High quality text-to-speech converter

Question

I'm looking for a text-to-speech converter, that

• produces a wave or mp3 file (no real-time needed)
• has no fundamental restrictions on commercial usage of the produced output

• comes with female / male speakers for at least

  • English and
  • German language

• puts output quality (pronunciation, tone and intonation) over ease of use (third-party product integration, nice GUI, etc.)

• is gratis (in the ideal case)

There is no need for a whole integrated (and expensive) software suite.

Preferred environments (in descending order): JVM (binaries) > Windows > Linux console > web app

Answer 1

You could take a look at IVONA Text to Speech. It's not free but it does meet all of your other requirements:

• Multiple voices (with additional voices available)
• Multiple languages (with additional languages available)
  • American English
  • British English
  • German
  • Spanish
  • French
  • Italian
• High quality (from my experience* anyway)

* One thing I will note, is that I have only used the IVONA Text to Speech voices for Android and I am basing my assessment of the quality level of the voices off this.

The software is available via their website and comes in a variety of shapes and sizes: for Windows, Android, as APIs for developers, as a Speech Server (for IVRs and the like), etc.

Answer 2

I found something: CoquiTTS (web demo)

There are multiple models, especially German (tts_models/de/thorsten/vits) and English (tts_models/en/ljspeech/vits) ones.

Some of the models have multiple speakers (see Python code below to get the details)

import sys


def gen(model_name: str, text: str, out_file: str) -> None:
    from TTS.api import TTS  # pip install TTS

    tts = TTS(model_name=model_name, progress_bar=False, gpu=False)
    # if you want to see which models are available: TTS.list_models()
    print("#" * 80)
    print(f"model={model_name}")
    print(f"speakers={tts.speakers}")
    print(f"languages={tts.languages}")
    speaker = tts.speakers[0] if tts.speakers else None
    language = "en" if tts.languages else None
    tts.tts_to_file(text=text, speaker=speaker, language=language, file_path=out_file)


def get_text(filepath: str) -> str:
    with open(filepath) as fp:
        text = fp.read()
    text = text.replace("”", '"').replace("“", '"')
    return text


if __name__ == "__main__":
    gen(
        model_name="tts_models/en/ljspeech/vits--neon",
        text=get_text(sys.argv[1]),
        out_file=sys.argv[2]
    )