Talkey无法切换eSpeak语音求助：葡萄牙语女声配置无效

Let me walk you through fixing this—I’ve run into similar Talkey configuration hiccups before!

First, let’s break down the potential issues and solutions step by step:

1. Verify MBROLA Voice Files Are Installed #

The brazil-mbrola-4 voice depends on the MBROLA engine and its corresponding voice package, which most systems don’t install by default. Let’s rule this out first:

• Test if eSpeak can use the voice directly from the command line:

  espeak -v brazil-mbrola-4 "Olá, esta é uma voz feminina de teste"
  

• If you get an error like Cannot find voice 'brazil-mbrola-4' or no sound, install the required packages:
  • For Debian/Ubuntu-based systems:

    sudo apt-get install mbrola mbrola-br4
    

  • For other systems, download the mb-br4 voice file from the MBROLA repository and place it in /usr/share/mbrola/ (or ~/.mbrola/ for user-specific installs).

2. Fix Talkey’s Language Code Mapping #

Talkey expects language keys that match eSpeak’s precise language codes. Your original config uses pt (generic Portuguese), but the brazil-mbrola-4 voice is tied to pt-br (Brazilian Portuguese). Let’s adjust that:

tts = talkey.Talkey(
    preferred_languages=['pt-br'],  # Use precise regional code
    engine_preference=['espeak'],
    espeak={
        'options':{
            'enabled': True,
        },
        'languages':{
            'pt-br': {  # Match the preferred language code
                'voice': 'brazil-mbrola-4'
            },
        }
    }
)

3. Add Extra Arguments to Force Voice Selection #

If the above still doesn’t work, Talkey’s internal voice mapping might be misfiring. Add direct eSpeak arguments to enforce the voice:

tts = talkey.Talkey(
    preferred_languages=['pt-br'],
    engine_preference=['espeak'],
    espeak={
        'options':{
            'enabled': True,
        },
        'languages':{
            'pt-br': {
                'voice': 'brazil-mbrola-4',
                'extra_args': ['-v', 'brazil-mbrola-4']  # Force eSpeak to use this voice
            },
        }
    }
)

Final Check #

After installing the MBROLA packages and adjusting the config, test with a simple line to confirm:

tts.say("Olá! Agora a voz deve ser feminina brasileira.")

内容的提问来源于stack exchange，提问作者Giovana Morais