media 0.1.5

A Media Toolkit. Text-to-speech is currently available.

Media

A Media Toolkit.

Text-to-speech is currently available.

Other features will be developed in the future.

Thanks for Microsoft Azure!

Release Note

• 0.1.4 - 0.1.5: Fixed some issues.

• 0.1.3:

  • Added phonemes to improve pronunciation.
  • Updated the documentation.

• 0.1.2:

  • Added a link to the official site (Azure) to get the key.
  • Fixed some issues.

• 0.1.1: Text-to-speech is currently available.

Quick Start

Get the keys

See "Get the keys for your resource - Azure"

Text-to-speech

from media import Voice

text = "One Piece! Dragon Ball! Doraemon! Naruto!"
voice = Voice("YourSubscriptionKey", "YourServiceRegion")
# use English
voice.speak(text)
# use Japanese
voice.speak(text, lang=Voice.LANG.JA_JP, voice_name=Voice.NAME.FEMALE.JA_JP_NANAMI)
# Save the voice file to the local
voice.save(text)

Use phonemes to improve pronunciation

from media import Voice, SSML

voice = Voice("YourSubscriptionKey", "YourServiceRegion")
ssml = SSML()
# If you want to modify the phoneme of a word, add '[]' in text
ssml.voice = {
    "text": "His name is Mike [Zhou]",
    "phonemes": [{
        "word": "Zhou",
        "alphabet": "ups",  # default: sapi
        "ph": "JH AU",
    }]
}
voice.speak(ssml)

The values ​​of alphabet and ph can refer to here, see: "Use phonemes to improve pronunciation - Azure"

Error detection

success = voice.speak()
if not success:
    # do something...
    print(voice.error)
    # do something...

View the generated XML for SSML

from media import SSML

ssml = SSML()
# for human
print(ssml)
# for program
ssml.dump()

Full Example

Text-to-speech, in a different tone.

from media import Voice, SSML

voice = Voice("YourSubscriptionKey", "YourServiceRegion")
ssml = SSML(lang=SSML.LANG.ZH_CN, voice_name=SSML.NAME.FEMALE.ZH_CN_XIAO_XUAN)
ssml.voice = "啊？"
ssml.voice = {
    "text": "这是可以说的吗？",
    "role": SSML.ROLE.YOUNG_ADULT_FEMALE,
    "style": SSML.STYLE.CHEERFUL,
    "rate": SSML.RATE.MEDIUM,
}
ssml.voice = {
    "text": "啊，可以可以",
    "name": SSML.NAME.FEMALE.ZH_CN_XIAO_MO,
    "style": SSML.STYLE.FEARFUL,
    "role": SSML.ROLE.OLDER_ADULT_FEMALE,
    "degree": "2",
}
ssml.voice = {
    "text": "没事没事",
    "name": SSML.NAME.FEMALE.ZH_CN_XIAO_MO,
    "style": SSML.STYLE.SAD,
    "role": SSML.ROLE.OLDER_ADULT_FEMALE,
    "degree": "2",
    "rate": SSML.RATE.FAST,
}
# It will play the generated speech
voice.speak(ssml)

If you want to save:

voice.save(ssml)

If you want to save and customize the name or location:

voice.save(ssml, path="这是可以说的吗.mp3")