Quick Start¶

from supertonic import TTS

# Note: First run downloads model automatically (~305MB)
tts = TTS(auto_download=True)

# Get a voice style
style = tts.get_voice_style(voice_name="F3")

# Generate speech (English - default)
text = "The train delay was announced at 4:45 PM on Wed, Apr 3, 2024 due to track maintenance."
wav, duration = tts.synthesize(text, voice_style=style, lang="en")
# wav: np.ndarray, shape = (1, num_samples)
# duration: np.ndarray, shape = (1,)

# Save to file
tts.save_audio(wav, "output.wav")

from supertonic import TTS

tts = TTS()
style = tts.get_voice_style("M1")

# English (default)
wav_en, _ = tts.synthesize("Hello, welcome to Supertonic!", voice_style=style, lang="en")

# Korean
wav_ko, _ = tts.synthesize("안녕하세요! 수퍼토닉에 오신 것을 환영합니다.", voice_style=style, lang="ko")

# Spanish
wav_es, _ = tts.synthesize("¡Hola! Bienvenido a Supertonic.", voice_style=style, lang="es")

# Portuguese
wav_pt, _ = tts.synthesize("Olá! Bem-vindo ao Supertonic.", voice_style=style, lang="pt")

# French
wav_fr, _ = tts.synthesize("Bonjour! Bienvenue sur Supertonic.", voice_style=style, lang="fr")

tts.save_audio(wav_ko, "output_korean.wav")

from supertonic import TTS

tts = TTS()

# List available voices
voice_list = tts.voice_style_names  # ex) ['M1', 'M2', 'F1', 'F2']

# Try different voices
for voice_name in voice_list:
    style = tts.get_voice_style(voice_name)
    wav, dur = tts.synthesize(f"This is the {voice_name} voice style demonstration.", voice_style=style)
    tts.save_audio(wav, f"output_{voice_name}.wav")

from supertonic import TTS
from pathlib import Path

tts = TTS()

# Load custom style from JSON
custom_style = tts.get_voice_style_from_path(Path("custom_voice.json"))
wav, dur = tts.synthesize("Using a custom voice style from JSON file.", voice_style=custom_style)

from supertonic import TTS

tts = TTS()
style = tts.get_voice_style("F1")

texts = {
    0.7: "This is slow speed demonstration.",
    1.0: "This is normal speed demonstration.",
    1.3: "This is fast speed demonstration.",
    2.0: "This is fastest speed demonstration.",
}

for speed, text in texts.items():
    wav, dur = tts.synthesize(text, voice_style=style, speed=speed)
    tts.save_audio(wav, f"output_{speed:.1f}_speed.wav")

from supertonic import TTS

tts = TTS()
style = tts.get_voice_style("M1")

texts = {
    2: "This is low steps demonstration.",
    5: "This is balanced steps demonstration.",
    10: "This is high steps demonstration.",
}

for steps, text in texts.items():
    wav, dur = tts.synthesize(text, voice_style=style, total_steps=steps)
    tts.save_audio(wav, f"output_{steps:02d}_steps.wav")

from supertonic import TTS

tts = TTS()
style = tts.get_voice_style("F1")

long_text = """
Artificial intelligence has transformed many fields.
From healthcare to transportation, AI systems are making impacts.
Natural language processing allows computers to understand human language.
These advances are opening up new possibilities.
"""

wav, dur = tts.synthesize(
    long_text,
    voice_style=style,
    max_chunk_length=300,
    silence_duration=0.3
)
tts.save_audio(wav, "output.wav")

from supertonic import TTS

tts = TTS()
text_processor = tts.model.text_processor

# Check if text is supported
text = "Hello World! Welcome to 世界."
is_valid, unsupported = text_processor.validate_text(text)

if not is_valid:
    print(f"Unsupported characters: {unsupported}")
    # Will show: ['世', '界'] (or similar)

# Get all supported characters
supported_chars = text_processor.supported_character_set
print(f"Supported characters: {sorted(list(supported_chars))}")

from supertonic import TTS

# Auto-detect (recommended)
tts = TTS()

# High-performance server
tts = TTS(intra_op_num_threads=12, inter_op_num_threads=12)

# Low-resource environment
tts = TTS(intra_op_num_threads=2, inter_op_num_threads=2)