483
Google does not stop trying to improve JPEG, despite the availability of other promising image formats. The company introduced Jpegli, an advanced JPEG encoding library that supports high backward compatibility. According to the company, Jpegli compresses images 35% more efficiently at high quality settings, which saves space and traffic, and speeds up web page loading. It is reported that web developers can easily integrate the new library into their workflows without compromising speed or resource consumption.
Google cites the following benefits of Jpegli. The library is fully compatible with the original JPEG standard, as well as API/ABI compatible with libjpeg-turbo and MozJPEG. That is, Jpegli can be easily embedded in applications that already use these libraries. Jpegli's encoding speed is comparable to libjpeg-turbo and MozJPEG. This means that web developers can integrate the new library into their workflows without compromising speed or resource consumption.
When images are compressed or decompressed with Jpegli, more accurate and "psychovisually efficient" calculations are performed, and images appear sharper with fewer artifacts.
Jpegli can be encoded using more than 10 bits per component. Traditional JPEG encoding solutions offer only 8-bit dynamics per component, resulting in visible banding artifacts on smooth gradients. At the same time, the resulting images are fully compatible with 8-bit viewing programs, in particular with browsers.
Jpegli compresses images more efficiently than traditional JPEG encoders, which saves space and traffic, and speeds up web page loading.
During the development of Jpegli, a number of new methods were used to reduce noise and improve image quality. The development of the JPEG XL format was partially used, which the company refused to support in Google Chrome after long-term testing due to the "lack of significant benefits and interest in the community".
The developers also ran a test where people compared pairs of images encoded using Jpegli, libjpeg-turbo, and MozJPEG. The results showed that Jpegli can compress high-quality images 35% better than traditional JPEG codecs.
