Mars with good seeing - Jan 7th (UK)

Hi all,

Not a lot of time to process the data set today so I'll put this wip
out first. The seeing was fairly steady last night from my location on
the UK's south coast. Here's the first quick process from the data set
put together last night when I should have gone to bed zzz.

http://www.digitalsky.org.uk/mars/20...59-49_RGB2.jpg

--
Pete Lawrence
http://www.digitalsky.org.uk

On 8 Jan, 09:15, Pete Lawrence wrote:
Hi all,

Not a lot of time to process the data set today so I'll put this wip
out first. The seeing was fairly steady last night from my location on
the UK's south coast. Here's the first quick process from the data set
put together last night when I should have gone to bed zzz.

http://www.digitalsky.org.uk/mars/20...59-49_RGB2.jpg

--
Pete Lawrencehttp://www.digitalsky.org.uk

A nice image and it is really good to see people posting astronomical
material, rather than spam, to the group!

Martin Nicholson, Daventry, England.
http://www.martin-nicholson.info/1/1a.htm
Visit the Astronomical Hall of Shame at http://www.geocities.com/queen5658/

"Pete Lawrence" wrote in message
...
Hi all,

Not a lot of time to process the data set today so I'll put this wip
out first. The seeing was fairly steady last night from my location on
the UK's south coast. Here's the first quick process from the data set
put together last night when I should have gone to bed zzz.

http://www.digitalsky.org.uk/mars/20...59-49_RGB2.jpg


Some nice detail you've caught there Pete, great image!

Pete Lawrence wrote:
Hi all,

Not a lot of time to process the data set today so I'll put this wip
out first. The seeing was fairly steady last night from my location on
the UK's south coast. Here's the first quick process from the data set
put together last night when I should have gone to bed zzz.

http://www.digitalsky.org.uk/mars/20...59-49_RGB2.jpg


Seeing good?? I'll say! A wonderful image.

On Tue, 08 Jan 2008 09:15:28 +0000, Pete Lawrence
wrote:

Thanks for the comments. Here's the final version of this particular
data set.

http://www.digitalsky.org.uk/Mars/20..._Mars+800n.jpg
--
Pete Lawrence
http://www.digitalsky.org.uk

http://www.digitalsky.org.uk/Mars/20..._Mars+800n.jpg

Excellent, Pete.

I have been wanting to ask you this for a while now: Why is
it that when you show the R, G, B components separately,
the resolution seems to get WORSE as wavelength goes from
red to blue?

That seems counter-intuitive.

Is it just that in this case, there IS no sharp detail in the
blue in this image? I don't think that's the case.

Also, I've noticed this same effect when you've posted images
of Saturn, I'm sure.

Another thing I'm confused by: If I were to take your composite
image into, say, Photoshop, and split it into R, G, B components,
I don't think the blue image would look like your "B" image. There
really is no sharp detail in your "B" image, yet I'd expect in Photoshop
to see sharp detail at the edge of the polar ice cap, where there is
a sharp transition between full white and a very dark colour approaching
black. So something else seems to be going on. How do you combine
your R, G, B images to get the composite?

(Couple of mins later I've just effectively done the above. I pulled
the R and G levels to zero, leaving just B active, and the composite image
in the blue channel looks much sharper than your blue component image.


Thanks.


Martin
--
M.A.Poyser Tel.: 07967 110890
Manchester, U.K. http://www.livejournal.com/userinfo.bml?user=fleetie

On Tue, 8 Jan 2008 18:09:44 -0000, "Fleetie"
wrote:

Hi Martin,

http://www.digitalsky.org.uk/Mars/20..._Mars+800n.jpg

Excellent, Pete.

I have been wanting to ask you this for a while now: Why is
it that when you show the R, G, B components separately,
the resolution seems to get WORSE as wavelength goes from
red to blue?

That seems counter-intuitive.

Is it just that in this case, there IS no sharp detail in the
blue in this image? I don't think that's the case.

Also, I've noticed this same effect when you've posted images
of Saturn, I'm sure.

It's because of the atmosphere. Red has an easier time of coming
through it than blue which is refracted (and scattered) considerably.

Another thing I'm confused by: If I were to take your composite
image into, say, Photoshop, and split it into R, G, B components,
I don't think the blue image would look like your "B" image. There
really is no sharp detail in your "B" image, yet I'd expect in Photoshop
to see sharp detail at the edge of the polar ice cap, where there is
a sharp transition between full white and a very dark colour approaching
black. So something else seems to be going on. How do you combine
your R, G, B images to get the composite?

(Couple of mins later I've just effectively done the above. I pulled
the R and G levels to zero, leaving just B active, and the composite image
in the blue channel looks much sharper than your blue component image.

The science in the image presentation is in the base R-G-B images
which are normally left as they come out of Registax. They are
labelled "base" images to indicate that they've not been fiddled with
although in this particular instance I've broken my own rule and
contrast enhanced them via PhotoShop's curve function to accentuate
the albedo features and the white (cloud) patches in the blue channel.

Once I've got an R-G-B set from Registax, I combine them and then
treat the RGB image as a single entity. Each set goes through about 30
stages of tweaking (in my regime anyway) before it's released as a
processed RGB image. I keep each process change as a layer in
PhotoShop and will frequently compare layers to make sure what's
happening doesn't introduce stuff that simply isn't there. It normally
takes several hours to do this properly. The upshot is that the final
processed RGB image bears little resemblance to the original R-G-B
combined image which looks pale and lack-lustre by comparison.

The aim of the colour image is to try and bring out as much hidden
detail in the combined image as possible while still attempting to
keep a sense of reality to the colour scheme. Whether I achieve this
isn't really for me to judge as I'm too close to the image. As far as
I can tell (and I have compared a few images to a base Hubble image as
reference) I don't introduce too many things which aren't there.



Martin

--
Pete Lawrence
http://www.digitalsky.org.uk

"Pete Lawrence" wrote
It's because of the atmosphere. Red has an easier time of coming
through it than blue which is refracted (and scattered) considerably.

Of course; that makes sense.

Actually when I first read what you wrote above, I assumed you
meant Mars' atmosphere; I was (and still am) dubious that Mars' atmosphere
could have that much effect, and then it dawned on me that it
may well be Earth's (much thicker) atmosphere that was blurring
the blue. I am satisfied now that blue scattering in Earth's
atmosphere is causing this severe blurring.

Another thing I'm confused by: If I were to take your composite
image into, say, Photoshop, and split it into R, G, B components,
I don't think the blue image would look like your "B" image. There
really is no sharp detail in your "B" image, yet I'd expect in Photoshop
to see sharp detail at the edge of the polar ice cap, where there is
a sharp transition between full white and a very dark colour approaching
black. So something else seems to be going on. How do you combine
your R, G, B images to get the composite?

(Couple of mins later I've just effectively done the above. I pulled
the R and G levels to zero, leaving just B active, and the composite image
in the blue channel looks much sharper than your blue component image.

The science in the image presentation is in the base R-G-B images
which are normally left as they come out of Registax. They are
labelled "base" images to indicate that they've not been fiddled with
although in this particular instance I've broken my own rule and
contrast enhanced them via PhotoShop's curve function to accentuate
the albedo features and the white (cloud) patches in the blue channel.
[...]

Ok; what I'm still a little confused about is that I would expect your
blue base image to contribute only to the blue component of the composite
image - AND the blue component of the composite image NOT to be affected
by anything in the R or G base images.

Yet this is demonstrably not the case, as revealed by experiment in
Photoshop.

It seems that something about the combination of the images, or (as seems
more likely to me at the moment) your many stages of subsequent processing on
the composite image as a whole, is causing some "crosstalk" between colour
channels, so that the blue component of the final composite image contains
some information from the R and G base images. (Demonstrably it does;
see above.)

This is not a criticism at all; merely an observation and a result of my
seeking to understand what's going on here.

Thank-you very much for taking the time to post such a detailed explanation.

Oh; unrelated, but I'm curious: Presumably you use 3 filters to get your
3 RGB images, yes? What are their transmission spectra like? Are they quite
narrow-band, or rather broad? Their responses will have considerable bearing
on the colour-fidelity (as compared to human eye response) of your composite
image, I suppose.


Martin
--
M.A.Poyser Tel.: 07967 110890
Manchester, U.K. http://www.livejournal.com/userinfo.bml?user=fleetie

On Tue, 8 Jan 2008 19:25:25 -0000, "Fleetie"
wrote:

The science in the image presentation is in the base R-G-B images
which are normally left as they come out of Registax. They are
labelled "base" images to indicate that they've not been fiddled with
although in this particular instance I've broken my own rule and
contrast enhanced them via PhotoShop's curve function to accentuate
the albedo features and the white (cloud) patches in the blue channel.
[...]

It seems that something about the combination of the images, or (as seems
more likely to me at the moment) your many stages of subsequent processing on
the composite image as a whole, is causing some "crosstalk" between colour
channels, so that the blue component of the final composite image contains
some information from the R and G base images. (Demonstrably it does;
see above.)

Yes I think that's exactly right. When the individual RGBs are mixed
and then the colour image is processed as a whole, there is a certain
transfer between channels depending on that processing is applied to
the image. As I said in my last post, the three base R G B images are
the ones that carry the real science as they are closest to the actual
result obtained from processing the original AVI (but there's still
sharpening applied). When they are added together to form the full
colour image, the gloves come off and a variety of tools are used to
try and accentuate the detail in the image. This obviously causes the
cross talk you mention but as the goal is to pull out the detail in
the colour image as a whole, little if any attention is made to what
the processes are doing to the individual colour channels as a whole.

Of course this could add in stuff which wasn't there but part of the
skill of imaging is to make sure, as far as is possible, that this
doesn't happen. I've just processed the next set taken 30 minutes
afterwards and the features shown in the first one all rotate as you
would expect (phew! ).

Oh; unrelated, but I'm curious: Presumably you use 3 filters to get your
3 RGB images, yes? What are their transmission spectra like? Are they quite
narrow-band, or rather broad? Their responses will have considerable bearing
on the colour-fidelity (as compared to human eye response) of your composite
image, I suppose.

Rather than me waffle on further, you can see for yourself...

http://www.astronomik.com/english/eng_rgbt2.html


--
Pete Lawrence
http://www.digitalsky.org.uk

On Jan 8, 7:12 pm, Pete Lawrence wrote:
On Tue, 08 Jan 2008 09:15:28 +0000, Pete Lawrence

wrote:

Thanks for the comments. Here's the final version of this particular
data set.

http://www.digitalsky.org.uk/Mars/20..._Mars+800n.jpg

Exceptional image Pete! This is your best one yet (I know I have said
it before) and you seem to raise the standard yet even further with
each new result.

Good work!

Anthony.

--
Pete Lawrencehttp://www.digitalsky.org.uk