Jon,<br /><br />There are several ways to address the issue of Martian sky color.<br />1. To have a digital color camera on MER, like I have on my computer, size 2 cm, to take a picture, and send out several pixels of it. - Simplest, and the most accurate.<br />No calibration neeeded.<br /><br />But we are stuck with the BW filtered cameras, hence:<br /><br />2. Phenomenological <br />You have a set of filtered intensities, as being sent out from the MER. You want to transform them into pixel values.<br />So, you look for a formula that would do it.<br />Lets say you have guessed a right formula. How do you check it? You look at the calibration target, apply the formula, and see, if it gives a good estimation for the calibration colors. If you are satisfied on that account, that is all. The formula is cleared, and is used for getting the pixel values for any other set of filtered intensities, sky including. - Simple, and as accurate as it can be for the filtering approach.<br /><br />3. Theoretical<br />You have a set of filtered intensities, as being sent out from the MER. You want to transform them into pixel values.<br />You are saying: There is an actual specter, a wavelength distribution, that was transformed into those intensities by the MER camera.<br />Each intensity is a result of filtering of the actual specter by the particular filter, and then by interaction of the filtered specter with the photosensitive target.<br /><br />Bottom line, that you have <b>3 measured values</b>, and an <b>infinite number of parameters</b> that describe the original specter, which you want to recover as a theoretician, on your first step. It is impossibility, so whichever way you think or pretend you invented of doing it, is wrong way. As a second step, you think, you will find a way of converting the "recovered" specter into pixel values. Which is also impossibility, since now you are going to convert infinity of values into 3 values, and you never can check all the inputs for right output