[Python] Coqui TTS: No espeak backend found. Install espeak-ng or espeak to your system

I have a test script for Coqui TTS under Windows. I moved the installation folder to the project directory and set the PATH variables.



import os
import torch
from TTS.api import TTS
from TTS.utils.manage import ModelManager

# Get device
device = "cuda" if torch.cuda.is_available() else "cpu"

espeak_path = r"eSpeakNG"

# Add eSpeak to PATH
os.environ["PATH"] = espeak_path + os.pathsep + os.environ.get("PATH", "")
os.environ["ESPEAK_DATA_PATH"] = os.path.join(espeak_path, "espeak-ng-data")

# Show models
manager = ModelManager()
models = manager.list_models()
print(models)

# Initialize TTS-Model
tts = TTS(model_name="tts_models/de/thorsten/tacotron2-DDC")
# Save tts
tts.tts_to_file(text="Das ist ein Test.", file_path="output.wav")


I always get the error message

File "C:\Python311\Lib\site-packages\TTS\tts\utils\text\phonemizers\espeak_wrapper.py", line 114, in __init__
raise Exception(" [!] No espeak backend found. Install espeak-ng or espeak to your system.")


Exception: [!] No espeak backend found. Install espeak-ng or espeak to your system.

I tried this code

import subprocess

espeak_path = r"eSpeakNG"
os.environ["PATH"] = espeak_path + os.pathsep + os.environ.get("PATH", "")
os.environ["ESPEAK_DATA_PATH"] = os.path.join(espeak_path, "espeak-ng-data")

try:
result = subprocess.run(
["espeak-ng.exe", "--version"],
capture_output=True,
text=True,
)
print("eSpeak NG output:", result.stdout)
except FileNotFoundError:
print("espeak-ng.exe not found.")


and the output is eSpeak NG output: eSpeak NG text-to-speech: 1.52-dev Data at: eSpeakNG\espeak-ng-data, so it is set correct, right? I also tried this code to ensure the Dll can be loaded:

import ctypes

try:
ctypes.CDLL(r"eSpeakNG\libespeak-ng.dll")
print("libespeak-ng.dll successfully loaded.")
except OSError as e:
print("Error loading libespeak-ng.dll:", e)


Is there anything Coqui uses I forgot to add to the PATH? Thank you.

Continue reading...