BIT EXACT ROUND TRIP IMAGE COMPRESSION

Pixspan Data’s Bit Exact Round Trip™ image compression is focused on four primary goals in compressing full resolution images:

• Ensuring that the fidelity of the original images is perfectly preserved
• Providing best-in-class image data savings (compression ratio reduction) to accelerate transfers and reduce storage costs
• Delivering the fastest decompress times to make compression seamless in imaging workflows
• Scaling in performance to the available compute platform

Full-Resolution imagary represents a fundamental asset and work product across many industries, for example:

Compression Savings

Pixspan Data’s patented compression software is a file format aware technology, focusing on compressing the raster portion of an image file, which represents almost the entire file size. Pixspan Data achieves best-in-class compression performance, significantly outpacing the performance of other mathematically lossless compression algorithms on image files. Pixspan Data compression generally saves 50% to 80% or more of storage space and transmission time for images produced by various industry verticals and raster formats including Bayer Pattern based camera raw images.

Decompress Speeds

Pixspan Data compression can be incorporated into real time workflows that require high throughput and low latency, for example satellite image analysis and special effect artistry. In workflows like these, Pixspan Data compression is transparent to the analysts/artists, i.e. they are unable to detect that image compression is being employed throughout the IT infrastructure. This seamless experience is due to Pixspan Data’s focus on providing the fastest decompression speeds.

Pixspan Data’s algorithms do not employ DCT or wavelet steps, which consume large processing resources. As a result, Pixspan Data runs several times faster than other methods, requiring less computing to achieve the same speed.

Performance Scaling

Pixspan Data Software operates on any Intel or AMD based off-the-shelf server or virtual instance. It also can be hosted on an ARM-based Mac processor natively. Compression and decompression speeds can be further accelerated with CUDA-compatible GPUs. Approximate performance metrics for various compute platforms are shown in the following table: