Tom Barry wrote: > But mtf must be more than just a multiplier factor. > If I take a sinusoidal signal and multiply it by .5 > then the amplitude is only half as much but no > information has yet been lost, barring quantization. > I could multiply it by 2 later and get it back. It > is only when errors (random info) are introduced > that information goes away. But I'm not sure how > this figures in gain or mtf and I'm not sure how > that part is supposed to be measured. > > I realize there are many sources of errors, A/D > noise, quantization noise, various optical, > electronic, and digital filters etc. and each > contributes in complex ways to lowered mtf (either > case). But I still can't quite visualize what is > being measured when "contrast" is lowered. I don't know if you're more comfortable with audio, Tom. Audio analogy: the gain-bandwidth curve. An amp might provide 100 watts of output at 1 KHz. Now you crank up the frequency, and you might find that for an input signal at 30 KHz, that same amp can only manage 50 watts. Crank the frequency up further, and pretty soon the output will drop to 0. At 1 KHz, where the amp gives outputs of 0 to 100 watts, the amp can reproduce 20 dB of "contrast." That is, the difference between softest and loudest sounds it can reproduce is 10*log(100) or 20 dB. But at 30 KHz, the difference is only 10*log(50), or 17 dB. The higher the frequency, the less contrast that amp can reproduce. So someone can tell you his audio gear reproduces 100 KHz, but at what output level? Perhaps only 5 watts by the time you're up to 100 KHz, or 7 dB of "contrast," compared with 20 dB of contrast at 1 KHz. The amp with the flattest bandwidth will sound better (sharper) than one where the bandwidth curve droops at higher frequencies. (Beyond 20 KHz or so, people don't care. Dogs might.) With a lens, film, sensor, etc., it's the same thing, but in two dimensions. If you measure the maximum lines/mm the medium can resolve, the next question is "but at what contrast level?" Just like the max frequency the amp can reproduce, "but at what output level?" In images, this number will vary over the area of the image. If your lens, say, can "resolve" 90 pairs/mm, just how much contrast are you getting between the black and white lines? If, in the 90 lines/mm chart, the black and white lines are both mostly gray, then even though in principle the image detail is there, the picture will not look very sharp because the small detail is washed out. So, the MTF curve is to images what the gain-bandwidth curve is to audio. Contrast vs frequency. I'm not sure how the error questions crept into this, but it seems to me that once the contrast has been lost by the medium, it can't be recreated accurately anymore. Just pushing the gain after the fact will amplify noise along with signal. Bert ---------------------------------------------------------------------- You can UNSUBSCRIBE from the OpenDTV list in two ways: - Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word unsubscribe in the subject line.