Spaces:
Sleeping
Sleeping
File size: 4,064 Bytes
35f1d7f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 | import gradio as gr
import edge_tts
import asyncio
import tempfile
from piper_engine import PiperEngine
# -----------------------------
# Load Piper Once
# -----------------------------
piper_engine = PiperEngine(
"models/en_US-amy-medium.onnx"
)
# -----------------------------
# Voice Loader (Edge)
# -----------------------------
async def get_voices():
voices = await edge_tts.list_voices()
return {
f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v["ShortName"]
for v in voices
}
# -----------------------------
# Edge TTS (Neutral)
# -----------------------------
async def edge_tts_engine(text, voice, rate, pitch):
if not text.strip():
return None, "Please enter text."
if not voice:
return None, "Please select a voice."
voice_short = voice.split(" - ")[0]
rate_str = f"{rate:+d}%"
pitch_str = f"{pitch:+d}Hz"
communicate = edge_tts.Communicate(
text,
voice_short,
rate=rate_str,
pitch=pitch_str,
)
with tempfile.NamedTemporaryFile(
delete=False,
suffix=".mp3",
) as tmp:
path = tmp.name
await communicate.save(path)
return path, None
# -----------------------------
# Piper TTS (Expressive)
# -----------------------------
def piper_engine_wrapper(text, rate, pitch):
if not text.strip():
return None, "Please enter text."
speed = 1.0 + rate / 100 # rate from slider
pitch_scale = 1.0 + pitch / 50
audio = piper_engine.synthesize(text, speed, pitch_scale)
return audio, None
# -----------------------------
# Unified Interface
# -----------------------------
async def unified_tts(
text,
voice,
rate,
pitch,
engine,
):
if engine == "Edge (Neutral)":
return await edge_tts_engine(
text,
voice,
rate,
pitch,
)
else:
return piper_engine_wrapper(
text,
rate,
pitch,
)
# -----------------------------
# UI Builder
# -----------------------------
async def create_demo():
voices = await get_voices()
with gr.Blocks(analytics_enabled=False) as demo:
gr.Markdown("""
# 🎙️ AI Text-to-Speech Lab
Compare traditional and expressive AI voices.
""")
# Text
text_input = gr.Textbox(
label="Text",
lines=6,
placeholder="Paste lecture notes, narration, or scripts here...",
)
# Engine Selector
engine_radio = gr.Radio(
["Edge (Neutral)", "Piper (Expressive)"],
value="Edge (Neutral)",
label="Generation",
)
# Voice (Edge Only)
voice_dropdown = gr.Dropdown(
choices=[""] + list(voices.keys()),
label="Edge Voice",
value="",
)
with gr.Row():
rate_slider = gr.Slider(
-50, 50, 0,
step=1,
label="Emotion / Speed",
)
pitch_slider = gr.Slider(
-20, 20, 0,
step=1,
label="Intonation / Pitch",
)
generate_btn = gr.Button(
"Generate Audio",
variant="primary",
)
audio_output = gr.Audio(
label="Output",
)
warning_md = gr.Markdown()
generate_btn.click(
fn=unified_tts,
inputs=[
text_input,
voice_dropdown,
rate_slider,
pitch_slider,
engine_radio,
],
outputs=[
audio_output,
warning_md,
],
)
return demo
# -----------------------------
# Runner
# -----------------------------
async def main():
demo = await create_demo()
demo.queue(default_concurrency_limit=20)
demo.launch()
if __name__ == "__main__":
asyncio.run(main())
|