( Blog )

New release boosts GPU JPEG2000 encoding by 20%

BRNO – Comprimato a tech startup has announced the new release of its JPEG2000 GPU encoder and decoder. The system, which has already been outperforming other compression methods, is now another 20% faster at encoding. Thanks to NVIDIA CUDA, Comprimato can encode 4K signal at speeds up to 83 frames per second (for further details, run the performance calculator).

Thanks to the new low-level code, the JPEG2000 GPU Codec is now even more parallel than before. This helps to speed up encoding by another 20% and shows that we still have the capacity to improve GPU-based JPEG2000 compression,“ said Martin Jirman, Comprimato’s head of development. He added that the company will continue to further develop the code to reach even higher numbers.

Comprimato is using NVIDIA’s CUDA a parallel programming and computing platform, for its code. In effect, it uses the capacity of the GPU to perform JPEG2000 encoding and decoding. This is not only effective, but helps to reserve the capacity of the computer’s processing unit for other tasks.

The philosophy of our company is created around the idea that many calculation and computing tasks can be performed by graphics cards. Thanks to CUDA, NVIDIA helps developers to reach this capacity,“ said Jiri Matela, Comprimato’s CEO and founder. „Comprimato is currently looking into ways in which other codecs could also be processed by GPU platforms,“ added Matela.

JPEG2000 is an image encoding system that uses state-of-the-art compression techniques based on wavelet technology. Its architecture lends itself to a wide range of uses from portable digital cameras to advanced pre-press, medical imaging and other key sectors. Compared to JPEG, JPEG2000 offers higher compression without compromising quality. Images compressed using the JPEG2000 format can be reverted back to their original size and quality without any loss of information, which makes it an excellent format for video processing.

by comprimato
January 19th, 2016