app.py

import streamlit as st
from transformers import pipeline
import sounddevice as sd
import numpy as np

# Set the title with a colorful header
st.markdown("<h1 style='text-align: center; color: #FF6347;'>🎤 Text-to-Speech with Bark Model 🎶</h1>", unsafe_allow_html=True)

# Add a description with a stylish subtitle
st.markdown("<h3 style='text-align: center; color: #4682B4;'>Convert your text into lifelike speech instantly!</h3>", unsafe_allow_html=True)
st.markdown("---")

# Initialize the text-to-speech pipeline with the Bark model
synthesizer = pipeline("text-to-speech", model="suno/bark")

# Add an input text box with a vibrant background
text = st.text_area(
    "Enter the text you want to convert to speech:",
    placeholder="Type something interesting...",
    height=150,
)

# Button to generate and play speech with a custom style
if st.button("🎙️ Convert to Speech"):
    if text.strip() == "":
        st.error("Please enter some text before converting!")
    else:
        with st.spinner("Generating speech... 🎶"):
            # Generate speech
            speech = synthesizer(text)

            # Convert the audio data to 16-bit PCM format
            audio_data = speech["audio"]
            audio_data = np.int16(audio_data / np.max(np.abs(audio_data)) * 32767)  # Normalize and convert to int16

            # Play the generated audio immediately
            sampling_rate = speech["sampling_rate"]
            sd.play(audio_data, sampling_rate)
            sd.wait()  # Wait until the audio has finished playing

            # Success message
            st.success("🎉 Speech generation complete!")

# Add a colorful footer
st.markdown("<footer style='text-align: center; color: #708090; font-size: 12px;'>Created with ❤️ using Streamlit and Transformers</footer>", unsafe_allow_html=True)