Quote: Mauro,
The variation in sensitivity of rods varies substantially more than 0.3 magnitudes. This is to be expected, since the dark adaptation process improves sensitivity by a factor of 90,000x ( I have no idea if that is an average or an extreme). I have, however, participated in many "field" experiments in which the visibility of faint stars in a field with known magnitudes varied by over a full magnitude from individual to individual. Momentary clearings could account for some of that difference, as could experience in simply focusing, but I think that degree of variation among experienced viewers is fairly common. Schaefer's work suggests that inexperienced viewers may lose a couple magnitudes and more. I have seen proof that details of magnitude 25 mpsas are visible in large scopes, despite the fact that this is literally 3 or more magnitudes fainter than the sky brightness at the same site (magnification being the reason, of course). Other people had a hard time seeing details at the magnitude 22 mpsas level on the same object There is reason to suspect the variation in the cones at daylit levels of illumination varies less. That makes sense from the standpoint of survivability in evolutionary terms.
What isn't well known from field tests (as opposed to lab tests) is if the threshold of mesopic vision varies by a similar amount. The red-green mix study to which you refer has a flaw in that the threshold for the yellow color of matching may have a large variation (not tested), and the color sensitivity at the lowest possible limit was not tested.
Since the eye is so much more sensitive to green light than red light, it's logical to assume that red has to be a lot brighter to be visible to the cones.
How deep into the twilight do you see the color red? My own car is red, and I cannot tell it is red under a dark sky (mag.21.5mpsas), but I definitely see it as red (not gray) at a brightness of sky of mag.18mpsas. I will do a test to see at what point the red becomes indeterminate as the twilight deepens.
The point is, I expect the threshold for red is indeed brighter than mag.20mpsas. But are your calculations for the nebula correct? I've observed a lot of galaxies where the peak brightness for the galaxy (per the RC3) is magnitude 18 mpsas, and this peak brightness is a lot fainter than 90% of the Orion nebula. I suppose what would help, here, is a drawing of the nebula marked with where each color is seen. Also what would help is a map showing surface brightness at each location in the nebula.
Perhaps there is no disagreement at all, and the nebula sections where color is seen are simply above the human threshold.
Don, maybe we agree more than it appears. The 0.3 mags is the result of the paper pointed out by DaveK in order to justify color vision. I read it and found that the genetic difference accounts for only 0.3 mags, at 1000 trolands retinal illuminatin (full photopic). What the paper measures is the difference of the response curves of two types L-cones at photopic level at 690 nm. Spectral relative sensitivity has nothing to do with mesopic threshold (this is another big mistake when infrared sensitivity -should it be proved- is invoked to demonstrate real red perceptions at faint levels). But since that paper was pointed out to provide a justification I thought it was worth to disclose what actually that paper says.
I too think that mesopic thresholds varies approximately 1 magnitude (somewhat less). In Fact I often write 17-18. That is also hinted by the data in the first two papers. Curiously the mesopic limit matches the full moon illumination. I think this is not casual.
The viewing of details at mpsas 25 is another matter. Even in a 22 mpsas sky background vieweing the details is more a matter of contrast detection that actual sensitivity. I understand that you own a SQM. Try in a closed room. I found I am able to see shadows cast by a small led onto the wall at surface brightness of 24+. I have not found exactly the thresholds of rods, but in a very dark room, the SQM read below the lower limit (once) and about 27 (If I recall correctly, another time) and that was absolutely dark. I think we can sense light at about 25 with the rods.
I read somewhere that the difference between rods and cones mostly lies in the fact that many rods are "binned" together and provides a cumulative stimulus to ganglion cells (thus they give low spatial resolution but high sensitivity). On the countrary cones are connect one to one (approximately I think) so that the maximum spatial resolution is achieved but at the expense of sensitivity. L-cones, being the most numerous, perhaps have some kind of "binning" and they are probaly the most sensitive.
The 90000 adaptation to light is achieved by regeneration of rhodospin. Again it has little to share with the maximum sensitivity (nobody has a 9.000.000 times range).
I have made tests with a pure monochromatic red light (633 nm) and with broadband red samples. I am not sure I can take the SQM readings as correct (SQM has a different color sensitivity than the conventional photopic curve, which in turn is different than that of L-cones). Thus I cannot say exactly what is the limit of my L-cones but two things are clear: monochromatic light provides a lesser coupling L-cones to M-cones and makes red visible to lower levels than the broadband red cards. A faint "red" say at 19 looks as brigh as a fain "blue" at 22 (quotions means the source color). If I mix a faint red 3 mags brighter that the blue I get an undoubtely gray. For example a monochromatic red at 18 appears colored. A blue at 21 appears "grue". When they are superimpose they no longer appear colored. Thus I suspect that even the faintest OIII can cancel real red percetion. If we see Ha somewhere with Ha filter, there is still the possibility that without filter some fainter green and blue (Hb and OIII) can cancel the red perception: more precisely: L-cones see red, rods see fainter grue but produce a stronger stimulus and the region is not recognized as red a low level sensorial fusion (but maybe then later it is assumed red because it is a mediul light).
I think we have the narrowband color photographs. Let us see where we see Ha.
About the computation of apparent brightness I will better explain in another reply.
PS in addition the V brightness of an object tells little of its Ha brightness. To evaluate red visibility we need to know brightness in the "red" and L-cones threshold. Noting the regions seen in Ha tells where L-cones are really stimuylated (which does not automatically mean that the region is red, beause the other colors are also to be accounted).
Previous
Index
Next
Flat
Re: Is it possible to see the nebula with colour?
Edit
Reply
Quote
