release: v2.0.0 (2026/10/13)
This commit is contained in:
AILab
@@ -1,15 +1,21 @@
|
||||
# ComfyUI Audio Nodes
|
||||
# ComfyUI-EdgeTTS
|
||||
|
||||
<p align="center">
|
||||
<img src="https://img.shields.io/badge/ComfyUI-Custom--Node-blue?style=for-the-badge&logo=comfyui" alt="ComfyUI">
|
||||
<img src="https://img.shields.io/badge/PyTorch-%3E=2.0-red?style=for-the-badge&logo=pytorch" alt="PyTorch">
|
||||
<img src="https://img.shields.io/badge/License-GPL--3.0-green?style=for-the-badge" alt="GPL-3.0">
|
||||
</p>
|
||||
|
||||
ComfyUI-EdgeTTS is a powerful text-to-speech node for ComfyUI, leveraging Microsoft's Edge TTS capabilities. It enables seamless conversion of text into natural-sounding speech, supporting multiple languages and voices. Ideal for enhancing user interactions, this node is easy to integrate and customize, making it perfect for various applications.
|
||||
|
||||
https://github.com/user-attachments/assets/a5b9165b-a413-49fd-989e-0ef3141afce7
|
||||
|
||||
## Updates
|
||||
- V1.2.2 (2026-01-25) - Voice ID update requirments [update log](update.md#v121-2026125).
|
||||
- V1.2.1 (2025-07-23) - Voice ID update & Bug Fixed [update log](update.md#v121-2025723).
|
||||
- V1.2.0 (2025-06-20) - Simplified voice display format, improved performance with lazy loading and caching, and reduced memory usage. For more information, please see the [update log](update.md#v120-2025620).
|
||||
|
||||
- V1.1.0 (2025-01-24) - Added 19 new languages and 38 new voices, with more detailed characteristics for existing Chinese voices. For more information, please see the [update log](update.md#v110-2025124).
|
||||
## News & Updates
|
||||
- **2026/10/13**: Update ComfyUI-EdgeTTS to **v2.0.0** ( [update.md](update.md#v200-20261013) )
|
||||
- **2026/01/25**: Update ComfyUI-EdgeTTS to **v1.2.2** ( [update.md](update.md#v122-2026125) )
|
||||
- **2025/07/23**: Update ComfyUI-EdgeTTS to **v1.2.1** ( [update.md](update.md#v121-2025723) )
|
||||
- **2025/06/20**: Update ComfyUI-EdgeTTS to **v1.2.0** ( [update.md](update.md#v120-2025620) )
|
||||
- **2025/01/24**: Update ComfyUI-EdgeTTS to **v1.1.0** ( [update.md](update.md#v110-2025124) )
|
||||
|
||||
## Features
|
||||
|
||||
@@ -24,6 +30,7 @@ https://github.com/user-attachments/assets/a5b9165b-a413-49fd-989e-0ef3141afce7
|
||||
- **Whisper STT**: High-accuracy speech recognition
|
||||
- Multiple language support with auto-detection
|
||||
- Multiple model sizes (tiny to large)
|
||||
- Dual output: raw transcription text (`TEXT`) and timecoded subtitles (`SRT`)
|
||||
- Supports ComfyUI audio format
|
||||
- Language detection confidence reporting
|
||||
|
||||
|
||||
+70
-69
@@ -1,69 +1,70 @@
|
||||
# ComfyUI-EdgeTTS
|
||||
|
||||
# ComfyUI-EdgeTTS is a powerful text-to-speech node for ComfyUI, leveraging Microsoft's Edge TTS capabilities.
|
||||
# It enables seamless conversion of text into natural-sounding speech, supporting multiple languages and voices.
|
||||
# Ideal for enhancing user interactions, this node is easy to integrate and customize, making it perfect for various applications.
|
||||
|
||||
# This integration script follows GPL-3.0 License.
|
||||
# When using or modifying this code, please respect both the original model licenses
|
||||
# and this integration's license terms.
|
||||
#
|
||||
# Source: https://github.com/1038lab/ComfyUI-EdgeTTS
|
||||
|
||||
import importlib.util
|
||||
import os
|
||||
import sys
|
||||
|
||||
current_dir = os.path.dirname(__file__)
|
||||
sys.path.insert(0, current_dir)
|
||||
|
||||
NODE_CLASS_MAPPINGS = {}
|
||||
NODE_DISPLAY_NAME_MAPPINGS = {}
|
||||
WEB_DIRECTORY = "./web"
|
||||
|
||||
def load_modules_from_directory(directory):
|
||||
if not os.path.exists(directory):
|
||||
return
|
||||
|
||||
for file in os.listdir(directory):
|
||||
if not file.endswith(".py"):
|
||||
continue
|
||||
|
||||
module_name = os.path.basename(file)[:-3]
|
||||
if module_name == os.path.basename(__file__)[:-3]:
|
||||
continue
|
||||
|
||||
file_path = os.path.join(directory, file)
|
||||
try:
|
||||
spec = importlib.util.spec_from_file_location(module_name, file_path)
|
||||
if spec is None or spec.loader is None:
|
||||
continue
|
||||
|
||||
module = importlib.util.module_from_spec(spec)
|
||||
sys.modules[module_name] = module
|
||||
spec.loader.exec_module(module)
|
||||
|
||||
if hasattr(module, "NODE_CLASS_MAPPINGS"):
|
||||
NODE_CLASS_MAPPINGS.update(module.NODE_CLASS_MAPPINGS)
|
||||
if hasattr(module, "NODE_DISPLAY_NAME_MAPPINGS"):
|
||||
NODE_DISPLAY_NAME_MAPPINGS.update(module.NODE_DISPLAY_NAME_MAPPINGS)
|
||||
|
||||
except Exception as e:
|
||||
print(f"Error loading module {module_name}: {e}")
|
||||
|
||||
def load_javascript(web_directory):
|
||||
return []
|
||||
|
||||
load_modules_from_directory(current_dir)
|
||||
load_modules_from_directory(os.path.join(current_dir, "py"))
|
||||
|
||||
NODE_CLASS_MAPPINGS = dict(sorted(
|
||||
NODE_CLASS_MAPPINGS.items(),
|
||||
key=lambda x: NODE_DISPLAY_NAME_MAPPINGS.get(x[0], x[0])
|
||||
))
|
||||
NODE_DISPLAY_NAME_MAPPINGS = dict(sorted(
|
||||
NODE_DISPLAY_NAME_MAPPINGS.items(),
|
||||
key=lambda x: x[1]
|
||||
))
|
||||
|
||||
__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY", "load_javascript"]
|
||||
from pathlib import Path
|
||||
import sys
|
||||
import importlib.util
|
||||
|
||||
__repo_name__ = "ComfyUI-EdgeTTS"
|
||||
__version__ = "2.0.0"
|
||||
|
||||
current_dir = Path(__file__).parent
|
||||
if str(current_dir) not in sys.path:
|
||||
sys.path.insert(0, str(current_dir))
|
||||
|
||||
scan_dirs = [current_dir]
|
||||
nodes_dir = current_dir / "py"
|
||||
if nodes_dir.exists() and nodes_dir.is_dir():
|
||||
scan_dirs.append(nodes_dir)
|
||||
if str(nodes_dir) not in sys.path:
|
||||
sys.path.insert(0, str(nodes_dir))
|
||||
|
||||
NODE_CLASS_MAPPINGS = {}
|
||||
NODE_DISPLAY_NAME_MAPPINGS = {}
|
||||
WEB_DIRECTORY = "./web"
|
||||
|
||||
|
||||
def load_nodes():
|
||||
models_dir = current_dir / "models"
|
||||
for base in scan_dirs:
|
||||
for file in base.rglob("*.py"):
|
||||
if file.stem == "__init__":
|
||||
continue
|
||||
try:
|
||||
if file.is_relative_to(models_dir):
|
||||
continue
|
||||
except Exception:
|
||||
if str(models_dir) in str(file):
|
||||
continue
|
||||
|
||||
module_name = file.stem
|
||||
try:
|
||||
spec = importlib.util.spec_from_file_location(module_name, file)
|
||||
if spec and spec.loader:
|
||||
module = importlib.util.module_from_spec(spec)
|
||||
sys.modules[module_name] = module
|
||||
spec.loader.exec_module(module)
|
||||
|
||||
if hasattr(module, "NODE_CLASS_MAPPINGS"):
|
||||
NODE_CLASS_MAPPINGS.update(module.NODE_CLASS_MAPPINGS)
|
||||
if hasattr(module, "NODE_DISPLAY_NAME_MAPPINGS"):
|
||||
NODE_DISPLAY_NAME_MAPPINGS.update(module.NODE_DISPLAY_NAME_MAPPINGS)
|
||||
except Exception as e:
|
||||
print(f"[{__repo_name__}] Error loading module {file.name}: {e}")
|
||||
|
||||
|
||||
load_nodes()
|
||||
|
||||
NODE_CLASS_MAPPINGS = dict(
|
||||
sorted(
|
||||
NODE_CLASS_MAPPINGS.items(),
|
||||
key=lambda x: NODE_DISPLAY_NAME_MAPPINGS.get(x[0], x[0]),
|
||||
)
|
||||
)
|
||||
NODE_DISPLAY_NAME_MAPPINGS = dict(
|
||||
sorted(NODE_DISPLAY_NAME_MAPPINGS.items(), key=lambda x: x[1])
|
||||
)
|
||||
|
||||
__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY"]
|
||||
|
||||
print(f'\033[36m[{__repo_name__}]\033[0m v'
|
||||
f'\033[93m{__version__}\033[0m | '
|
||||
f'\033[37m{len(NODE_CLASS_MAPPINGS)} nodes\033[0m '
|
||||
f'\033[92mLoaded\033[0m')
|
||||
+6
-6
@@ -28,21 +28,21 @@ class Save_Audio:
|
||||
def INPUT_TYPES(s):
|
||||
return {
|
||||
"required": {
|
||||
"audio": ("AUDIO", ),
|
||||
"audio": ("AUDIO", {"tooltip": "Input audio stream to save"}),
|
||||
"filepath": ("STRING", {
|
||||
"multiline": True,
|
||||
"default": "",
|
||||
"placeholder": "custom_path/filename",
|
||||
"tooltip": "by default, the audio will be saved to output/TTS/{date}/TTS-xxxx.{format}"
|
||||
"tooltip": "Custom output path or filename (default: output/TTS/{date}/TTS-xxxx.{format})"
|
||||
}),
|
||||
"format": (s.AUDIO_FORMATS, {"default": "mp3"}),
|
||||
"quality": (s.QUALITY_PRESETS, {"default": "high"}),
|
||||
"overwrite": ("BOOLEAN", {"default": False})
|
||||
"format": (s.AUDIO_FORMATS, {"default": "mp3", "tooltip": "Audio format (mp3, wav, flac)"}),
|
||||
"quality": (s.QUALITY_PRESETS, {"default": "high", "tooltip": "Audio quality bitrate preset"}),
|
||||
"overwrite": ("BOOLEAN", {"default": False, "tooltip": "Overwrite existing file if set to true"})
|
||||
},
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("STRING", "AUDIO",)
|
||||
RETURN_NAMES = ("filepath", "audio",)
|
||||
RETURN_NAMES = ("FILEPATH", "AUDIO",)
|
||||
FUNCTION = "save_audio"
|
||||
OUTPUT_NODE = True
|
||||
CATEGORY = "🧪AILab/🔊Audio"
|
||||
|
||||
+62
-29
@@ -1,8 +1,3 @@
|
||||
# ComfyUI-EdgeTTS V1.2.1
|
||||
# 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
|
||||
@@ -10,6 +5,37 @@ import re
|
||||
import torch
|
||||
import torchaudio
|
||||
import json
|
||||
from ailab_utils import setup_audio_service
|
||||
|
||||
setup_audio_service()
|
||||
|
||||
|
||||
def decode_audio(filepath):
|
||||
try:
|
||||
waveform, sample_rate = torchaudio.load(filepath)
|
||||
return waveform, sample_rate
|
||||
except Exception:
|
||||
try:
|
||||
from comfy_extras.nodes_audio import load as comfy_audio_load
|
||||
return comfy_audio_load(filepath)
|
||||
except Exception:
|
||||
import av
|
||||
with av.open(filepath) as af:
|
||||
stream = af.streams.audio[0]
|
||||
sr = stream.codec_context.sample_rate
|
||||
n_channels = stream.channels
|
||||
frames = []
|
||||
for frame in af.decode(streams=stream.index):
|
||||
buf = torch.from_numpy(frame.to_ndarray())
|
||||
if buf.shape[0] != n_channels:
|
||||
buf = buf.view(-1, n_channels).t()
|
||||
frames.append(buf)
|
||||
wav = torch.cat(frames, dim=1)
|
||||
if wav.dtype.is_floating_point:
|
||||
wav = wav.to(torch.float32)
|
||||
else:
|
||||
wav = wav.to(torch.float32) / 32768.0
|
||||
return wav, sr
|
||||
|
||||
class EdgeTTS:
|
||||
_voice_data_cache = None
|
||||
@@ -78,7 +104,7 @@ class EdgeTTS:
|
||||
voices, _, _ = cls.get_voice_data()
|
||||
return {
|
||||
"required": {
|
||||
"text": ("STRING", {"multiline": True, "placeholder": "Enter text to convert to speech"}),
|
||||
"text": ("STRING", {"multiline": True, "placeholder": "Enter text to convert to speech", "tooltip": "Text content to convert into speech"}),
|
||||
"voice": (voices, {"default": voices[0], "tooltip": "Select a voice for text-to-speech"}),
|
||||
},
|
||||
"optional": {
|
||||
@@ -88,57 +114,64 @@ class EdgeTTS:
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("AUDIO",)
|
||||
RETURN_NAMES = ("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"
|
||||
|
||||
temp_file = f"temp_tts_{os.getpid()}.mp3"
|
||||
try:
|
||||
text = text.strip()
|
||||
if not text:
|
||||
raise ValueError("Input text cannot be empty")
|
||||
|
||||
setup_audio_service()
|
||||
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"
|
||||
)
|
||||
for attempt in range(2):
|
||||
try:
|
||||
await communicate.save(temp_file)
|
||||
else:
|
||||
break
|
||||
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)
|
||||
break
|
||||
raise
|
||||
|
||||
waveform, sample_rate = torchaudio.load(temp_file)
|
||||
except Exception as e:
|
||||
if attempt == 0:
|
||||
await asyncio.sleep(0.3)
|
||||
setup_audio_service()
|
||||
continue
|
||||
raise e
|
||||
|
||||
waveform, sample_rate = decode_audio(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:
|
||||
except Exception:
|
||||
pass
|
||||
|
||||
def tts(self, text, voice, speed=1.0, pitch=0):
|
||||
|
||||
+397
@@ -0,0 +1,397 @@
|
||||
import secrets
|
||||
import struct
|
||||
from datetime import datetime as dt
|
||||
from datetime import timezone as tz
|
||||
from functools import lru_cache
|
||||
from typing import List, Optional, Tuple
|
||||
|
||||
import edge_tts
|
||||
import edge_tts.communicate
|
||||
import edge_tts.constants
|
||||
import edge_tts.drm
|
||||
|
||||
TRUSTED_CLIENT_TOKEN = "6A5AA1D4EAFF4E9FB37E23D68491D6F4"
|
||||
CHROMIUM_FULL_VERSION = "143.0.3650.75"
|
||||
SEC_MS_GEC_VERSION = f"2-{CHROMIUM_FULL_VERSION}"
|
||||
|
||||
WIN_EPOCH = 11644473600
|
||||
S_TO_NS = 1e9
|
||||
_U32_MASK = 0xFFFFFFFF
|
||||
|
||||
_SHA256_INITIAL = (
|
||||
0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
|
||||
0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19,
|
||||
)
|
||||
_SHA256_STATE = (
|
||||
0xFD9FBEA2, 0x91243CE7, 0xBE499094, 0xCD2E1E0A,
|
||||
0x6C006F49, 0x04BD35AB, 0xFEB0A7E4, 0x03EFFBD2,
|
||||
)
|
||||
_SHA256_SCHEDULE = (
|
||||
0xEB24BF9E, 0x1558691F, 0xA0C09EE2, 0x0CFA32BA,
|
||||
0xA622EE8C, 0xD1DD9491, 0xC0DD25A0, 0xF9E70AB6,
|
||||
0xF0C31573, 0x7058CE4E, 0x7188E606, 0x00E5AB33,
|
||||
0xD258B060, 0x72766564,
|
||||
)
|
||||
_SHA256_ROUND_CONSTANTS = (
|
||||
0x428A2F98, 0x71374491, 0xB5C0FBCF, 0xE9B5DBA5,
|
||||
0x3956C25B, 0x59F111F1, 0x923F82A4, 0xAB1C5ED5,
|
||||
0xD807AA98, 0x12835B01, 0x243185BE, 0x550C7DC3,
|
||||
0x72BE5D74, 0x80DEB1FE, 0x9BDC06A7, 0xC19BF174,
|
||||
0xE49B69C1, 0xEFBE4786, 0x0FC19DC6, 0x240CA1CC,
|
||||
0x2DE92C6F, 0x4A7484AA, 0x5CB0A9DC, 0x76F988DA,
|
||||
0x983E5152, 0xA831C66D, 0xB00327C8, 0xBF597FC7,
|
||||
0xC6E00BF3, 0xD5A79147, 0x06CA6351, 0x14292967,
|
||||
0x27B70A85, 0x2E1B2138, 0x4D2C6DFC, 0x53380D13,
|
||||
0x650A7354, 0x766A0ABB, 0x81C2C92E, 0x92722C85,
|
||||
0xA2BFE8A1, 0xA81A664B, 0xC24B8B70, 0xC76C51A3,
|
||||
0xD192E819, 0xD6990624, 0xF40E3585, 0x106AA070,
|
||||
0x19A4C116, 0x1E376C08, 0x2748774C, 0x34B0BCB5,
|
||||
0x391C0CB3, 0x4ED8AA4A, 0x5B9CCA4F, 0x682E6FF3,
|
||||
0x748F82EE, 0x78A5636F, 0x84C87814, 0x8CC70208,
|
||||
0x90BEFFFA, 0xA4506CEB, 0xBEF9A3F7, 0xC67178F2,
|
||||
)
|
||||
_SHA256_ADJUSTMENT = 0xC9DC06A7
|
||||
_SHA256_BLOCK_CONSTANTS = (
|
||||
_SHA256_ROUND_CONSTANTS[:14] + (_SHA256_ADJUSTMENT,) + _SHA256_ROUND_CONSTANTS[15:]
|
||||
)
|
||||
_SHA256_SIGMA0_ADJUSTMENT = 0x059C0B80
|
||||
_SHA256_SIGMA1_ADJUSTMENT = 0x000B92C0
|
||||
|
||||
_CS64_COEFFICIENTS = (
|
||||
(0xE82E91B1, 0xB32700D0),
|
||||
(0x44DE24C6, 0x88DFF51D),
|
||||
(0x3BEB65BC, 0x29514038),
|
||||
(0xC89C5933, 0xB328CBDF),
|
||||
(0x8D1EC211, 0xD405B4D5),
|
||||
(0x314EF5D5, 0x26E55F3C),
|
||||
(0xE92491B1, 0x607801C6),
|
||||
(0xF0A5B9D5, 0x82B2C1FB),
|
||||
(0x06064CF3, 0xFA51A3DA),
|
||||
(0xDA79FA60, 0x207CC464),
|
||||
(0xBED20560, 0x3C3C74F3),
|
||||
(0x89F9C9F8, 0xD2D14B61),
|
||||
(0x35FA38C6, 0xBB04A2C1),
|
||||
(0xCB49E3C6, 0xC5626167),
|
||||
)
|
||||
_CS64_FACTOR = 0x82893933
|
||||
_CS64_ADJUSTMENT0 = 0xF2F90B1A
|
||||
_CS64_ADJUSTMENT1 = 0x0B1A0000
|
||||
|
||||
|
||||
def _rol16(value: int) -> int:
|
||||
value &= _U32_MASK
|
||||
return ((value << 16) | (value >> 16)) & _U32_MASK
|
||||
|
||||
|
||||
def _words_le4_even(data: bytes) -> List[int]:
|
||||
padded = data + b"\x00" * ((-len(data)) % 4)
|
||||
if not padded:
|
||||
return []
|
||||
words = list(struct.unpack(f"<{len(padded) // 4}I", padded))
|
||||
if len(words) % 2:
|
||||
words.append(0)
|
||||
return words
|
||||
|
||||
|
||||
def _sigma0(value: int) -> int:
|
||||
return (
|
||||
(value >> 7) ^ (value << 25) ^ (value >> 18) ^ (value << 14) ^ (value >> 3)
|
||||
) & _U32_MASK
|
||||
|
||||
|
||||
def _sigma1(value: int) -> int:
|
||||
return (
|
||||
(value >> 17) ^ (value << 15) ^ (value >> 19) ^ (value << 13) ^ (value >> 10)
|
||||
) & _U32_MASK
|
||||
|
||||
|
||||
def _sha256_rounds(
|
||||
state: Tuple[int, ...], words: List[int], first_round: int = 0
|
||||
) -> Tuple[int, ...]:
|
||||
a, b, c, d, e, f, g, h = state
|
||||
mask = _U32_MASK
|
||||
constants = (
|
||||
_SHA256_BLOCK_CONSTANTS if first_round == 14 else _SHA256_ROUND_CONSTANTS
|
||||
)
|
||||
aligned = min(64, ((first_round + 7) // 8) * 8)
|
||||
for index in range(first_round, aligned):
|
||||
rot_e = e * 0x100000001
|
||||
rot_a = a * 0x100000001
|
||||
t1 = (
|
||||
h
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (g ^ (e & (f ^ g)))
|
||||
+ constants[index]
|
||||
+ words[index]
|
||||
)
|
||||
t2 = ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22)) + ((a & b) | (c & (a | b)))
|
||||
a, b, c, d, e, f, g, h = (t1 + t2) & mask, a, b, c, (d + t1) & mask, e, f, g
|
||||
|
||||
for index in range(aligned, 64, 8):
|
||||
rot_e = e * 0x100000001
|
||||
rot_a = a * 0x100000001
|
||||
t1 = (
|
||||
h
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (g ^ (e & (f ^ g)))
|
||||
+ constants[index]
|
||||
+ words[index]
|
||||
)
|
||||
d = (d + t1) & mask
|
||||
h = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((a & b) | (c & (a | b)))
|
||||
) & mask
|
||||
rot_e = d * 0x100000001
|
||||
rot_a = h * 0x100000001
|
||||
t1 = (
|
||||
g
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (f ^ (d & (e ^ f)))
|
||||
+ constants[index + 1]
|
||||
+ words[index + 1]
|
||||
)
|
||||
c = (c + t1) & mask
|
||||
g = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((h & a) | (b & (h | a)))
|
||||
) & mask
|
||||
rot_e = c * 0x100000001
|
||||
rot_a = g * 0x100000001
|
||||
t1 = (
|
||||
f
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (e ^ (c & (d ^ e)))
|
||||
+ constants[index + 2]
|
||||
+ words[index + 2]
|
||||
)
|
||||
b = (b + t1) & mask
|
||||
f = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((g & h) | (a & (g | h)))
|
||||
) & mask
|
||||
rot_e = b * 0x100000001
|
||||
rot_a = f * 0x100000001
|
||||
t1 = (
|
||||
e
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (d ^ (b & (c ^ d)))
|
||||
+ constants[index + 3]
|
||||
+ words[index + 3]
|
||||
)
|
||||
a = (a + t1) & mask
|
||||
e = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((f & g) | (h & (f | g)))
|
||||
) & mask
|
||||
rot_e = a * 0x100000001
|
||||
rot_a = e * 0x100000001
|
||||
t1 = (
|
||||
d
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (c ^ (a & (b ^ c)))
|
||||
+ constants[index + 4]
|
||||
+ words[index + 4]
|
||||
)
|
||||
h = (h + t1) & mask
|
||||
d = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((e & f) | (g & (e | f)))
|
||||
) & mask
|
||||
rot_e = h * 0x100000001
|
||||
rot_a = d * 0x100000001
|
||||
t1 = (
|
||||
c
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (b ^ (h & (a ^ b)))
|
||||
+ constants[index + 5]
|
||||
+ words[index + 5]
|
||||
)
|
||||
g = (g + t1) & mask
|
||||
c = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((d & e) | (f & (d | e)))
|
||||
) & mask
|
||||
rot_e = g * 0x100000001
|
||||
rot_a = c * 0x100000001
|
||||
t1 = (
|
||||
b
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (a ^ (g & (h ^ a)))
|
||||
+ constants[index + 6]
|
||||
+ words[index + 6]
|
||||
)
|
||||
f = (f + t1) & mask
|
||||
b = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((c & d) | (e & (c | d)))
|
||||
) & mask
|
||||
rot_e = f * 0x100000001
|
||||
rot_a = b * 0x100000001
|
||||
t1 = (
|
||||
a
|
||||
+ ((rot_e >> 6) ^ (rot_e >> 11) ^ (rot_e >> 25))
|
||||
+ (h ^ (f & (g ^ h)))
|
||||
+ constants[index + 7]
|
||||
+ words[index + 7]
|
||||
)
|
||||
e = (e + t1) & mask
|
||||
a = (
|
||||
t1
|
||||
+ ((rot_a >> 2) ^ (rot_a >> 13) ^ (rot_a >> 22))
|
||||
+ ((b & c) | (d & (b | c)))
|
||||
) & mask
|
||||
return a, b, c, d, e, f, g, h
|
||||
|
||||
|
||||
def _expand_sha256_schedule(words: List[int], first_word: int = 16) -> None:
|
||||
mask = _U32_MASK
|
||||
for index in range(first_word, 64):
|
||||
x, y = words[index - 2], words[index - 15]
|
||||
sigma1 = (x >> 17) ^ (x << 15) ^ (x >> 19) ^ (x << 13) ^ (x >> 10)
|
||||
sigma0 = (y >> 7) ^ (y << 25) ^ (y >> 18) ^ (y << 14) ^ (y >> 3)
|
||||
words.append((sigma1 + words[index - 7] + sigma0 + words[index - 16]) & mask)
|
||||
|
||||
|
||||
@lru_cache(maxsize=1)
|
||||
def _sha256_init(data: bytes) -> Tuple[int, ...]:
|
||||
mask = _U32_MASK
|
||||
words = [0] * 14 + [int.from_bytes(data[:3], "big")]
|
||||
words.append(int.from_bytes(data[3:7], "big"))
|
||||
for index in range(16, 30):
|
||||
value = _SHA256_SCHEDULE[index - 16]
|
||||
if index - 2 >= 14:
|
||||
value += (
|
||||
(_sigma1(words[index - 2]) ^ _SHA256_SIGMA1_ADJUSTMENT)
|
||||
if index == 16
|
||||
else _sigma1(words[index - 2])
|
||||
)
|
||||
if index - 7 >= 14:
|
||||
value += words[index - 7]
|
||||
if index - 15 >= 14:
|
||||
value += (
|
||||
(_sigma0(words[index - 15]) ^ _SHA256_SIGMA0_ADJUSTMENT)
|
||||
if index == 29
|
||||
else _sigma0(words[index - 15])
|
||||
)
|
||||
words.append(value & mask)
|
||||
words.append(
|
||||
(
|
||||
_sigma1(words[28])
|
||||
+ words[23]
|
||||
+ _sigma0(words[15])
|
||||
+ words[14]
|
||||
+ _SHA256_ADJUSTMENT
|
||||
- _SHA256_ROUND_CONSTANTS[14]
|
||||
)
|
||||
& mask
|
||||
)
|
||||
_expand_sha256_schedule(words, 31)
|
||||
working = _sha256_rounds(_SHA256_STATE, words, 14)
|
||||
return tuple((x + y) & mask for x, y in zip(_SHA256_INITIAL, working))
|
||||
|
||||
|
||||
def _sha256_digest(data: bytes) -> bytes:
|
||||
if len(data) < 7:
|
||||
raise ValueError("Token salt must contain at least seven bytes")
|
||||
mask = _U32_MASK
|
||||
length = 57 + len(data)
|
||||
padded = data[7:] + b"\x80" + b"\x00" * ((55 - length) % 64)
|
||||
padded += struct.pack(">Q", length * 8)
|
||||
state = _sha256_init(data[:7])
|
||||
for offset in range(0, len(padded), 64):
|
||||
words = list(struct.unpack(">16I", padded[offset : offset + 64]))
|
||||
_expand_sha256_schedule(words)
|
||||
working = _sha256_rounds(state, words)
|
||||
state = tuple((x + y) & mask for x, y in zip(state, working))
|
||||
return struct.pack(">8I", *state)
|
||||
|
||||
|
||||
def _cs64_digest(data: bytes, digest: bytes) -> Tuple[int, int]:
|
||||
mask = _U32_MASK
|
||||
key_a, key_b = struct.unpack_from("<II", digest)
|
||||
key_a |= 1
|
||||
key_b |= 1
|
||||
factors = (
|
||||
(key_a * _CS64_FACTOR) & mask,
|
||||
(key_b * _CS64_FACTOR) & mask,
|
||||
)
|
||||
terms = [
|
||||
(coefficient * factors[index % 2]) & mask
|
||||
for index, (coefficient, _) in enumerate(_CS64_COEFFICIENTS)
|
||||
]
|
||||
offsets = [offset for _, offset in _CS64_COEFFICIENTS]
|
||||
for index, word in enumerate(_words_le4_even(b"\x00" + data)):
|
||||
term = word * (key_a if index % 2 == 0 else key_b)
|
||||
offset = _rol16(word) * (0xEF0569FB if index % 2 == 0 else 0xC31713DB)
|
||||
if index == 0:
|
||||
term += _CS64_ADJUSTMENT0 * factors[0]
|
||||
offset += _CS64_ADJUSTMENT1
|
||||
terms.append(term & mask)
|
||||
offsets.append(offset & mask)
|
||||
swap_acc = swap_last = state = acc = 0
|
||||
for index in range(0, len(terms), 2):
|
||||
x, y = terms[index], terms[index + 1]
|
||||
t = _rol16((offsets[index] + x) & mask)
|
||||
q = (t * 0x76F50F34) & mask
|
||||
swap_acc += (_rol16(q) * 0xC3EFEA97 + q) & mask
|
||||
t = _rol16((offsets[index + 1] + y) & mask)
|
||||
q = (t * 0x1881CE3C) & mask
|
||||
swap_last = (_rol16(q) * 0x35BD1EC9 + q) & mask
|
||||
swap_acc = (swap_acc + swap_last) & mask
|
||||
|
||||
state = (state * key_a + x) & mask
|
||||
state = (_rol16(state) * 0xCF98B111) & mask
|
||||
state = (_rol16(state) * 0x87085B9F) & mask
|
||||
state = (_rol16(state) * 0x12CEB96D) & mask
|
||||
state = (_rol16(state) * 0x257E1D83) & mask
|
||||
acc += state
|
||||
t = (state * key_b + y) & mask
|
||||
t = (_rol16(t) * 0xA27416F5) & mask
|
||||
t = (_rol16(t) * 0xD38396FF) & mask
|
||||
state = (_rol16((_rol16(t) * 0x7C932B89) & mask) * 0xBFA49F69) & mask
|
||||
acc = (acc + state) & mask
|
||||
return swap_last ^ state, swap_acc ^ acc
|
||||
|
||||
|
||||
@lru_cache(maxsize=1)
|
||||
def _token_hash(data: bytes) -> Tuple[str, str]:
|
||||
digest = _sha256_digest(data)
|
||||
cs64 = struct.pack("<II", *_cs64_digest(data, digest)).hex().upper()
|
||||
return cs64, digest.hex().upper()
|
||||
|
||||
|
||||
def generate_sec_ms_gec_v2(unix_timestamp: Optional[float] = None) -> str:
|
||||
if unix_timestamp is None:
|
||||
ticks = dt.now(tz.utc).timestamp()
|
||||
else:
|
||||
ticks = float(unix_timestamp)
|
||||
ticks += WIN_EPOCH
|
||||
ticks -= ticks % 300
|
||||
ticks *= S_TO_NS / 100
|
||||
str_to_hash = f"{ticks:.0f}{TRUSTED_CLIENT_TOKEN}".encode("ascii")
|
||||
cs64, sha256 = _token_hash(str_to_hash)
|
||||
return f"{cs64}-{sha256}"
|
||||
|
||||
|
||||
def setup_audio_service() -> None:
|
||||
current_version = getattr(edge_tts.constants, "SEC_MS_GEC_VERSION", "")
|
||||
if current_version.startswith("2-"):
|
||||
return
|
||||
|
||||
edge_tts.constants.SEC_MS_GEC_VERSION = SEC_MS_GEC_VERSION
|
||||
if hasattr(edge_tts.communicate, "SEC_MS_GEC_VERSION"):
|
||||
edge_tts.communicate.SEC_MS_GEC_VERSION = SEC_MS_GEC_VERSION
|
||||
|
||||
if hasattr(edge_tts, "drm") and hasattr(edge_tts.drm, "DRM"):
|
||||
edge_tts.drm.DRM.generate_sec_ms_gec = staticmethod(generate_sec_ms_gec_v2)
|
||||
|
||||
if hasattr(edge_tts.communicate, "DRM"):
|
||||
edge_tts.communicate.DRM.generate_sec_ms_gec = staticmethod(generate_sec_ms_gec_v2)
|
||||
+11
-4
@@ -1,7 +1,9 @@
|
||||
import io
|
||||
import os
|
||||
import torch
|
||||
import torchaudio
|
||||
import whisper
|
||||
from whisper.utils import WriteSRT
|
||||
import json
|
||||
|
||||
class WhisperSTT:
|
||||
@@ -32,7 +34,7 @@ class WhisperSTT:
|
||||
def INPUT_TYPES(s):
|
||||
return {
|
||||
"required": {
|
||||
"audio": ("AUDIO",),
|
||||
"audio": ("AUDIO", {"tooltip": "Input audio stream to transcribe"}),
|
||||
"model_size": (s.WHISPER_MODELS, {
|
||||
"default": "base",
|
||||
"tooltip": "Whisper model size - larger = more accurate but slower"
|
||||
@@ -44,7 +46,8 @@ class WhisperSTT:
|
||||
}
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("STRING",)
|
||||
RETURN_TYPES = ("STRING", "STRING",)
|
||||
RETURN_NAMES = ("TEXT", "SRT",)
|
||||
FUNCTION = "stt"
|
||||
CATEGORY = "🧪AILab/🔊Audio"
|
||||
|
||||
@@ -66,8 +69,12 @@ class WhisperSTT:
|
||||
detected_lang = result.get("language", "unknown")
|
||||
confidence = result.get("language_probability", 0.0)
|
||||
print(f"[Whisper STT] Detected: {detected_lang} (conf: {confidence:.2f})")
|
||||
|
||||
return (result["text"].strip(),)
|
||||
|
||||
sio = io.StringIO()
|
||||
WriteSRT(".").write_result(result, sio)
|
||||
srt_text = sio.getvalue().strip()
|
||||
|
||||
return (result["text"].strip(), srt_text)
|
||||
|
||||
except Exception as e:
|
||||
print(f"[Whisper STT] Error: {str(e)}")
|
||||
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 116 KiB |
@@ -0,0 +1,362 @@
|
||||
{
|
||||
"id": "0f97e221-e840-4849-8fbf-b9f07bcf2a12",
|
||||
"revision": 0,
|
||||
"last_node_id": 7,
|
||||
"last_link_id": 7,
|
||||
"nodes": [
|
||||
{
|
||||
"id": 1,
|
||||
"type": "EdgeTTS",
|
||||
"pos": [
|
||||
120,
|
||||
150
|
||||
],
|
||||
"size": [
|
||||
380,
|
||||
260
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "AUDIO",
|
||||
"type": "AUDIO",
|
||||
"slot_index": 0,
|
||||
"links": [
|
||||
3,
|
||||
5
|
||||
]
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"cnr_id": "comfyui-edgetts",
|
||||
"ver": "7b166c5897ad802b116352d15f227275bb5f829d",
|
||||
"Node name for S&R": "EdgeTTS",
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"version": "7.8",
|
||||
"input_ue_unconnectable": {}
|
||||
}
|
||||
},
|
||||
"widgets_values": [
|
||||
"Hello! This is Microsoft Edge Text to Speech running in ComfyUI. We are testing audio synthesis and speech recognition with automatic SRT subtitles.",
|
||||
"[English] en-US Jenny",
|
||||
1,
|
||||
0
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"text": "Hello! This is Microsoft Edge Text to Speech running in ComfyUI. We are testing audio synthesis and speech recognition with automatic SRT subtitles.",
|
||||
"voice": "[English] en-US Jenny",
|
||||
"speed": 1,
|
||||
"pitch": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"id": 2,
|
||||
"type": "Save_Audio",
|
||||
"pos": [
|
||||
890.3295245655077,
|
||||
150
|
||||
],
|
||||
"size": [
|
||||
292.3765127144609,
|
||||
180
|
||||
],
|
||||
"flags": {},
|
||||
"order": 5,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"link": 4
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "FILEPATH",
|
||||
"type": "STRING",
|
||||
"links": null
|
||||
},
|
||||
{
|
||||
"name": "AUDIO",
|
||||
"type": "AUDIO",
|
||||
"links": null
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"cnr_id": "comfyui-edgetts",
|
||||
"ver": "7b166c5897ad802b116352d15f227275bb5f829d",
|
||||
"Node name for S&R": "Save_Audio",
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"version": "7.8",
|
||||
"input_ue_unconnectable": {}
|
||||
}
|
||||
},
|
||||
"widgets_values": [
|
||||
"",
|
||||
"mp3",
|
||||
"high",
|
||||
false
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"filepath": "",
|
||||
"format": "mp3",
|
||||
"quality": "high",
|
||||
"overwrite": false
|
||||
}
|
||||
},
|
||||
{
|
||||
"id": 4,
|
||||
"type": "WhisperSTT",
|
||||
"pos": [
|
||||
539.2942229626224,
|
||||
397.2078115425857
|
||||
],
|
||||
"size": [
|
||||
291.1443589154411,
|
||||
102
|
||||
],
|
||||
"flags": {},
|
||||
"order": 1,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"link": 3
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "TEXT",
|
||||
"type": "STRING",
|
||||
"links": [
|
||||
6
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "SRT",
|
||||
"type": "STRING",
|
||||
"links": [
|
||||
7
|
||||
]
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"input_ue_unconnectable": {}
|
||||
},
|
||||
"cnr_id": "comfyui-edgetts",
|
||||
"ver": "7b166c5897ad802b116352d15f227275bb5f829d",
|
||||
"Node name for S&R": "WhisperSTT"
|
||||
},
|
||||
"widgets_values": [
|
||||
"base",
|
||||
"auto"
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"model_size": "base",
|
||||
"language": "auto"
|
||||
}
|
||||
},
|
||||
{
|
||||
"id": 5,
|
||||
"type": "PreviewAudio",
|
||||
"pos": [
|
||||
542.9514251545787,
|
||||
150
|
||||
],
|
||||
"size": [
|
||||
270,
|
||||
88
|
||||
],
|
||||
"flags": {},
|
||||
"order": 2,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"link": 5
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"links": [
|
||||
4
|
||||
]
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"input_ue_unconnectable": {}
|
||||
},
|
||||
"cnr_id": "comfy-core",
|
||||
"ver": "0.39.1",
|
||||
"Node name for S&R": "PreviewAudio"
|
||||
},
|
||||
"widgets_values": [],
|
||||
"widgets_values_named": {}
|
||||
},
|
||||
{
|
||||
"id": 6,
|
||||
"type": "iToolsPreviewText",
|
||||
"pos": [
|
||||
890.3295245655077,
|
||||
397.2078115425857
|
||||
],
|
||||
"size": [
|
||||
302.1176470588234,
|
||||
100.37646005667887
|
||||
],
|
||||
"flags": {},
|
||||
"order": 3,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "text",
|
||||
"type": "STRING",
|
||||
"link": 6
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "text",
|
||||
"shape": 6,
|
||||
"type": "STRING",
|
||||
"links": null
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"input_ue_unconnectable": {}
|
||||
},
|
||||
"cnr_id": "comfyui-itools",
|
||||
"ver": "a8f2fef227ca5b0cb9321c39ac7b874a906b1a5b",
|
||||
"Node name for S&R": "iToolsPreviewText"
|
||||
},
|
||||
"widgets_values": [
|
||||
"Hello. This is Microsoft Edge Text to speech running in Comfy UI. We are testing audio synthesis and speech recognition with automatic SRT subtitles."
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"text2": "Hello. This is Microsoft Edge Text to speech running in Comfy UI. We are testing audio synthesis and speech recognition with automatic SRT subtitles."
|
||||
}
|
||||
},
|
||||
{
|
||||
"id": 7,
|
||||
"type": "iToolsPreviewText",
|
||||
"pos": [
|
||||
891.0197627701644,
|
||||
556.6776902033889
|
||||
],
|
||||
"size": [
|
||||
297.97642846200984,
|
||||
208.58036056219373
|
||||
],
|
||||
"flags": {},
|
||||
"order": 4,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "text",
|
||||
"type": "STRING",
|
||||
"link": 7
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "text",
|
||||
"shape": 6,
|
||||
"type": "STRING",
|
||||
"links": []
|
||||
}
|
||||
],
|
||||
"properties": {
|
||||
"ue_properties": {
|
||||
"widget_ue_connectable": {},
|
||||
"input_ue_unconnectable": {}
|
||||
},
|
||||
"cnr_id": "comfyui-itools",
|
||||
"ver": "a8f2fef227ca5b0cb9321c39ac7b874a906b1a5b",
|
||||
"Node name for S&R": "iToolsPreviewText"
|
||||
},
|
||||
"widgets_values": [
|
||||
"1\n00:00:00,000 --> 00:00:01,000\nHello.\n\n2\n00:00:01,000 --> 00:00:05,880\nThis is Microsoft Edge Text to speech running in Comfy UI.\n\n3\n00:00:05,880 --> 00:00:10,800\nWe are testing audio synthesis and speech recognition with automatic SRT subtitles."
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"text2": "1\n00:00:00,000 --> 00:00:01,000\nHello.\n\n2\n00:00:01,000 --> 00:00:05,880\nThis is Microsoft Edge Text to speech running in Comfy UI.\n\n3\n00:00:05,880 --> 00:00:10,800\nWe are testing audio synthesis and speech recognition with automatic SRT subtitles."
|
||||
}
|
||||
}
|
||||
],
|
||||
"links": [
|
||||
[
|
||||
3,
|
||||
1,
|
||||
0,
|
||||
4,
|
||||
0,
|
||||
"AUDIO"
|
||||
],
|
||||
[
|
||||
4,
|
||||
5,
|
||||
0,
|
||||
2,
|
||||
0,
|
||||
"AUDIO"
|
||||
],
|
||||
[
|
||||
5,
|
||||
1,
|
||||
0,
|
||||
5,
|
||||
0,
|
||||
"AUDIO"
|
||||
],
|
||||
[
|
||||
6,
|
||||
4,
|
||||
0,
|
||||
6,
|
||||
0,
|
||||
"STRING"
|
||||
],
|
||||
[
|
||||
7,
|
||||
4,
|
||||
1,
|
||||
7,
|
||||
0,
|
||||
"STRING"
|
||||
]
|
||||
],
|
||||
"groups": [],
|
||||
"config": {},
|
||||
"extra": {
|
||||
"ue_links": [],
|
||||
"links_added_by_ue": [],
|
||||
"ds": {
|
||||
"scale": 1.1590909090909092,
|
||||
"offset": [
|
||||
42.24245340003889,
|
||||
28.257345869366866
|
||||
]
|
||||
},
|
||||
"frontendVersion": "1.53.10",
|
||||
"VHS_latentpreview": false,
|
||||
"VHS_latentpreviewrate": 0,
|
||||
"VHS_MetadataImage": true,
|
||||
"VHS_KeepIntermediate": true
|
||||
},
|
||||
"version": 0.4,
|
||||
"seed_widgets": {}
|
||||
}
|
||||
+2
-2
@@ -1,13 +1,13 @@
|
||||
[project]
|
||||
name = "comfyui-edgetts"
|
||||
description = "ComfyUI-EdgeTTS is a powerful text-to-speech node for ComfyUI, leveraging Microsoft's Edge TTS capabilities. It enables seamless conversion of text into natural-sounding speech, supporting multiple languages and voices. Ideal for enhancing user interactions, this node is easy to integrate and customize, making it perfect for various applications."
|
||||
version = "1.2.2"
|
||||
version = "2.0.0"
|
||||
|
||||
license = {file = "LICENSE"}
|
||||
dependencies = [
|
||||
"edge-tts>=7.0.0",
|
||||
"torchaudio",
|
||||
"torchcodec==0.9",
|
||||
"torchcodec>=0.9",
|
||||
"openai-whisper>=20231117",
|
||||
"numpy",
|
||||
"torch",
|
||||
|
||||
+1
-1
@@ -1,6 +1,6 @@
|
||||
edge-tts>=7.0.0
|
||||
torchaudio
|
||||
torchcodec==0.9
|
||||
torchcodec>=0.9
|
||||
openai-whisper>=20231117
|
||||
numpy
|
||||
#mutagen>=1.47.0 # For audio metadata handling
|
||||
|
||||
@@ -1,5 +1,26 @@
|
||||
# ComfyUI-EdgeTTS Update Log
|
||||
|
||||
## v2.0.0 (2026/10/13)
|
||||
|
||||
### Service Reliability & Protocol Update
|
||||
- **Next-Gen Speech Handshake Protocol**:
|
||||
- **Why We Changed It**: Upstream service protocol changes previously caused speech synthesis requests to fail with authorization rejection errors.
|
||||
- **Seamless Continuity**: Fully upgraded internal connection negotiations to maintain seamless, continuous access to all Edge TTS neural voices without requiring creators to reconfigure environments or modify existing workflows.
|
||||
- **Automated Fallback & Retry**: Added rapid network jitter recovery to automatically retry transient failed handshakes, preventing pipeline interruptions during batch generation.
|
||||
|
||||
### Audio Decoding Resilience
|
||||
- **Multi-Backend Stream Decoding**: Integrated robust audio decoding with automatic format fallbacks, eliminating runtime crashes caused by system audio codec mismatches.
|
||||
|
||||
### New Features & Subtitle Generation
|
||||
- **Whisper SRT Subtitle Output**:
|
||||
- **Why We Added It**: Creators frequently need precise timecoded subtitles to feed into video captioning and multimodal animation workflows.
|
||||
- **Dual Output**: Added a dedicated `SRT` output socket to the Whisper STT node alongside raw `TEXT`, producing clean time-synced subtitle blocks (`00:00:01,000 --> 00:00:03,500`) with zero external dependencies.
|
||||
|
||||
### Node Architecture & Standardization
|
||||
- **Standardized Port Casing**: Aligned all node output sockets with 1038lab standards (`AUDIO`, `FILEPATH`, `TEXT`, `SRT`) for seamless chaining across audio and video workflows.
|
||||
- **Enhanced Parameter Guidance**: Added descriptive tooltips to all input controls across Edge TTS, Save Audio, and Whisper STT nodes.
|
||||
- **Dynamic Node Discovery**: Upgraded internal module loader for faster startup and clean alphabetical node organization in ComfyUI menus.
|
||||
|
||||
## v1.2.2 (2026/1/25)
|
||||
|
||||
### Dependency Compatibility Fixes
|
||||
|
||||
Reference in New Issue
Block a user