alekssamos / msspeech Goto Github PK

View Code? Open in Web Editor NEW

55.0 3.0 10.0 429 KB

not official API for Microsoft speech synthesis from Microsoft Edge web browser read aloud

Home Page: https://pypi.org/project/msspeech/

Python 95.47% Makefile 4.53%

speech-synthesis speech-synthesis-api speech-synthesis-library tts-api

msspeech's Introduction

msspeech

not official API for Microsoft speech synthesis from Microsoft Edge web browser read aloud

Installation

pip install --upgrade msspeech

poetry add msspeech

After updating an already installed library To update the list of voices, run the command in your terminal:

msspeech-update-voices

poetry run msspeech-update-voices

Notes

Bad news

Since the first of July 2022, the list of voices and the API as a whole has been very much limited!

But there is also good news

They returned back some male voices and added new languages, as well as made support for emotional styles. Despite the fact that styles appeared in JSON, you still won't be able to use them, SSML does not perceive them. SSML is very limited here, so there is no point in supporting it.

The official documentation is not suitable for this API. It seems this API uses undocumented SSML markup.

https://docs.microsoft.com/ru-ru/azure/cognitive-services/speech-service/language-support#text-to-speech

Using

the pitch and rate values are set as a percentage from -100 to +100, that is, it can be a negative, positive number, or zero for the default value.

examples: -30, 40, 0

The volume should be a fractional number from 0.1 to 1.0, but in fact it doesn't work for some reason.

The maximum synthesize text length is approximately 31000 characters per request.

from CLI

synthesize text:

msspeech Guy hello --filename audio.mp3

update voices list:

msspeech-update-voices

From python

import asyncio
from msspeech import MSSpeech


async def main():
	mss = MSSpeech()
	print("Geting voices...")
	voices = await mss.get_voices_list()
	print("searching Russian voice...")
	for voice in voices:
		if voice["Locale"] == "ru-RU":
			print("Russian voice found:", voice["FriendlyName"])
			await mss.set_voice(voice["Name"])


	print("*" * 10)
	filename = "audio.mp3"
	# with open("s.txt", encoding="UTF8") as f: text:str = f.read()
	text = "Или написать текст здесь"
	print("waiting...")
	await mss.set_rate(10)
	await mss.set_pitch(0)
	await mss.set_volume(1.0)
	await mss.synthesize(text.strip(), filename)
	print("*"*10)
	print("SUCCESS! OK!")
	print("*"*10)

if __name__ == "__main__":
	asyncio.run(main())

msspeech's People

Contributors

Stargazers

Watchers

Forkers

ishine xyx208 homebrew-startup pendave gdtiti faithless35 18639522576 1zero8 leos-code pixu2019

msspeech's Issues

Does the CLI support reading text from a file?

There is a text file containing several sentences. When specifying this file through the CLI, multiple audio files will be outputted, with each sentence separated by a newline character, and each line representing one audio.

I can't use many new narators but exist on microsoft.

https://learn.microsoft.com/en-us/azure/cognitive-services/Speech-Service/language-support?tabs=tts
for voice in voices: if voice["Locale"] == "zh-CN": print("Chinese voices found: ", voice["ShortName"]) await mss.set_voice(voice["Name"])

I can't use:
such as
zh-CN-henan
zh-CN-liaoning
.........

Option to use the Narrator voices?

In new Windows 11 builds, Narrator can use the exact same voices that are available in Edge, but offline, and seemingly more customizable. These voices install as store apps, so we don't need an internet connection either, like you seem to in order to use the Edge voices. Is it possible to get these integrated into this project?

(problem and solution) save the audio in data buffer

@alekssamos Thank you very much for this tool, I barely know it and it is perfect for some projects I am working on.

Problem:
When trying to save the audio file in a data buffer, the following error is displayed

msspeech/__init__.py, line 520, in _synthesize

     bc += await f.write(resp[1])

TypeError: object int can't be used in 'await' expression

My code

file = BytesIO()
await mss.synthesize(data.text.strip(), file)
file.seek(0)

Solution:
remove the await and it worked for me.

Doesn't work with SoundLoader (kivy)

I already tried to save the file as .mp3, .wav, .ogg, doesn't matter, kivy isn't able to load properly.
b'Unrecognized audio format'
from kivy.app import App
from kivy.core.audio import SoundLoader

class MusicPlayerApp(App):
def build(self):
# load the mp3 file using SoundLoader
sound = SoundLoader.load(R"D:\Documents\aaaa.ogg")
# play the audio file
sound.play()
return

if name == 'main':
MusicPlayerApp().run()

the voices_list_plus.json is outdated, how can I update it myself?

Hello I find the voices_list_plus.json is outdated, as I can't find this voice for example:
zh-CN-YunjianNeural

https://raw.githubusercontent.com/alekssamos/msspeech/da5904e14e7e8b383f4230c57585eaa92271a4bb/msspeech/voices_list_plus.json

:)
How can I update it myself?

And where can I add custom SSML?