text/sample-text-jp.txt

$ python tts-openai.py -h                                                                                                    254 ↵
usage: tts-openai.py [-h] [--model {tts-1,tts-1-hd}] text_file_path

Text-to-Speech script using OpenAI's API.

positional arguments:
  text_file_path        Path to the text file.

options:
  -h, --help            show this help message and exit
  --model {tts-1,tts-1-hd}
                        Model to use for text-to-speech. Choices are tts-1 or tts-1-hd. Default is tts-1.

Usage example: python tts-openai.py example.txt --model tts-1-hd

ffprobe mp3 mp3/sample-text-jp.mp3

文字数計算

wc -m text/report_XXXX.txt