I have a problem with Coqui TTS API

[AliA6565]

Hello, I have a porblem with Coqui TTS API, I tried to put it to my chatbot, but every time it's generating a voice, I get this error: Error with Coqui TTS: Expected tuple of length 2. Received: None

[eginhard]

What's the command you're trying to run and what's the full error message?

[AliA6565]

import os
import time
import pygame
import speech_recognition as sr
from gtts import gTTS
from dotenv import load_dotenv
from gradio_client import Client

Initialize Pygame mixer for audio playback

pygame.mixer.init()

Load the environment variables for RVC

load_dotenv()

Load the LLM (Qwen API) for chat responses

qwen_client = Client("https://qwen-qwen2-5-72b-instruct.hf.space/")

Function to select microphone

def select_microphone():
r = sr.Recognizer()
mic_list = sr.Microphone.list_microphone_names()

print("Available microphones:")
for i, mic_name in enumerate(mic_list):
    print(f"{i}: {mic_name}")

mic_index = int(input("Enter the number of the microphone you want to use: "))
return mic_index

Function to transcribe speech

def transcribe_speech(mic_index):
r = sr.Recognizer()
with sr.Microphone(device_index=mic_index) as source:
r.adjust_for_ambient_noise(source, duration=1)
print("Listening... please speak.")
audio = r.listen(source)

try:
    print("Transcribing audio...")
    return r.recognize_google(audio)
except sr.UnknownValueError:
    return "Sorry, I couldn't understand."
except sr.RequestError as e:
    return f"Error with the service; {e}"

Function to generate response from the Qwen API

def generate_response(user_input):
result = qwen_client.predict(
query=user_input,
history=[],
system="You are Hatsune Miku, a friendly virtual assistant who loves music and chatting with friends. Please reply sweetly and gently!",
api_name="/model_chat"
)
return result[1][0][1]

Function to use GTTS for text-to-speech and run through Coqui TTS

def coqui_tts(text, reference_audio_url):
coqui_client = Client("https://coqui-xtts.hf.space/--replicas/mvdvr/")
print(f"Sending text to Coqui TTS API: {text}")
print(f"Using reference audio URL: {reference_audio_url}")
try:
result = coqui_client.predict(
text, # Text to synthesize
"en,en", # Language
reference_audio_url, # Reference audio URL
reference_audio_url, # Reference mic audio (using the same URL here)
True, # Use microphone
True, # Cleanup reference voice
True, # Disable language auto-detect
True, # Agree to terms
fn_index=1
)
print("TTS result:", result)
except Exception as e:
print(f"Error with Coqui TTS: {e}")

Main chatbot loop

def chatbot():
mic_index = select_microphone()
reference_audio_url = "https://raw.githubusercontent.com/AliA6565/audio/main/1.wav"

print("Welcome! I'm Hatsune Miku, your virtual friend. Let's chat! 😊")

while True:
    user_input = transcribe_speech(mic_index).strip()
    print(f"You said: {user_input}")

    if user_input.lower() in ['exit', 'bye', 'goodbye']:
        print("Hatsune Miku: Goodbye! It was nice chatting with you! 🎶")
        break

    # Generate response from Qwen API
    ai_response = generate_response(user_input)
    print(f"Hatsune Miku is responding: {ai_response}")

    # Convert text to speech using Coqui TTS with voice cloning from GitHub-hosted reference audio
    coqui_tts(ai_response, reference_audio_url)

if name == "main":
chatbot()

[eginhard]

Ok, so you're not actually using this library, so I can't help you. If you want to use Coqui TTS from Python, see the docs: https://coqui-tts.readthedocs.io/en/latest/inference.html#python-tts-api

[AliA6565]

thank you

[AliA6565]

Thanks