Galileo Imagery, I couldn't find a topic not specific to one moon.... Options

[eoincampbell]

Really brilliant work here, much appreciated...

[volcanopele]

I thought that this one was the correct color and gamma
and have been planing on redoing the color on my map ( i white balanced it )
http://pirlwww.lpl.arizona.edu/~perry/io_images/c21.htm
http://pirlwww.lpl.arizona.edu/~perry/io_i...1ISCOLOR_01.png

Yes, true color...umm... Well gamma, quite frankly, gamma correction has always made images looks too washed out and bright, which maybe more correct, but I don't know, you lose what you gain with our fancy CCD and CMOS detectors.

As far as true color. No it isn't. That uses an Infrared image centered at 756 nm for red, though I tried to not stretch the colors unnecessarily.

[JohnVV]

color ???? "that is the question " when i did my map i did what i would do in the photo darkroom and balanced for white - then tweeked it for ascetics .I always thought that there was too much green in them
and used the
http://pdsimg.jpl.nasa.gov/data/cassini/ca...0K_0_0_SIMP.IMG
the eastern half
[attachment=22189:tx_1_0.jpg]

[volcanopele]

Wow, yeah, that is WAYYYY too red and dark.

Bjorn has a decent tutorial on Io's color at http://www.mmedia.is/bjj/3dtest/io/index.html .

Another true color approximation that is quite good is at http://photojournal.jpl.nasa.gov/catalog/PIA02308

[ugordan]

gamma correction has always made images looks too washed out and bright, which maybe more correct

It is indeed more correct. The images look more washed out because the objects of interest really are that washed out in reality. If you want to scrutinize the surfaces, higher contrast is great, but if you want more realistic looking images you need to apply gamma correction. I personally don't like the way the terminator appears in higher phase, uncorrected images. It's barely visible and erodes much of the visible disc toward the sunlit terrain.

you lose what you gain with our fancy CCD and CMOS detectors

It's actually more demanding of those fancy CCDs as it exposes any low level noise that would otherwise be drowned out in the darks. Voyager 8 bit is barely workable this way, Galileo is a bit better.

It's hard to illustrate what gamma correction does or why it's important with these distant objects as it's hard to relate to them. I'll give a more down to "Earth" example with a Phoenix image. The left side is uncorrected data, the right side is sRGB correct gamma, same calibrated image, click to enlarge:



You'll notice that apart from vastly higher contrast, there is color shifting present - the surface is redder (not simply more saturated color). It makes talking about "true color" uncorrected images sort of moot.

[tedstryk]

It is indeed more correct. The images look more washed out because the objects of interest really are that washed out in reality.

Not true. On any computer monitor, the dynamic range is much smaller than in real life. Because of this, the "washed out areas" would not appear that way to the eye. These gamma corrected versions are not "more realistic" than other versions. It is simply a matter of picking your poison and deciding which trade-offs you are willing to make.

[ugordan]

On any computer monitor, the dynamic range is much smaller than in real life. Because of this, the "washed out areas" would not appear that way to the eye.

Based on what? The original data is 8 bit. That means a difference of 1 DN translates into 1/256 of total dynamic range covered - determined by exposure, etc. There isn't dynamic fidelity that the eye might otherwise be able to pick out in the original data to begin with.

You are mixing display brightness output and target object contrast. The fact a monitor can not display true luminance of anything other than perhaps Uranus or Neptune systems does not diminish the value of accurate contrast portrayal. Looking at a gamma correct display of a bright target on a monitor is equal to looking at the target through a neutral density filter that dims the object. There is no such filter for the effect straight up RGB substitution gives because it's not a natural effect. It's an artifact of the processing just as incomplete calibration would be.

sRGB gamma was introduced with the sole purpose of mapping more DNs to lower brightness levels because that's where the eye is more sensitive and banding would otherwise be present with 8 bit data. Gamma correction is just a inverse of that so that linear radiometrically calibrated data is presented in the proper way on the screen. It doesn't change the contrast, it brings the actual contrast on screen to par with reality.

You're telling me the above left Phoenix image is just as real as the right one because the computer screen is too dim? All your digital camera images are sRGB gamma correct by default, even though the things you take pictures on Earth are vastly brighter than even Io. So how come they don't appear washed out? Cramming wider dynamic range into a computer screen output is the domain of HDR processing and even it doesn't inherently affect contrast. See example in this site.

The irony is that if the sRGB standard never introduced gamma, i.e. if it left the linear 8 bit approach, we wouldn't be having this discussion. We'd be talking about either natural color and/or contrast-enhanced or saturation-enhanced versions. When an image would be produced and it looked too bland (most of the time), it would be deliberately contrast-enhanced and clearly labeled as such.

[tedstryk]

I'm not confusing anything. Not only is the monitor not as bright, the darkest blacks aren't all that dark. As for the two Phoenix images, yes, I am saying that. One is faithful to the brightness of the reflection, one suppresses it in the name of preserving more interesting parts of the image.

[4th rock from th...]

My non technical argument would be that we use monitors to view every day photographs and video (of ordinary subjects) and the images look correct and "as it should be".
So if we replicate the same processing that our normal digital cameras do and apply it to raw image data, the results should in practice be the same as is if the images were originally taken with that camera.
As for the Phoenix image posted, I'm forced to say that the right version does look like a normal digital photo taken on Earth. Therefore, it's as real as any photo viewed on a PC screen. But this is just my opinion of course.


But more technically speaking, the left Phoenix image is NOT faithful to image reflexion values unless you forced your monitor to display also with a linear response!
If the monitor is sRGB, it will apply gamma to that image, resulting in a non-linear display. So the right version has the correct linear response.

[DrShank]

color ???? "that is the question " when i did my map i did what i would do in the photo darkroom and balanced for white - then tweeked it for ascetics .I always thought that there was too much green in them
and used the
http://pdsimg.jpl.nasa.gov/data/cassini/ca...0K_0_0_SIMP.IMG
the eastern half
[attachment=22189:tx_1_0.jpg]

nice job! what was the color data source / images?

[ugordan]

Not only is the monitor not as bright, the darkest blacks aren't all that dark.

Neither are they dark on the monitor in the non-corrected images, yet it makes areas that are actually not that dark in the real scene darker than they are - terminator region and darker albedo features.

As for the two Phoenix images, yes, I am saying that.

I would once again echo 4th rock's technical argument in that the left image is not technically correct. The same argument holds for every day images as it holds for planetary objects, be it bland objects like Venus or highly dynamic range like Earth.

One is faithful to the brightness of the reflection, one suppresses it in the name of preserving more interesting parts of the image.

I maintain my point that the gamma-correct version is more correct. It preserves brightness, contrast and hence color relationships of the real world object. Anything else is either a contrast enhancement or dilution by definition. Other, enhanced versions are certainly helpful to show minute detail, I use them a lot myself too, especially for low phase bland objects, but that doesn't change the fact those objects are visually bland (i.e. low albedo differences and those details really are minute differences). I would also add a point that having a monitor with a really bright output, sufficient to mimic the real scene would not make the images much more detailed to the eye because the eye is more sensitive to low light changes not bright objects - the reason gamma was introduced in the first place.

[volcanopele]

My biggest issue at this point seems to be that the way my images look in Photoshop CS4 looks quite a bit different when displayed in a browser. In Photoshop, the display looks non-linear, with my image not washed out by any stretch of the imagination, but the terminator region is quite as dark as my images appear in a browser.

So I would like to find a way to make the images that I make in photoshop show up the exact same way in a browser (even on the same computer).

[ugordan]

So I would like to find a way to make the images that I make in photoshop show up the exact same way in a browser (even on the same computer).

You could have a custom color profile for your PS different from sRGB. Have you tried seeing if you're allowed to do Edit->Convert to profile-> then select sRGB IEC... ?

A longer response would need you to determine which of the two looks you like better. If it's the browser one, they prefer sRGB (many don't even handle embedded profiles correctly) so you might want to make that the default working color space in Photoshop. If you prefer the Photoshop look, then the above method of conversion should do the trick, possibly as the last step before you save the image for web.

EDIT: Need to read more carefully what you write...

[volcanopele]

hmm, that doesn't seem to work, but then again, Firefox has always been a bit odd for a while in the way it handles color.

I'll give you a good example. Compare the thumbnail version of 10ISIOGLOC03 on http://pirlwww.lpl.arizona.edu/~perry/io_images/c10.htm with the full size version at http://pirlwww.lpl.arizona.edu/~perry/io_i...0ISIOGLOC03.png . On my browser, I like the way the thumbnail looks because it looks exactly the same way in Photoshop, and that is how I prefer it to appear. But in the full-size version, the terminator is way too dark.

And converting to a different profile (or using JPEG rather than PNG) didn't seem to work.

[ugordan]

Hmm. I'm stuck with a laptop right now with lousy contrast so can't make out any differences in the way the terminator appears, though the thumbnail image does seem to appear slightly more greenish compared to the full res. It could be the two have different embedded profiles. Do you generate both in Photoshop or?

I think there was a thread somewhere here about browser issues with color profiles, maybe one of the images is stuck with a profile (even though it's sRGB and FF can't handle it properly or something) instead of bare data.

Edit: yeah, I can see the difference in Photoshop if I resize the thumbnail. When opening the JPG thumbnail it says there's no embedded profile, but when opening the full PNG it says nothing.