Home
This project is an image format that seems to have no use for now. It is built for people who want high-precision data storage, even if no standard software opens it yet
It uses actual math instead of just guessing pixels:
- Haar Wavelet Engine: Multi-level frequency layers. It prioritizes image structure over noise. When you compress it too much, it fails with painterly blurs instead of square JPEG blocks
- Reversible YCoCg-R: A high-efficiency color transform that provides better decorrelation and color accuracy than standard YCbCr
- Variable Bit-Depth: Fully supports 8-bit, 10-bit, and 16-bit pipelines. Good for HDR10 and professional photography
- LZMA Entropy Coding: Deep dictionary-based compression for maximum data density
- Progressive Loading: Chunked Bitstream structure allows for the low-resolution image to improve gradually as the file loads
- Self-Healing (Anti-Rot): XOR parity blocks can repair the file if your hard drive starts dying
- "True" ROI Decoding: Extract specific 512px regions from giant images without decoding the whole file
- Chrono-Layers: Store your undo history in one file so you can see how an image changed over time
- Watermarking: Hide invisible text secrets directly within the frequency layers