RGB Color Balance with Ex Calibrator Take

RGB Color Balance with Ex. Calibrator – “Take 2” SIG Presentation B. Waddington 5/21/2013

Options for Getting Color Balance • “Prescriptive” tool like Goldman’s RGB calculator • CCDStack background and white point calibration • G 2 V calibration • Ad-hoc – “make it look like Adam Block’s” • Something better…

Ex. Calibrator • Automated tool written by Bob Franke – V 4 is *much* more automated • Used to determine R: G: B color ratios • Based on quantitative measures of star color • Uses your DSO target FOV rather than a separate G 2 V calibration • Freeware

What is Ex. Calibrator Doing? • Downloads professional survey (photometry) data for your field-of-view • Scans your image and identifies nonsaturated stars that are in the survey data • Measures the corresponding R, G, B star brightness values in your image • Computes average R, G, B ratios to best match the survey data

Ex. Calibrator Installation • Just download the zip file from Bob Franke’s site: http: //bfastro. com/excalibrator. htm • Unzip it to a location you want – there is no installer required • Run it…

Using the Program • Select your combined and aligned red, green, and blue images • Select an aligned file that has been platesolved – filter doesn’t matter • Choose your calibration method • Be sure you’ve specified a “server” – Strasbourg, France is fine • Click on ‘Calibrate Image’

Calibration Options • Ex. Calibrator “classic” – Locates the stars in your image using a fairly simple algorithm – Selects only those stars that are “near-white” – Computes the ratios so the white stars in your image look white – Works well when there are many stars to choose from

Calibration Options • SExtractor – Uses a background, professional-grade program for locating the stars in your image – Still selects only “near-white” stars and computes the RGB ratios accordingly – A good option if the ‘classic’ approach is finding few stars to use

Calibration Options • SExtractor/ Linear regression – Uses a background, professional-grade program for locating the stars in your image – Uses a much broader range of star colors in your image – Applies linear regression to compute the optimal RGB ratio across all star colors – Can only be used with SDSS data

Calibration Options • Use multiple options to see a range of results • Use the ‘remove outliers’ command to converge on the final result

What Survey Data Are Used? • Two sources of data for photometric (not spectroscopic) data • Sloan Digital Sky Survey (SDSS) – Highly accurate, all-digital data – Does not cover 100% of the sky – http: //www. sdss 3. org/dr 9/index. php#coverage • NOMAD (US Naval Observatory) – Merged data from 6 catalogs – Highly variable (often poor) photometry – Over 1 billion stars

What Survey Data Are Used? • Photometric data represent magnitude measures through various filters – B, V(green), and R are useful to us • B-V is the primary “color index” • Empirical data show that Nomad values are “too blue” compared to SDSS – so Ex. Calibrator includes a “Nomad adjustment”

Possible Hiccups • You get a nasty error message: function load. Fits. Key. Words failed for Sum File. – You didn’t specify a plate-solved image! – Be sure you saved the file after plate-solving – Look in the file header to be sure the plate solution data are present

Possible Hiccups • You see an error message saying something like: • Check your file names – avoid blanks and most special characters…

Possible Hiccups • Ex. Cal can’t find any stars – Are the color files aligned with each other and with the plate-solved file? – For SDSS data, try the linear regression option – Try the Nomad catalog – if it fails, there’s something wrong with your images

Possible Hiccups • You get a message saying the FOV is not covered by the SDSS database – Well, it isn’t – Use the Nomad catalog – this may require resetting your calibration method to ‘classic’

Possible Hiccups • You are getting results but there are very few stars being used – If you’re using SDSS data, you’re probably ok – You can “gently” widen the ‘Min’ and ‘Max’ color values – Use linear regression option – it uses all the stars

White Balance: What’s Going On • Instrumentation effects – Filter band-pass – Camera sensitivity • Sky – Altitude of DSO – Transparency (moisture, dust, clouds) – Reddening due to interstellar dust • Goals is to set RGB ratios so “neutral is neutral”

Review of Options • CCDStack background/white adjustment – Very quick and easy and will usually get you “close” – Can be a challenge when a white point is not evident – Normal spiral galaxies make a very good white point, but nebulas can be difficult

Review of Options • G 2 V imaging system calibration – Uses very accurate spectral classification of individual stars – G 2 V is ‘white’ – Used to characterize your imaging system, not your sky • Stars are relatively sparse (< 800) so they won’t be in your DSO field of view • RGB ratios must be adjusted for “atmospheric extinction” based on DSO altitude • Doesn’t handle transparency or reddening effects

System Calibration with Ex. Cal (G 2 V’) • Image a star-field near the zenith on a good, clear night • Be sure the FOV is covered by SDSS data • Do a linear regression calibration to compute your “G 2 V-equivalent” ratios

When Would You Use G 2 V’? • To understand color response of your system for planning exposure times • To sidestep having to use Nomad data or to sanity-check Nomad results • Remember that you must correct for atmospheric extinction • http: //www. kellysky. net/White%20 Balancin g%20 RGB%20 Filters. pdf

Review of Options • Ex. Calibrator Strengths – Uses DSO field of view – so it handles both instrumentation and sky effects – For digital SDSS data, results are likely to be very good – Has no impact on data capture – purely a postprocessing step – Relatively easy to use and nothing to lose in trying it – Much easier than G 2 V with equal or better results

Review of Options • Ex. Calibrator Challenges – Can be problematic for small FOVs – Requires a plate-solved image – Nomad data can produce “iffy” results – Ratios produce an “apparent” view, not necessarily an “intrinsic” view (reddening)

The Weird Magnitude System (Useless historical sidebar)