Humans are visual creatures, and video is currently the dominant choice for online information consumption, accounting for over 80% of all online traffic. However, images will continue to play a significant role in the digital experience. Despite the rising costs associated with data bandwidth consumption, companies like Google are interested in optimizing this cost. Unfortunately, Google’s attempt to replace JPG, PNG, and GIF with its WebP image format did not succeed as intended.
Google decided to take a different approach and improve the JPEG format rather than compete with its widespread use. This led to the development of jpegli, a JPEG encoding library containing both an encoder and a decoder. Importantly, both the encoder and decoder comply with the original JPEG standard, ensuring compatibility with existing decoders.
While the technical aspects of jpegli may be complex, its benefits are more straightforward. Jpegli compresses images about 35% more than traditional JPEG codecs while maintaining visual quality, resulting in significant bandwidth savings. Jpegli also provides more precise and psychovisually effective calculations, enhancing image brightness and reducing perceptible artifacts. Additionally, it can encode images with 10+ bits per component without compatibility issues and reduces visible banding artifacts in slow gradients. Jpegli offers comparable speed to other encoding libraries without requiring additional computing resources.
A higher ELO score indicates a better overall performance
Google has released the full jpegli source code on GitHub, allowing interested individuals to explore and potentially utilize it today.