GET3D (Nvidia)

3d shape learning from images

The Pop2Piano tool is a generative model for creating high quality 3D textured shapes from images, specifically for piano covers of popular modern songs. The system uses a deep learning algorithm to map audio features from the original song onto the piano cover, rendering a stereo audio output with the piano on one side and the original song on the other. The Pop2Piano tool can be accessed using an intuitive demo interface that allows the user to select from a range of popular songs and arranger styles to create their own unique covers. The Pop2Piano tool is backed by a research paper and comes with fully documented source code that is accessible through the provided links. With a high-quality output sound and a wide range of customization options, the Pop2Piano tool is a powerful creative resource for musicians, music producers, and anyone else looking to experiment with audio generation and cover creation.

Features:
- Generates high quality 3D textured shapes for piano cover generation from images
- Uses a deep learning algorithm to map audio features from the original song onto the piano cover
- Renders a stereo audio output with the piano on one side and the original song on the other
- Can be accessed using an intuitive demo interface
- Allows the user to select from a range of popular songs and arranger styles
- The Pop2Piano tool is backed by a research paper
- Comes with fully documented source code that is accessible through provided links
- Has a high-quality output sound and a wide range of customization options
- A powerful creative resource for musicians, music producers and anyone else looking to experiment with audio generation and cover creation
- Provides qualitative comparisons with samples and demo videos.