Verbatim

import streamlit as st import speech_recognition as sr import pyttsx3 import numpy as np import sounddevice as sd import noisereduce as nr import wave import threading

def noise_reduce(audio_data, sample_rate): #Apply noise reduction to improve speech recognition accuracy. reduced_audio = nr.reduce_noise(y=audio_data, sr=sample_rate) return reduced_audio

def recognize_speech(stop_event): recognizer = sr.Recognizer() with sr.Microphone() as source: st.write("Listening...") while not stop_event.is_set(): audio = recognizer.listen(source) try: text = recognizer.recognize_google(audio, language='en-NG') st.session_state['transcription'] += text + " " except sr.UnknownValueError: st.session_state['transcription'] += "[Unrecognized Speech] " except sr.RequestError: st.session_state['transcription'] += "[Error: Connection Issue] "

def text_to_speech(text): engine = pyttsx3.init() engine.say(text) engine.runAndWait()

def start_recording(): if 'stop_event' in st.session_state and st.session_state['stop_event'].is_set(): st.session_state['stop_event'].clear() else: st.session_state['stop_event'] = threading.Event() st.session_state['transcription'] = "" threading.Thread(target=recognize_speech, args=(st.session_state['stop_event'],)).start()

def stop_recording(): if 'stop_event' in st.session_state: st.session_state['stop_event'].set()

def main(): st.set_page_config(page_title="Verbatim - Speech to Text", layout="wide")

st.sidebar.title("Menu")
st.sidebar.button("User Profile")
st.sidebar.button("Settings")
st.sidebar.button("Session History")

st.title("Verbatim")
st.write("Speech To Text Program Also Supports Nigerian English")

col1, col2 = st.columns(2)

with col1:
    if st.button("Start Transcription"):
        start_recording()
    if st.button("Stop Transcription"):
        stop_recording()

with col2:
    if st.button("Read Aloud"):
        text_to_speech(st.session_state.get('transcription', ''))

st.session_state['transcription'] = st.text_area("Transcribed Text:", st.session_state.get('transcription', ''), height=300)
st.download_button("Download Transcript", st.session_state.get('transcription', ''), "transcription.txt")

if name == "main": main() import streamlit as st import speech_recognition as sr import pyttsx3 import time import pyaudio

def recognize_speech(): recognizer = sr.Recognizer() with sr.Microphone() as source: recognizer.adjust_for_ambient_noise(source) st.session_state['status'] = "Listening..." try: audio = recognizer.listen(source, timeout=5) text = recognizer.recognize_google(audio, language='en-NG')

        if 'transcription' not in st.session_state:
            st.session_state['transcription'] = ""
        
        st.session_state['transcription'] += text + " "
        st.session_state['updated'] = True
    
    except sr.UnknownValueError:
        st.session_state['transcription'] += "[Unrecognized Speech] "
    except sr.RequestError:
        st.session_state['transcription'] += "[Error: Connection Issue] "
    except Exception as e:
        st.session_state['transcription'] += f"[Error: {str(e)}] "

def text_to_speech(): engine = pyttsx3.init() engine.say(st.session_state.get('transcription', '')) engine.runAndWait()

def start_recording(): st.session_state['recording'] = True st.session_state['transcription'] = "" st.session_state['status'] = "Recording..."

def stop_recording(): st.session_state['recording'] = False st.session_state['status'] = "Stopped"

def main(): st.set_page_config(page_title="Verbatim - Speech to Text", layout="wide")

st.sidebar.title("Menu")
st.sidebar.button("User Profile")
st.sidebar.button("Settings")
st.sidebar.button("Session History")

st.title("Verbatim")
st.write("Speech To Text Program Also Supports Nigerian English")

col1, col2 = st.columns(2)

with col1:
    if st.button("Record & Transcribe"):
        start_recording()
    if st.button("Stop Transcription"):
        stop_recording()

with col2:
    if st.button("Read Aloud"):
        text_to_speech()

st.text_area("Transcribed Text:", st.session_state.get('transcription', ''), height=300, key="transcription")
st.download_button("Download Transcript", st.session_state.get('transcription', ''), "transcription.txt")

st.write(f"Status: {st.session_state.get('status', 'Idle')}")

if st.session_state.get('recording', False):
    recognize_speech()
    time.sleep(0.1)
    st.rerun()

if name == "main": if 'transcription' not in st.session_state: st.session_state['transcription'] = "" if 'recording' not in st.session_state: st.session_state['recording'] = False if 'status' not in st.session_state: st.session_state['status'] = "Idle" main()