NarraPhon: Automated Generation of Educational Videos from PowerPoint Slides

NarraPhon is a system designed to automatically convert PowerPoint (.pptx) presentations into engaging educational videos. It leverages AI technologies, including the StyleTTS2 text-to-speech model and the Whisper speech recognition model, to generate natural-sounding audio narrations and synchronized subtitles. This project aims to simplify the creation of multimedia educational content, making it more accessible to educators and learners alike.

Project Description

This project addresses the challenge of efficiently converting existing PowerPoint-based educational materials into video formats. Creating educational videos often requires significant time, resources, and technical expertise in video editing. NarraPhon automates this process, enabling educators to easily transform their presentations into dynamic and accessible video content.

Key Features:

• Automated Video Generation: Converts PowerPoint slides into video format with synchronized audio and optional subtitles.
• High-Quality Speech Synthesis: Utilizes the StyleTTS2 model to generate natural-sounding speech from slide notes or content, with customizable voice styles and parameters.
• Subtitle Generation: Employs the Whisper model for automatic speech recognition to create accurate and synchronized subtitles, enhancing accessibility.
• Modular Architecture: Designed with modular components (Extraction, Speech Generation, Subtitle Generation, Video/Audio Composition) for maintainability and scalability.
• User-Friendly Interface: Provides a Gradio-based web interface for easy file upload, parameter customization, and video generation.
• Efficient Content Creation: Significantly reduces the time and effort required to produce educational videos compared to manual methods.

Motivation:

The increasing demand for online learning resources necessitates efficient content creation methods. This project is motivated by the need to:

• Enhance Accessibility: Make educational content available to a wider audience, including those who prefer video-based learning.
• Improve Engagement: Offer a more dynamic and engaging learning experience compared to static slides.
• Save Time and Resources: Reduce the manual effort involved in video production.
• Increase Content Reach: Facilitate the distribution of educational materials on various online platforms.

Potential Impact:

NarraPhon has the potential to:

• Revolutionize educational content creation: By providing an easy-to-use tool for transforming PowerPoint presentations into videos.
• Democratize video production: Enable educators with limited video editing experience to create high-quality videos.
• Support diverse learning styles: Cater to visual and auditory learners.
• Expand the reach of educational materials: Facilitate sharing across digital platforms.

For a more in-depth look at the project's architecture, implementation details, and code examples, please refer to the [project report](link to your report if you have one).

Prerequisites

Before you begin, ensure you have the following installed on your system:

• Git: For cloning the repository and managing large files.
• wget: For downloading Miniconda.
• Bash/Zsh: The installation instructions assume you are using either Bash or Zsh as your shell.

Installation

Follow these steps to set up the environment and install NarraPhon:

1. Install Miniconda

This will create a miniconda3 directory in your home folder and install Miniconda into it.

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh
~/miniconda3/bin/conda init bash
~/miniconda3/bin/conda init zsh

Important: Restart your shell (e.g., close and reopen your terminal) for the changes to take effect.

2. Create and Activate Conda Environment

This creates a new Conda environment named "mooc" with Python 3.10.12 and activates it.

conda create -n mooc python=3.10.12
conda activate mooc

3. Clone the NarraPhon Repository and Install System Dependencies

Update the package list and install necessary system packages.

apt update

Clone the project repository and navigate into it.

git clone https://github.com/Tugaytalha/NarraPhon.git
cd NarraPhon

Install necessary system dependencies for text to speech, audio processing, and handling documents.

sudo apt-get install espeak-ng
sudo apt-get install ffmpeg
apt install poppler-utils
sudo apt-get install libreoffice

4. Download StyleTTS2 Model and Reference Audio

Install Git Large File Storage (LFS).

apt install git-lfs

Clone the StyleTTS2 model from Hugging Face (this may take some time due to the model's size).

git-lfs clone https://huggingface.co/yl4579/StyleTTS2-LibriTTS

Move the downloaded model and reference audio files to their appropriate locations within the project.

mv StyleTTS2-LibriTTS/Models .
mv StyleTTS2-LibriTTS/reference_audio.zip .
unzip reference_audio.zip
mv reference_audio Demo/reference_audio

5. Install Python Packages

Install the required Python packages, including the OpenAI Whisper model.

pip install -r requirements.txt
pip install git+https://github.com/openai/whisper.git

Usage

After completing the installation steps, you are now ready to use NarraPhon.

(Add instructions on how to use your application here. For example, you might have a script that needs to be run or specific commands to generate speech).

Example:

python app.py

After running the script, you should see the link to the local server and a live DNS where you can access the web interface.

Troubleshooting

• Conda environment issues: If you encounter problems with the Conda environment, ensure you have restarted your shell after installing Miniconda. You can also try deactivating and reactivating the environment:

  conda deactivate
  conda activate mooc

• Model not found: Double-check that you have correctly moved the Models directory and reference_audio files to the specified locations in step 4.
• Dependency errors: If you face errors during the installation of Python packages, make sure you are using the correct Python version (3.10.12 as specified in the Conda environment).
• ImageMagick (libmagic) Issues: Some features may require ImageMagick. If you encounter errors related to ImageMagick, follow these steps:
  1. Installation:

     sudo apt-get update
     sudo apt-get install imagemagick

     Verify installation:

     convert --version

  2. Configuration: Add to your ~/.bashrc or ~/.zshrc:

     export IMAGE_MAGICK_BINARY="/usr/bin/convert"

     Reload configuration:

     source ~/.bashrc  # or source ~/.zshrc

  3. Security Policy (if needed): If you still have issues, it might be due to ImageMagick's security policy.
     • Locate the policy file (usually /etc/ImageMagick-6/policy.xml or /etc/ImageMagick/policy.xml).
     • Edit with root privileges (e.g., sudo nano /etc/ImageMagick-6/policy.xml).
     • Change rights="none" to rights="read|write" for relevant policies like path and delegate. Example:

       <policy domain="path" rights="read|write" pattern="@*" />
       <policy domain="delegate" rights="read|write" pattern="*" />

     • Save the file and restart any services using ImageMagick.

Contributing

Contributions are welcome! If you encounter any issues or have suggestions for improvements, please open an issue or submit a pull request on the GitHub repository.

Acknowledgements

• StyleTTS2: We thank the creators of StyleTTS2 for their excellent work.
• OpenAI Whisper: We acknowledge the developers of Whisper for their contribution to speech recognition.