Loading ...
Sorry, an error occurred while loading the content.

Re: Another question about color calibration

Expand Messages
  • Stan
    ... When creating the color images, were all of the color ratios set to 1.0 for both images? Those unweighted images look pretty good, though not exactly
    Message 1 of 13 , Jul 13, 2013
    • 0 Attachment
      --- Mark Striebeck <mark.striebeck@...> wrote:
      > I then stacked them with CCDStack and color combined them.

      When creating the color images, were all of the color ratios set to 1.0 for both images?

      Those "unweighted" images look pretty good, though not exactly alike in terms of the relative ratios of star colors. The nebulas do look realistic and compare well with other color pics - the Bubble is near-monotone red whereas the Crescent has a bluish haze. IMHO those images' color balances are very nearly "correct" and should be tweaked only slightly (if at all).

      > ...weights from XCalibrator applied (without any stretching)

      I have no experience with XCalibrator. It seems to me that if the conditions and sky alt were similar for both images then the same (or very similar) ratios should be applied to both images. To my eye, both of the XCalibrator weighted images look wrong for nebula and star colors (there are some green stars in the Crescent which is a no-no).

      So maybe you are not operating XCalibrator correctly?

      Stan
    • Bob Franke
      Okay folks, here are my eXcalibrator test results with Mark s images. Bubble RGB ratios are... 1.000 1.155 and 1.467 Crescent RGB ratios are... 1.000, 1.305
      Message 2 of 13 , Jul 14, 2013
      • 0 Attachment
        Okay folks, here are my eXcalibrator test results with Mark's images.

        Bubble RGB ratios are...
        1.000 1.155 and 1.467

        Crescent RGB ratios are...
        1.000, 1.305 and 1.533.

        Using the sometimes-inaccurate NOMAD data may account for the difference.

        The below webpage has a complete description of the processing steps and mouse-over images showing the difference between uncorrected and eXcalibrator corrected results.
        http://bf-astro.com/colorTest/markTest.htm

        Regards,
        Bob
        Shameless freeware plug,
        http://bf-astro.com/eXcalibrator/excalibrator.htm


        --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@...> wrote:
        >
        > On Sat, Jul 13, 2013 at 9:27 AM, Stan <stan_ccd@...> wrote:
        >
        > > **
        > >
        > >
        > > --- Mark Striebeck <mark.striebeck@> wrote:
        > > > I then stacked them with CCDStack and color combined them.
        > >
        > > When creating the color images, were all of the color ratios set to 1.0
        > > for both images?
        > >
        >
        > Yes, the non-weighted images had all color ratios set to 1.0.
        >
        > >
        > > Those "weighted" images look pretty good, though not exactly alike in
        > > terms of the relative ratios of star colors. The nebulas do look realistic
        > > and compare well with other color pix - the Bubble is near-monotone red
        > > whereas the Crescent has a bluish haze. IMHO those images' color balances
        > > are very neasrly "correct" and should be tweaked only slightly (if at all).
        > >
        >
        > If found some images of the Bubble nebula where the bubble nebula has a
        > blueish hue - which is what the weighted image looks like.
        >
        > >
        > > > ...weights from XCalibrator applied (without any stretching)
        > >
        > > I have no experience with XCalibrator. It seems to me that if the
        > > conditions and sky alt were similar for both images then the same (or very
        > > similar) ratios should be applied to both images. To my eye, the
        > > XCalibrator (weighted) images look wrong for nebula and star colors (there
        > > are some green stars in the Crescent which is a no-no).
        > >
        > > So maybe you are not operating XCalibrator correctly?
        > >
        >
        > That is absolutely possible. And yes, I expected too that the color ratios
        > for both images would have been somewhat similar.
        >
        > How do others here adjust the colors in their images? I'd love to try out
        > different techniques!
        >
        > MarkS
        >
        > >
        > > Stan
        > >
        > >
        > >
        >
        >
        > [Non-text portions of this message have been removed]
        >
      • Mike Dodd
        ... it looks like eXcalibrator favors more teal. To my eye, the eXcalibrator images are more pleasing, but I think the difference in color could easily boil
        Message 3 of 13 , Jul 14, 2013
        • 0 Attachment
          Bob Franke wrote:
          > Okay folks, here are my eXcalibrator test results with Mark's images.
          >
          > Bubble RGB ratios are...
          > 1.000 1.155 and 1.467
          >
          > Crescent RGB ratios are...
          > 1.000, 1.305 and 1.533.
          >
          > Using the sometimes-inaccurate NOMAD data may account for the difference.
          >
          > The below webpage has a complete description of the processing steps and mouse-over images showing the difference between uncorrected and eXcalibrator corrected results.
          > http://bf-astro.com/colorTest/markTest.htm

          it looks like eXcalibrator favors more teal. To my eye, the eXcalibrator
          images are more pleasing, but I think the difference in color could
          easily boil down to personal taste. Some people might prefer red, some
          might prefer teal. I think they're both nice images.

          --
          Mike

          Mike Dodd
          http://astronomy.mdodd.com
          Louisa County, Virginia USA N37.58.23 W77.56.24
        • Bob Franke
          Yes Mike, eXcalibrator favors the adjustment towards cyan, or teal, with these two images. However, that is not the general case. With the eXcalibrator
          Message 4 of 13 , Jul 14, 2013
          • 0 Attachment
            Yes Mike, eXcalibrator favors the adjustment towards cyan, or teal, with these two images. However, that is not the general case.

            With the "eXcalibrator classic" routine, the program finds stars with color similar to the Sun. The white balance is then calculated with the assumption that we perceive the Sun as white. The NOMAD data use the Johnson UBVRI filter system. This gives a (b-v) magnitude of about 0.65 and a (v-r) of 0.40 for the Sun. As I stated earlier, it is unfortunate that the accuracy of this data is often unreliable.

            With the accurate Sloan Digital Sky Survey, we use the UGRIZ filters. This gives a (u-g) magnitude of 1.43 and a (g-r) of 0.44 for the Sun. The problem with this data is that only about 25% of the sky is covered.

            With the Linear Regression routine, SDSS (g-r) measurements are converted to Johnson (b-v) values. Because of the B and V filters are similar to our red and green we can use any star color in the SDSS data for calculating proper color. With this routine eXcalibrator looks for stars from yellow to white to cyan. These are compared with the color in the user's image to compute the white balance.

            If the rumors are correct, PanSTARRS will release part of their data at the end of this year or early next year. So by then, or just a little later, we should have a 75% sky coverage with SDSS filters and can finally do away with NOMAD. Hopefully this data will be available to load into eXcalibrator.

            The matter of color will always be subjective. The eXcalibrator process provides a means for consistency. I find it nice to at least be consistent with myself.

            Sorry for the long post. I tend to get wordy with Dragon dictation. <g>

            Cheers,
            Bob


            --- In ccd-newastro@yahoogroups.com, Mike Dodd <mike@...> wrote:
            >
            > Bob Franke wrote:
            > > Okay folks, here are my eXcalibrator test results with Mark's images.
            > >
            > > Bubble RGB ratios are...
            > > 1.000 1.155 and 1.467
            > >
            > > Crescent RGB ratios are...
            > > 1.000, 1.305 and 1.533.
            > >
            > > Using the sometimes-inaccurate NOMAD data may account for the difference.
            > >
            > > The below webpage has a complete description of the processing steps and mouse-over images showing the difference between uncorrected and eXcalibrator corrected results.
            > > http://bf-astro.com/colorTest/markTest.htm
            >
            > it looks like eXcalibrator favors more teal. To my eye, the eXcalibrator
            > images are more pleasing, but I think the difference in color could
            > easily boil down to personal taste. Some people might prefer red, some
            > might prefer teal. I think they're both nice images.
            >
            > --
            > Mike
            >
            > Mike Dodd
            > http://astronomy.mdodd.com
            > Louisa County, Virginia USA N37.58.23 W77.56.24
            >
          • Mark Striebeck
            Hi Bob, Thanks for taking a closer look. How did you process the bubble nebula images to get the much more realistic color ratios? MarkS ... [Non-text portions
            Message 5 of 13 , Jul 14, 2013
            • 0 Attachment
              Hi Bob,

              Thanks for taking a closer look. How did you process the bubble nebula
              images to get the much more realistic color ratios?

              MarkS
              On Jul 14, 2013 1:05 AM, "Bob Franke" <bfranke@...> wrote:

              > **
              >
              >
              > Okay folks, here are my eXcalibrator test results with Mark's images.
              >
              > Bubble RGB ratios are...
              > 1.000 1.155 and 1.467
              >
              > Crescent RGB ratios are...
              > 1.000, 1.305 and 1.533.
              >
              > Using the sometimes-inaccurate NOMAD data may account for the difference.
              >
              > The below webpage has a complete description of the processing steps and
              > mouse-over images showing the difference between uncorrected and
              > eXcalibrator corrected results.
              > http://bf-astro.com/colorTest/markTest.htm
              >
              > Regards,
              > Bob
              > Shameless freeware plug,
              > http://bf-astro.com/eXcalibrator/excalibrator.htm
              >
              > --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@...>
              > wrote:
              > >
              > > On Sat, Jul 13, 2013 at 9:27 AM, Stan <stan_ccd@...> wrote:
              > >
              > > > **
              > > >
              > > >
              > > > --- Mark Striebeck <mark.striebeck@> wrote:
              > > > > I then stacked them with CCDStack and color combined them.
              > > >
              > > > When creating the color images, were all of the color ratios set to 1.0
              > > > for both images?
              > > >
              > >
              > > Yes, the non-weighted images had all color ratios set to 1.0.
              > >
              > > >
              > > > Those "weighted" images look pretty good, though not exactly alike in
              > > > terms of the relative ratios of star colors. The nebulas do look
              > realistic
              > > > and compare well with other color pix - the Bubble is near-monotone red
              > > > whereas the Crescent has a bluish haze. IMHO those images' color
              > balances
              > > > are very neasrly "correct" and should be tweaked only slightly (if at
              > all).
              > > >
              > >
              > > If found some images of the Bubble nebula where the bubble nebula has a
              > > blueish hue - which is what the weighted image looks like.
              > >
              > > >
              > > > > ...weights from XCalibrator applied (without any stretching)
              > > >
              > > > I have no experience with XCalibrator. It seems to me that if the
              > > > conditions and sky alt were similar for both images then the same (or
              > very
              > > > similar) ratios should be applied to both images. To my eye, the
              > > > XCalibrator (weighted) images look wrong for nebula and star colors
              > (there
              > > > are some green stars in the Crescent which is a no-no).
              > > >
              > > > So maybe you are not operating XCalibrator correctly?
              > > >
              > >
              > > That is absolutely possible. And yes, I expected too that the color
              > ratios
              > > for both images would have been somewhat similar.
              > >
              > > How do others here adjust the colors in their images? I'd love to try out
              > > different techniques!
              > >
              > > MarkS
              > >
              > > >
              > > > Stan
              > > >
              > > >
              > > >
              > >
              > >
              > > [Non-text portions of this message have been removed]
              > >
              >
              >
              >


              [Non-text portions of this message have been removed]
            • Mike Dodd
              ... I tried Dragon early-on in my recovery from a 2007 stroke, but it made too many mistakes, so I spent a lot of time proofing and correcting. I finally
              Message 6 of 13 , Jul 14, 2013
              • 0 Attachment
                Bob Franke wrote:
                > Sorry for the long post. I tend to get wordy with Dragon dictation.
                > <g>

                I tried Dragon early-on in my recovery from a 2007 stroke, but it made
                too many mistakes, so I spent a lot of time proofing and correcting. I
                finally decided to go back to keyboarding, first with one hand, and
                eventually with both. (My left hand is now about 90% for typing -- still
                not easy, but at least functional.)

                I'm pleased to hear from another Dragon user. It's possible that the
                errors I experienced were due to my slightly-impaired speech at the
                time. I should try it again now to see if it works better.
                --
                Mike

                Mike Dodd
                http://astronomy.mdodd.com
                Louisa County, Virginia USA N37.58.23 W77.56.24
              • Bob Franke
                Hello Mark, I loaded your color and the WCS files into CCDStack. The three color files were registered with each other but not to the WCS. So the first process
                Message 7 of 13 , Jul 14, 2013
                • 0 Attachment
                  Hello Mark,

                  I loaded your color and the WCS files into CCDStack. The three color files were registered with each other but not to the WCS. So the first process was to register the color files to the WCS file. Then I saved the four files as 32-bit floating-point.

                  I then loaded the four files into eXcalibrator, set then Min/Max magnitudes to 15/19 and clicked the NOMAD radio button. I selected "eXcalibrator classic" and clicked the Calibrate button. Two clicks of the Remove Outliers button reduced the star count to 9 and the standard deviation to 0.155. This gave a similar answer to the SExtractor routine with just one click of Remove Outliers. I decided to go with the "eXcalibrator classic" RGB ratios of 1.000, 1.155 and 1.467.

                  Then I simply used CCDStack to create the RGB image with the above ratios. For setting the background, I selected an area in the lower right-hand corner and checked the Desaturate Background box. Final image stretching was done with PhotoShop.

                  If this were my data, I probably would not use the NOMAD stars. I would simply use my standard RGB image train calibration of 1.00, 0.95 and 1.05. This calibration was determined with eXcalibrator and stars from the Sloan Digital Sky Survey. I only consider a NOMAD calibration when I have many calibration stars.

                  Hope this helps,
                  Bob


                  --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@...> wrote:
                  >
                  > Hi Bob,
                  >
                  > Thanks for taking a closer look. How did you process the bubble nebula
                  > images to get the much more realistic color ratios?
                  >
                  > MarkS
                  > On Jul 14, 2013 1:05 AM, "Bob Franke" <bfranke@...> wrote:
                  >
                  > > **
                  > >
                  > >
                  > > Okay folks, here are my eXcalibrator test results with Mark's images.
                  > >
                  > > Bubble RGB ratios are...
                  > > 1.000 1.155 and 1.467
                  > >
                  > > Crescent RGB ratios are...
                  > > 1.000, 1.305 and 1.533.
                  > >
                  > > Using the sometimes-inaccurate NOMAD data may account for the difference.
                  > >
                  > > The below webpage has a complete description of the processing steps and
                  > > mouse-over images showing the difference between uncorrected and
                  > > eXcalibrator corrected results.
                  > > http://bf-astro.com/colorTest/markTest.htm
                  > >
                  > > Regards,
                  > > Bob
                  > > Shameless freeware plug,
                  > > http://bf-astro.com/eXcalibrator/excalibrator.htm
                  > >
                  > > --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@>
                  > > wrote:
                  > > >
                  > > > On Sat, Jul 13, 2013 at 9:27 AM, Stan <stan_ccd@> wrote:
                  > > >
                  > > > > **
                  > > > >
                  > > > >
                  > > > > --- Mark Striebeck <mark.striebeck@> wrote:
                  > > > > > I then stacked them with CCDStack and color combined them.
                  > > > >
                  > > > > When creating the color images, were all of the color ratios set to 1.0
                  > > > > for both images?
                  > > > >
                  > > >
                  > > > Yes, the non-weighted images had all color ratios set to 1.0.
                  > > >
                  > > > >
                  > > > > Those "weighted" images look pretty good, though not exactly alike in
                  > > > > terms of the relative ratios of star colors. The nebulas do look
                  > > realistic
                  > > > > and compare well with other color pix - the Bubble is near-monotone red
                  > > > > whereas the Crescent has a bluish haze. IMHO those images' color
                  > > balances
                  > > > > are very neasrly "correct" and should be tweaked only slightly (if at
                  > > all).
                  > > > >
                  > > >
                  > > > If found some images of the Bubble nebula where the bubble nebula has a
                  > > > blueish hue - which is what the weighted image looks like.
                  > > >
                  > > > >
                  > > > > > ...weights from XCalibrator applied (without any stretching)
                  > > > >
                  > > > > I have no experience with XCalibrator. It seems to me that if the
                  > > > > conditions and sky alt were similar for both images then the same (or
                  > > very
                  > > > > similar) ratios should be applied to both images. To my eye, the
                  > > > > XCalibrator (weighted) images look wrong for nebula and star colors
                  > > (there
                  > > > > are some green stars in the Crescent which is a no-no).
                  > > > >
                  > > > > So maybe you are not operating XCalibrator correctly?
                  > > > >
                  > > >
                  > > > That is absolutely possible. And yes, I expected too that the color
                  > > ratios
                  > > > for both images would have been somewhat similar.
                  > > >
                  > > > How do others here adjust the colors in their images? I'd love to try out
                  > > > different techniques!
                  > > >
                  > > > MarkS
                  > > >
                  > > > >
                  > > > > Stan
                  > > > >
                  > > > >
                  > > > >
                  > > >
                  > > >
                  > > > [Non-text portions of this message have been removed]
                  > > >
                  > >
                  > >
                  > >
                  >
                  >
                  > [Non-text portions of this message have been removed]
                  >
                • Mark Striebeck
                  ... Ah - I forgot that (to register the wcs file to the color files)! Thanks!!! ... [Non-text portions of this message have been removed]
                  Message 8 of 13 , Jul 15, 2013
                  • 0 Attachment
                    On Sun, Jul 14, 2013 at 8:55 PM, Bob Franke <bfranke@...> wrote:

                    > **
                    >
                    >
                    > Hello Mark,
                    >
                    > I loaded your color and the WCS files into CCDStack. The three color files
                    > were registered with each other but not to the WCS. So the first process
                    > was to register the color files to the WCS file. Then I saved the four
                    > files as 32-bit floating-point.
                    >
                    Ah - I forgot that (to register the wcs file to the color files)! Thanks!!!

                    >
                    > I then loaded the four files into eXcalibrator, set then Min/Max
                    > magnitudes to 15/19 and clicked the NOMAD radio button. I selected
                    > "eXcalibrator classic" and clicked the Calibrate button. Two clicks of the
                    > Remove Outliers button reduced the star count to 9 and the standard
                    > deviation to 0.155. This gave a similar answer to the SExtractor routine
                    > with just one click of Remove Outliers. I decided to go with the
                    > "eXcalibrator classic" RGB ratios of 1.000, 1.155 and 1.467.
                    >
                    > Then I simply used CCDStack to create the RGB image with the above ratios.
                    > For setting the background, I selected an area in the lower right-hand
                    > corner and checked the Desaturate Background box. Final image stretching
                    > was done with PhotoShop.
                    >
                    > If this were my data, I probably would not use the NOMAD stars. I would
                    > simply use my standard RGB image train calibration of 1.00, 0.95 and 1.05.
                    > This calibration was determined with eXcalibrator and stars from the Sloan
                    > Digital Sky Survey. I only consider a NOMAD calibration when I have many
                    > calibration stars.
                    >
                    > Hope this helps,
                    > Bob
                    >
                    >
                    > --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@...>
                    > wrote:
                    > >
                    > > Hi Bob,
                    > >
                    > > Thanks for taking a closer look. How did you process the bubble nebula
                    > > images to get the much more realistic color ratios?
                    > >
                    > > MarkS
                    > > On Jul 14, 2013 1:05 AM, "Bob Franke" <bfranke@...> wrote:
                    > >
                    > > > **
                    >
                    > > >
                    > > >
                    > > > Okay folks, here are my eXcalibrator test results with Mark's images.
                    > > >
                    > > > Bubble RGB ratios are...
                    > > > 1.000 1.155 and 1.467
                    > > >
                    > > > Crescent RGB ratios are...
                    > > > 1.000, 1.305 and 1.533.
                    > > >
                    > > > Using the sometimes-inaccurate NOMAD data may account for the
                    > difference.
                    > > >
                    > > > The below webpage has a complete description of the processing steps
                    > and
                    > > > mouse-over images showing the difference between uncorrected and
                    > > > eXcalibrator corrected results.
                    > > > http://bf-astro.com/colorTest/markTest.htm
                    > > >
                    > > > Regards,
                    > > > Bob
                    > > > Shameless freeware plug,
                    > > > http://bf-astro.com/eXcalibrator/excalibrator.htm
                    > > >
                    > > > --- In ccd-newastro@yahoogroups.com, Mark Striebeck <mark.striebeck@>
                    > > > wrote:
                    >
                    > > > >
                    > > > > On Sat, Jul 13, 2013 at 9:27 AM, Stan <stan_ccd@> wrote:
                    > > > >
                    > > > > > **
                    > > > > >
                    > > > > >
                    > > > > > --- Mark Striebeck <mark.striebeck@> wrote:
                    > > > > > > I then stacked them with CCDStack and color combined them.
                    > > > > >
                    > > > > > When creating the color images, were all of the color ratios set
                    > to 1.0
                    > > > > > for both images?
                    > > > > >
                    > > > >
                    > > > > Yes, the non-weighted images had all color ratios set to 1.0.
                    > > > >
                    > > > > >
                    > > > > > Those "weighted" images look pretty good, though not exactly alike
                    > in
                    > > > > > terms of the relative ratios of star colors. The nebulas do look
                    > > > realistic
                    > > > > > and compare well with other color pix - the Bubble is
                    > near-monotone red
                    > > > > > whereas the Crescent has a bluish haze. IMHO those images' color
                    > > > balances
                    > > > > > are very neasrly "correct" and should be tweaked only slightly (if
                    > at
                    > > > all).
                    > > > > >
                    > > > >
                    > > > > If found some images of the Bubble nebula where the bubble nebula
                    > has a
                    > > > > blueish hue - which is what the weighted image looks like.
                    > > > >
                    > > > > >
                    > > > > > > ...weights from XCalibrator applied (without any stretching)
                    > > > > >
                    > > > > > I have no experience with XCalibrator. It seems to me that if the
                    > > > > > conditions and sky alt were similar for both images then the same
                    > (or
                    > > > very
                    > > > > > similar) ratios should be applied to both images. To my eye, the
                    > > > > > XCalibrator (weighted) images look wrong for nebula and star colors
                    > > > (there
                    > > > > > are some green stars in the Crescent which is a no-no).
                    > > > > >
                    > > > > > So maybe you are not operating XCalibrator correctly?
                    > > > > >
                    > > > >
                    > > > > That is absolutely possible. And yes, I expected too that the color
                    > > > ratios
                    > > > > for both images would have been somewhat similar.
                    > > > >
                    > > > > How do others here adjust the colors in their images? I'd love to
                    > try out
                    > > > > different techniques!
                    > > > >
                    > > > > MarkS
                    > > > >
                    > > > > >
                    > > > > > Stan
                    > > > > >
                    > > > > >
                    > > > > >
                    > > > >
                    > > > >
                    > > > > [Non-text portions of this message have been removed]
                    > > > >
                    > > >
                    > > >
                    > > >
                    > >
                    > >
                    > > [Non-text portions of this message have been removed]
                    > >
                    >
                    >
                    >


                    [Non-text portions of this message have been removed]
                  Your message has been successfully submitted and would be delivered to recipients shortly.