Hacker News Reader

Comment by kurthr | Hacker News Reader

With 100+Hz displays it's not that hard to do temporal dithering as well. Your cones are surprisingly low bandwidth (why old color TVs even worked at 30Hz), while your rods provide danger/flicker cues outside the fovea.

Getting an extra 2bits of hue (ab) while maintaining luminance (L) is quite doable except at the chroma and brightness extremes where your eye mostly ignores them anyway. That could be done pretty high in the display stack. I'd also say that the DACs in many displays are capable of higher chroma resolution, but gamma non-linearity eats up a bit dynamic range.

View on HN

Old TV wasn't 30Hz. 60i isn't the same as 30p.

Some TVs and monitors do temporal dithering for you... Accept 8-bit input, and temporal dither it for the 6-bit display doesn't look that bad. Probably extends well to 10-bit input.

Well, aktuuallly it's 59.94Hz and 29.97Hz, exactly BECAUSE the NTSC color "burst" carrier is only 3.579545MHz (note these are relatively prime 5×7×9/(8×11) MHz) and it has to be an exact multiple of 15734.2637Hz to generate the 30Hz color frames at ~487 lines (excluding vertical blanking lines which take it to 525). Of course those were interlaced to two sets of 262.5 lines.

Now, I'm hoping you're an old school CRT nerd, and if so then you'd also agree that the original color phosphor glow down periods were several ms and the eye's response to changing color is marginal at best. Uncorrelated subpixel scale dithering works just fine at 30Hz.

You're downplaying "Of course those were interlaced to two sets of 262.5 lines". That is what makes interlaced video 59.94 different images per second, and the difference between 30hz updates and 60hz updates is most definitely noticeable.

The eye's response to changing color is slow, but the response to changing luminance is very fast.

You have expertly summarized the first and last lines of my original post about Color NTSC. Interlacing is effectively dithered color super-resolution at low frame rate with blur from fast motion line tearing at the interlace rate.

   Your cones are surprisingly low bandwidth (why old color TVs even worked at 
   30Hz), while your rods provide danger/flicker cues outside the fovea.
   
   Uncorrelated subpixel scale dithering works just fine at 30Hz.

Reply on news.ycombinator.com