dave@imax.com (Dave Martindale) (11/20/89)
In article <10135@ucsd.Edu> brian@ucsd.edu (Brian Kantor) writes: > >Recall that the broadcast NTSC luminance bandwidth is about 4.2 MHz, >I is 1.5 MHz, and Q is .5 MHz, so it's not necessary to store lots of >color information if you're dealing with NTSC in and out. > >A back-of-the-envelope calculation (divide the bandwidth by the >horizontal scan frequency) will give you 267 pixels per line luminance, >95 I, and 31 Q: a rough indication that these pixel numbers are >probably about right - you lose just a bit of luminance, and you're >actually storing the chroma at slightly more resolution than needed. >(Don't forget Dr. Nyquist: sample at twice the max frequency.) I assume that you mean 267 *cycles*/line, which is 534 *pixels*/line? But you don't need to store pixels for the horizontal blanking interval - each scan line has 52.656 microseconds of picture and 10.9 microseconds of picture. So, for a 4.2 MHz bandwidth, you need to sample at 8.4 MHz, which is 119 ns/sample. That gives 442 pixels/scanline minimum, so 512 isn't a bad choice at all. The same calculation gives 158 pixels for I and 53 for Q. >Note that to save money, most television receivers do NOT directly >demodulate I and Q; instead a 33 degree shift of the chroma carrier >phase is used to demodulate R-Y, G-Y, and B-Y, with a reasonably good >approximation of the original RGB camera signal resulting. More precisely: to *properly* decode the chrominance signal, the receiver should demodulate along the I and Q axes, then use separate filters to band-limit I and Q to 1.5 MHz and 500 KHz, then use a delay line to compensate for the difference in delay between the two filters, then "rotate the coordinate system" from I/Q back to R-Y/B-Y using amplifiers and a few resistors. Instead, almost every consumer receiver in existence filters the whole chrominance signal to 500 KHz, then demodulates along the R-Y and B-Y directions directly. No delay line, only one filter instead of two, no inverting amplifier or resistor matrix - but 2/3 of the detail in the I signal has been discarded. Most consumer receivers also don't actually decode R-Y and B-Y at 90 degrees to each other. They use a wider angle, so that colour shifts due to poor transmission conditions cause less change in "flesh tones" in the picture - but at the cost of altering almost all colours in the picture, all the time. In article <1192@radius.UUCP> pierce@radius.UUCP writes: > There are additional problems with NTSC as well but I'm not sure I > want to go in to them. Heres a fun fact, though: Broadcast NTSC is > specced for a gamma of 2.2 and most TVs have a gamma of 2.8. This is deliberate. The idea is that the picture you see on the receiver has a gamma of 2.8/2.2 = 1.27 compared to the original scene. Subjective tests indicated that this artificially-high contrast was necessary to make the picture look subjectively "normal" when viewed in a room with dim lighting. For comparison, photographic transparencies (slides) and motion picture films are designed to have a gamma of about 1.5 to look "normal" in a very dark room. This has interesting implications for people who are doing "gamma correction" for film recorders - what is the value of gamma you should "correct" for? Depends on the application. Dave