This repository provides a comprehensive tool for converting audio files to text. It leverages state-of-the-art machine learning models to deliver accurate and efficient transcriptions. Key features include:
- Audio Format Conversion: Easily convert audio files between different formats (e.g., MP3 to WAV).
- Segment Extraction: Extract specific segments from audio files for targeted transcription.
- Speaker Identification: Identify and label different speakers in the audio.
- Easy-to-use Interface: Simplified commands for audio processing and transcription.
- High Accuracy: Utilizes advanced models like OpenAI's Whisper for precise transcriptions.
- Customizable Parameters: Adjust settings like language, model size, and number of speakers.
- Memory Efficient: Handles large audio files efficiently by processing in segments.
- Clone the repository:
git clone https://github.com/manuelarguelles/speechtotext_v2.git cd speechtotext_v2 - Install the necessary packages:
pip install -r requirements.txt
from audio_converter import convert_to_wav
# Convert an MP3 file to WAV format
convert_to_wav('path/to/input.mp3', 'path/to/output.wav')
## Extracting Audio Segment
from audio_extractor import extract_audio_segment
# Extract a segment from an audio file
extract_audio_segment('path/to/input.wav', start_time=120, end_time=720, output_path='path/to/output_segment.wav')
#Transcribing Audio
from transcriber import AudioTranscriber
## Create an instance of the AudioTranscriber
transcriber = AudioTranscriber('path/to/audio.wav', model_size='tiny', num_speakers=1)
# Transcribe the audio file
segments = transcriber.transcribe()
# Save the transcription to a text file
transcriber.save_transcription(segments, 'path/to/transcription.txt')