White Paper: Test & Measurement
H.264/H.265 VS JPEG XS
There are three ways to see video: uncompressed, lightly compressed, and highly compressed. For the small screen, highly compressed video is used. For the big screen and for content generation, uncompressed or lightly compressed video is used.
Two compression strategies emerged over the past 30 years. The Motion Picture Experts Group (MPEG) focused on transport in narrow band low-storage-capacity environments while maintaining acceptable image quality. The Joint Photographic Experts Group (JPEG) remained image quality focused thus requiring more storage capacity and wider transport bandwidths.
Endpoint distribution to the masses had to be delivered over narrow band channels (e.g., television, cable, streaming), which is better addressed by the MPEG approach. To achieve narrow band transport, compromises have been made on image quality. Image quality is good enough if a viewer of a movie on a small screen at a normal viewing distance does not notice the compression effects. However, if one stops on a single frame or watches it up close, errors will start to become obvious.
MPEG compression is computationally complex and requires a lot of buffering, resulting in a high latency between camera and display; many seconds is typical. JPEG XS is fast, low complexity encoder and delivers visually lossless images frame‑to‑frame. It has low buffering needs, resulting in low latency; milliseconds. The compromise to employ this technology is much more storage capacity and higher transport bandwidth; 150 mb/S (1080p/60 @ 20:1) to 600 mb/S (4K/60 @ 20:1). Over the past 10 years, bandwidth has exploded even for the cellular networks (5G) and storage cost is pennies.
The illustration compares the original uncompressed image with JPEG and MPEG 100:1 compression result. This paper explores these two encoding schemes to explain their differences, benefits, and compromises.
Don't have an account? Sign up here.