nodes/1038lab-ComfyUI-EdgeTTS
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
AI Lab
2025-06-20 10:30:39 -07:00
committed by GitHub
parent 867a5b6106
commit e4efa9d7e6
2 changed files with 243 additions and 148 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ailab_edgeTTS.py
+173 -140
View File
@@ -1,141 +1,174 @@
# ComfyUI-EdgeTTS V1.1.0
# A simplified Edge TTS node for ComfyUI
# Uses Microsoft Edge's online text-to-speech service
# Outputs standard ComfyUI audio format
import os
import edge_tts
import asyncio
import re
import torch
import torchaudio
import json
class EdgeTTS:
@staticmethod
def load_voices():
try:
config_path = os.path.join(os.path.dirname(__file__), "config.json")
with open(config_path, 'r', encoding='utf-8') as f:
config = json.load(f)
voices = []
tooltips = {}
default_voice = config.get("default_voice")
for language, voice_list in config["edge_tts_voices"].items():
for voice, description in voice_list:
voices.append(voice)
tooltips[voice] = f"{language}: {description}"
if default_voice in voices:
voices.remove(default_voice)
voices.insert(0, default_voice)
return voices, tooltips
except:
return (
["zh-CN-XiaoxiaoNeural", "en-US-JennyNeural", "ja-JP-NanamiNeural"],
{
"zh-CN-XiaoxiaoNeural": "Chinese: Female, cheerful",
"en-US-JennyNeural": "English: Female, casual",
"ja-JP-NanamiNeural": "Japanese: Female, natural"
}
)
DEFAULT_VOICES, VOICE_TOOLTIPS = load_voices.__func__()
@classmethod
def INPUT_TYPES(s):
return {
"required": {
"text": ("STRING", {"multiline": True, "placeholder": "Enter text to convert to speech"}),
"voice": (s.DEFAULT_VOICES, {"default": s.DEFAULT_VOICES[0], "tooltip": "Select a voice for text-to-speech"}),
},
"optional": {
"speed": ("FLOAT", {"default": 1.0, "min": 0.5, "max": 2.0, "step": 0.1, "tooltip": "Speech rate (0.5 to 2.0)"}),
"pitch": ("INT", {"default": 0, "min": -20, "max": 20, "step": 1, "tooltip": "Voice pitch adjustment (-20 to +20 Hz)"})
}
}
RETURN_TYPES = ("AUDIO",)
FUNCTION = "tts"
CATEGORY = "🧪AILab/🔊Audio"
async def generate_speech(self, text, voice, speed, pitch):
"""Generate speech from text using Edge TTS"""
speed_percent = int((speed - 1.0) * 100)
rate = "+0%" if speed_percent == 0 else f"{speed_percent:+d}%"
temp_file = f"temp_tts_{os.getpid()}.wav"
try:
text = text.strip()
if not text:
raise ValueError("Input text cannot be empty")
communicate = edge_tts.Communicate(
text=text,
voice=voice,
rate=rate,
pitch=f"{pitch:+d}Hz"
)
try:
await communicate.save(temp_file)
except edge_tts.exceptions.NoAudioReceived:
default_voice = self.DEFAULT_VOICES[0]
if voice != default_voice:
print(f"Warning: Failed with voice {voice}, trying default voice {default_voice}")
communicate = edge_tts.Communicate(
text=text,
voice=default_voice,
rate=rate,
pitch=f"{pitch:+d}Hz"
)
await communicate.save(temp_file)
else:
raise
waveform, sample_rate = torchaudio.load(temp_file)
if waveform.shape[0] > 1:
waveform = waveform.mean(dim=0, keepdim=True)
waveform = waveform / (waveform.abs().max() + 1e-6)
return {"waveform": waveform.unsqueeze(0), "sample_rate": sample_rate}
finally:
if os.path.exists(temp_file):
try:
os.remove(temp_file)
except:
pass
def tts(self, text, voice, speed=1.0, pitch=0):
"""Convert text to speech"""
if not text.strip():
raise ValueError("Text cannot be empty")
text = re.sub(r'\s+', ' ', text).strip()
try:
try:
loop = asyncio.get_event_loop()
except RuntimeError:
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
audio_data = loop.run_until_complete(self.generate_speech(text, voice, speed, pitch))
return (audio_data,)
except Exception as e:
print(f"TTS Error: {str(e)}")
empty_waveform = torch.zeros((1, 1, 16000))
return ({"waveform": empty_waveform, "sample_rate": 16000},)
NODE_CLASS_MAPPINGS = {
"EdgeTTS": EdgeTTS
}
NODE_DISPLAY_NAME_MAPPINGS = {
"EdgeTTS": "Edge TTS 🔊"
# ComfyUI-EdgeTTS V1.2.0
# A simplified Edge TTS node for ComfyUI
# Uses Microsoft Edge's online text-to-speech service
# Outputs standard ComfyUI audio format
import os
import edge_tts
import asyncio
import re
import torch
import torchaudio
import json
class EdgeTTS:
_voice_data_cache = None
@classmethod
def get_voice_data(cls):
if cls._voice_data_cache is None:
cls._voice_data_cache = cls.load_voices()
return cls._voice_data_cache
@staticmethod
def load_voices():
try:
config_path = os.path.join(os.path.dirname(__file__), "config.json")
with open(config_path, 'r', encoding='utf-8') as f:
config = json.load(f)
voices = []
tooltips = {}
voice_ids = {}
default_voice = config.get("default_voice")
default_display_name = None
for language, voice_list in config["edge_tts_voices"].items():
for voice, description in voice_list:
parts = voice.split('-')
if len(parts) >= 3:
lang_name = language.split('-')[0] if '-' in language else language
display_name = f"[{lang_name}] {parts[0]}-{parts[1]} {parts[-1].replace('Neural', '').replace('Multilingual', '')}"
else:
display_name = voice
voices.append(display_name)
tooltips[display_name] = f"{language}: {description}"
voice_ids[display_name] = voice
if voice == default_voice:
default_display_name = display_name
if default_display_name and default_display_name in voices:
voices.remove(default_display_name)
voices.insert(0, default_display_name)
return voices, tooltips, voice_ids
except:
return (
["[English] en-US Jenny", "[Chinese] zh-CN Xiaoxiao", "[Japanese] ja-JP Nanami"],
{"[English] en-US Jenny": "English-US: Female, casual", "[Chinese] zh-CN Xiaoxiao": "Chinese-Mainland: Female, cheerful", "[Japanese] ja-JP Nanami": "Japanese: Female, natural"},
{"[English] en-US Jenny": "en-US-JennyNeural", "[Chinese] zh-CN Xiaoxiao": "zh-CN-XiaoxiaoNeural", "[Japanese] ja-JP Nanami": "ja-JP-NanamiNeural"}
)
@property
def DEFAULT_VOICES(self):
return self.get_voice_data()[0]
@property
def VOICE_TOOLTIPS(self):
return self.get_voice_data()[1]
@property
def VOICE_IDS(self):
return self.get_voice_data()[2]
@classmethod
def INPUT_TYPES(cls):
voices, _, _ = cls.get_voice_data()
return {
"required": {
"text": ("STRING", {"multiline": True, "placeholder": "Enter text to convert to speech"}),
"voice": (voices, {"default": voices[0], "tooltip": "Select a voice for text-to-speech"}),
},
"optional": {
"speed": ("FLOAT", {"default": 1.0, "min": 0.5, "max": 2.0, "step": 0.1, "tooltip": "Speech rate (0.5 to 2.0)"}),
"pitch": ("INT", {"default": 0, "min": -20, "max": 20, "step": 1, "tooltip": "Voice pitch adjustment (-20 to +20 Hz)"})
}
}
RETURN_TYPES = ("AUDIO",)
FUNCTION = "tts"
CATEGORY = "🧪AILab/🔊Audio"
async def generate_speech(self, text, voice, speed, pitch):
"""Generate speech from text using Edge TTS"""
speed_percent = int((speed - 1.0) * 100)
rate = "+0%" if speed_percent == 0 else f"{speed_percent:+d}%"
temp_file = f"temp_tts_{os.getpid()}.wav"
try:
text = text.strip()
if not text:
raise ValueError("Input text cannot be empty")
communicate = edge_tts.Communicate(
text=text,
voice=voice,
rate=rate,
pitch=f"{pitch:+d}Hz"
)
try:
await communicate.save(temp_file)
except edge_tts.exceptions.NoAudioReceived:
default_display_name = self.DEFAULT_VOICES[0]
default_voice = self.VOICE_IDS.get(default_display_name, default_display_name)
if voice != default_voice:
print(f"Warning: Failed with voice {voice}, trying default voice {default_voice}")
communicate = edge_tts.Communicate(
text=text,
voice=default_voice,
rate=rate,
pitch=f"{pitch:+d}Hz"
)
await communicate.save(temp_file)
else:
raise
waveform, sample_rate = torchaudio.load(temp_file)
if waveform.shape[0] > 1:
waveform = waveform.mean(dim=0, keepdim=True)
waveform = waveform / (waveform.abs().max() + 1e-6)
return {"waveform": waveform.unsqueeze(0), "sample_rate": sample_rate}
finally:
if os.path.exists(temp_file):
try:
os.remove(temp_file)
except:
pass
def tts(self, text, voice, speed=1.0, pitch=0):
"""Convert text to speech"""
if not text.strip():
raise ValueError("Text cannot be empty")
text = re.sub(r'\s+', ' ', text).strip()
actual_voice = self.VOICE_IDS.get(voice, voice)
try:
try:
loop = asyncio.get_event_loop()
except RuntimeError:
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
audio_data = loop.run_until_complete(self.generate_speech(text, actual_voice, speed, pitch))
return (audio_data,)
except Exception as e:
print(f"TTS Error: {str(e)}")
empty_waveform = torch.zeros((1, 1, 16000))
return ({"waveform": empty_waveform, "sample_rate": 16000},)
NODE_CLASS_MAPPINGS = {
"EdgeTTS": EdgeTTS
}
NODE_DISPLAY_NAME_MAPPINGS = {
"EdgeTTS": "Edge TTS 🔊"
}
config.json
+70 -8
View File
@@ -3,7 +3,7 @@
"default_language": "en-US",
"default_voice": "en-US-JennyNeural",
"default_speed": 1.0,
"default_audio_format": "wav",
"default_audio_format": "mp3",
"ffmpeg_path": "ffmpeg",
"edge_tts_voices": {
"Chinese-Mainland": [
@@ -12,7 +12,10 @@
["zh-CN-YunjianNeural", "Male, Sports, Novel, Passion"],
["zh-CN-YunxiNeural", "Male, Novel, Lively, Sunshine"],
["zh-CN-YunxiaNeural", "Male, Cartoon, Novel, Cute"],
["zh-CN-YunyangNeural", "Male, News, Professional, Reliable"]
["zh-CN-YunyangNeural", "Male, News, Professional, Reliable"],
["zh-CN-XiaoxiaoMultilingualNeural", "Female, Friendly, Positive"],
["zh-CN-YunfanMultilingualNeural", "Male, Friendly, Positive"],
["zh-CN-YunxiaoMultilingualNeural", "Male, Friendly, Positive"]
],
"Chinese-Cantonese": [
["zh-HK-HiuGaaiNeural", "Female, Friendly, Positive"],
@@ -37,22 +40,74 @@
["en-US-EricNeural", "Male, Rational"],
["en-US-MichelleNeural", "Female, Friendly, Pleasant"],
["en-US-RogerNeural", "Male, Lively"],
["en-US-SteffanNeural", "Male, Rational"]
["en-US-SteffanNeural", "Male, Rational"],
["en-US-AvaMultilingualNeural", "Female, Expressive, Caring, Pleasant, Friendly"],
["en-US-AndrewMultilingualNeural", "Male, Warm, Confident, Authentic, Honest"],
["en-US-EmmaMultilingualNeural", "Female, Cheerful, Clear, Conversational"],
["en-US-BrianMultilingualNeural", "Male, Approachable, Casual, Sincere"]
],
"English-GB": [
["en-GB-LibbyNeural", "Female, Friendly, Positive"],
["en-GB-MaisieNeural", "Female, Friendly, Positive"],
["en-GB-RyanNeural", "Male, Friendly, Positive"],
["en-GB-SoniaNeural", "Female, Friendly, Positive"],
["en-GB-ThomasNeural", "Male, Friendly, Positive"]
["en-GB-ThomasNeural", "Male, Friendly, Positive"],
["en-GB-AdaMultilingualNeural", "Female, Friendly, Positive"],
["en-GB-OllieMultilingualNeural", "Male, Friendly, Positive"]
],
"English-AU": [
["en-AU-NatashaNeural", "Female, Friendly, Positive"],
["en-AU-WilliamNeural", "Male, Friendly, Positive"]
],
"English-CA": [
["en-CA-ClaraNeural", "Female, Friendly, Positive"],
["en-CA-LiamNeural", "Male, Friendly, Positive"]
],
"English-HK": [
["en-HK-SamNeural", "Male, Friendly, Positive"],
["en-HK-YanNeural", "Female, Friendly, Positive"]
],
"English-IN": [
["en-IN-NeerjaExpressiveNeural", "Female, Friendly, Positive"],
["en-IN-NeerjaNeural", "Female, Friendly, Positive"],
["en-IN-PrabhatNeural", "Male, Friendly, Positive"]
],
"English-IE": [
["en-IE-ConnorNeural", "Male, Friendly, Positive"],
["en-IE-EmilyNeural", "Female, Friendly, Positive"]
],
"English-KE": [
["en-KE-AsiliaNeural", "Female, Friendly, Positive"],
["en-KE-ChilembaNeural", "Male, Friendly, Positive"]
],
"English-NZ": [
["en-NZ-MitchellNeural", "Male, Friendly, Positive"],
["en-NZ-MollyNeural", "Female, Friendly, Positive"]
],
"English-NG": [
["en-NG-AbeoNeural", "Male, Friendly, Positive"],
["en-NG-EzinneNeural", "Female, Friendly, Positive"]
],
"English-PH": [
["en-PH-JamesNeural", "Male, Friendly, Positive"],
["en-PH-RosaNeural", "Female, Friendly, Positive"]
],
"English-SG": [
["en-SG-LunaNeural", "Female, Friendly, Positive"],
["en-SG-WayneNeural", "Male, Friendly, Positive"]
],
"English-ZA": [
["en-ZA-LeahNeural", "Female, Friendly, Positive"],
["en-ZA-LukeNeural", "Male, Friendly, Positive"]
],
"English-TZ": [
["en-TZ-ElimuNeural", "Male, Friendly, Positive"],
["en-TZ-ImaniNeural", "Female, Friendly, Positive"]
],
"Japanese": [
["ja-JP-NanamiNeural", "Female, Friendly, Positive"],
["ja-JP-KeitaNeural", "Male, Friendly, Positive"]
["ja-JP-KeitaNeural", "Male, Friendly, Positive"],
["ja-JP-MasaruMultilingualNeural", "Male, Friendly, Positive"]
],
"Korean": [
["ko-KR-SunHiNeural", "Female, Friendly, Positive"],
@@ -94,7 +149,10 @@
"Spanish": [
["es-ES-ElviraNeural", "Female, Friendly, Positive"],
["es-ES-AlvaroNeural", "Male, Friendly, Positive"],
["es-ES-XimenaNeural", "Female, Friendly, Positive"]
["es-ES-XimenaNeural", "Female, Friendly, Positive"],
["es-ES-IsidoraMultilingualNeural", "Female, Friendly, Positive"],
["es-ES-ArabellaMultilingualNeural", "Female, Friendly, Positive"],
["es-ES-TristanMultilingualNeural", "Male, Friendly, Positive"]
],
"Spanish-MX": [
["es-MX-DaliaNeural", "Female, Friendly, Positive"],
@@ -109,12 +167,16 @@
["it-IT-ElsaNeural", "Female, Friendly, Positive"],
["it-IT-DiegoNeural", "Male, Friendly, Positive"],
["it-IT-IsabellaNeural", "Female, Friendly, Positive"],
["it-IT-GiuseppeMultilingualNeural", "Male, Friendly, Positive"]
["it-IT-GiuseppeMultilingualNeural", "Male, Friendly, Positive"],
["it-IT-IsabellaMultilingualNeural", "Female, Friendly, Positive"],
["it-IT-MarcelloMultilingualNeural", "Male, Friendly, Positive"],
["it-IT-AlessioMultilingualNeural", "Male, Friendly, Positive"]
],
"Portuguese-BR": [
["pt-BR-FranciscaNeural", "Female, Friendly, Positive"],
["pt-BR-AntonioNeural", "Male, Friendly, Positive"],
["pt-BR-ThalitaMultilingualNeural", "Female, Friendly, Positive"]
["pt-BR-ThalitaMultilingualNeural", "Female, Friendly, Positive"],
["pt-BR-MacerioMultilingualNeural", "Male, Friendly, Positive"]
],
"Portuguese-PT": [
["pt-PT-RaquelNeural", "Female, Friendly, Positive"],