Question

I am writing a medical application, where i have packets coming in IP interface. Which has following definition for pixel skeleton decoding. I can decide what i pick in the main frame to have different pixel algorithm. But i do not understand the logic properly.

What does this 3 formats mean? Which one would you use?

 8 bit  YUV (4:2:2) 
 10 bit YUV (4:2:2)
 10 bit RGB (4:4:4)

Answer 1

It exists to manage the bandwidth used by color images. Since the typical human eye is more sensitive to luminosity than color, the two color channels (U and V) can be given reduced bandwidth compared to the luminosity channel (Y) while maintaining a certain level of overall quality.

I cannot answer which format is best for your application unless I know more about your application.

• 8 bit YUV 422 is the most compatible format.
• 10 bit YUV 422 achieves a slightly higher quality (especially w.r.t. banding) for a given bandwidth, but support is limited and processing is more difficult. (This assumes some form of lossy compression — it of course uses more bandwidth if uncompressed.)
• 10 bit RGB 444 is often preferable for writing video effects, but undesirable for distributing finished video.