iraf.net

Hi,I've used dopcor to shift a spectrum by a velocity in km/s. Then I've been trying to measure that shift by using fxcor with the newly shifted spectrum as object and the original non-shifted spectrum as template. My assumption is that the vrel (or vhel) that fxcor computes will correspond with the velocity that I fed into dopcor. However, fxcor seems to return a systematically lower shift (dopcor shift was 8 km/s, and fxcor measures -0.8 km/s).I know dopcor works correctly because I measured pre and post shift lines in splot. I therefore don't understand why fxcor is not detecting the shift applied by dopcor. Thank you for your help.
Steph

Steph,Did you set the 'isvelocity' parameter to DOPCOR when you ran it? Using an artificial spectrum I was able to DOPCOR it with a shift of 8km/s and recover that value, but note that if it thought the shift was z=8 then the value is something like -0.88, curiously close to your report. The tip-off is whether the status bar says something like Vrel or Zrel when printing the value.Otherwise, please post the command you used when running DOPCOR.Cheers,
-Mike

I reran the dopcor shift and then fxcor-check this morning, and as you can see, the vrel that fxcor calculates is not equivalent to the velocity that I redshifted by. By the way, I have been setting isvelocity=yes, and as I mentioned I know that dopcor works correctly, as I could measure the shifts in splot. Thank you once again for your help!Here are the commands I used for dopcor:BD+332642_Jul20_comb_s.fits: Doppler correction - apertures=all, redshift=2.669E-5, flux factor=0.
--> lpar dopcor
input = BD+332642_Jul20_comb.fits List of input spectra
output = BD+332642_Jul20_comb_s.fits List of output spectra
redshift = 8 Redshift or velocity (Km/s)
(isvelocity = yes) Is the redshift parameter a velocity?
(add = no) Add to previous dispersion correction?
(dispersion = yes) Apply dispersion correction?
(flux = no) Apply flux correction?
(factor = 3.0) Flux correction factor (power of 1+z)
(apertures = ) List of apertures to correct
(verbose = yes) Print corrections performed?
(mode = al)Here are the parameters and output for these two files in fxcor:--> lpar fxcor
objects = BD+332642_Jul20_comb_s.fits List of object spectra
templates = BD+332642_Jul20_comb.fits List of template spectra
(apertures = *) Apertures to be used
(cursor = ) Graphics input cursor
(continuum = none) Continuum subtract spectra?
(filter = both) Fourier filter the spectra?
(rebin = smallest) Rebin to which dispersion?
(pixcorr = no) Do a pixel-only correlation?
(osample = 3965.00-3973.00,4090.00-4125.00,4331.00-4352.00,4855.00-4865.00) Object regions to be correlated ('*' => all)
(rsample = 3965.00-3973.00,4090.00-4125.00,4331.00-4352.00,4855.00-4865.00) Template regions to be correlated
(apodize = 0.2) Apodize end percentage
(function = parabola) Function to fit correlation
(width = INDEF) Width of fitting region in pixels
(height = 0.0) Starting height of fit
(peak = no) Is height relative to ccf peak?
(minwidth = 3.0) Minimum width for fit
(maxwidth = 400.0) Maximum width for fit
(weights = 1.0) Power defining fitting weights
(background = 0.0) Background level for fit
(window = 2000.0) Size of window in the correlation plot
(wincenter = INDEF) Center of peak search window
(output = fxcor.swBD33orig.ex2) Root spool filename for output
(verbose = long) Verbose output to spool file?
(imupdate = no) Update the image header?
(graphics = stdgraph) Graphics output device
(interactive = yes) Interactive graphics?
(autowrite = yes) Automatically record results?
(autodraw = yes) Automatically redraw fit results?
(ccftype = image) Output type of ccf
(observatory = kpno) Observation location database
(continpars = ) Continuum processing parameters
(filtpars = ) Filter parameters pset
(keywpars = ) Header keyword translation pset
(mode = ql)--> lpar filtpars
(f_type = square) Filter window type
(cuton = 12) Cuton component for filter
(cutoff = 700) Cutoff component for filter
(fullon = 9) Component at which (ramp) filter reaches full value
(fulloff = 700) Component at which (ramp) filter reaches zero
(mode = ql)the status bar says
HJD=3937.8129 FWHM=499.70 Vr=0.700 Vo=18.562 Vh=0.700 +/- 0.102The output (.txt):#K IRAF = NOAO/IRAF V2.13-BETA version
#K USER = swortel name
#K HOST = jollyroger.internal.amnh.org computer
#K DATE = 2007-03-23 yyyy-mm-dd
#K TIME = 11:02:58 hh:mm:ss
#K PACKAGE = rv name
#K TASK = fxcor name
#
#T Template ' A' -- Object = 'BD_+33_2642__2006_Jul_20-21' \*
# Image = 'BD+332642_Jul20_comb.fits' Vhelio = INDEF
#
# Velocity Dispersion = 3.07 Km/sec/pixel Rebinned WPC = 4.44514E-6
#
#N OBJECT IMAGE REF HJD AP CODES SHIFT HGHT FWHM TDR VOBS VREL VHELIO VERR
#U name image days cfr/fun pixel km/s km/s km/s km/s
#
BD_+33_2642__20 BD+332642_Jul20_comb_s.fits A 3937.8129 1 NBS/par 0.228 0.81 499.70 3834.1 18.5620 0.7000 0.7001 0.102

Hi Steph,Would you mind mailing me the original image (fitz@iraf.net) so I can try to reproduce it, I'm not seeing anything unusual in your output. Does using the entire spectrum instead of regions, or not filtering, make any difference in the output? The computed shift corresponds to the velocity as I'd expect, I'm just wondering whether this extra processing is affecting the peak at all.-Mike

Hi Mike,I sent you the spectra I've been shifting/checking this morning. I also had another question:Is there a way to "zoom in" or change the windowing when looking at the original object/filtered object in the fourier mode of fxcor (keystrokes f, then o)?Thank you for your continued help!Steph

Hi Steph,Thanks for the data file. I'm still trying to figure out exactly what's happening, but using the default FXCOR parameters does resolve to a Vhelio of the 8 km/s I shifted with DOPCOR. For some reason the filtering and use of sample regions are "degrading" the result.As for your second question, the ":zoom" command in FFT mode can be used to specify a magnification factor for the fft plot display (e.g. ":zoom 2" will show only the first half of the power spectrum, etc).Cheers,
-Mike