% @(#)responirsp.hlq 17.1.1.1 (ESO-IPG) 01/25/02 17:52:53 % @(#) %++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ %.COPYRIGHT (c) 1992 European Southern Observatory %.IDENT rectifirsp.hlq %.AUTHOR E.Oliva, Arcetri (Florence) %.KEYWORDS MIDAS, help files, RESPONSE/IRSPEC %.PURPOSE On-line help file for the command: RESPONSE/IRSPEC %.VERSION 1.0 24-SEP-1992 : Creation, E.Oliva %---------------------------------------------------------------- \se SECTION./IRSPEC \es\co RESPONSE/IRSPEC 24-SEP-1992 EO \oc\su RESPONSE/IRSPEC in_ima flux_table out_response_ima [yrows=y1,y2,y3,y4] [obs=observation_mode] [norm=normalize_option] [rect=rectify_option] Create a response frame from a standard star 2D spectrum and a flux table. \us\pu Purpose: To a create a (1D) response frame which contains the conversion from counts/sec to flux units for the measurements taken at a given central wavelength (grating position). It requires a "sky-subtracted" and wavelength calibrated 2D frame of a "standard" star plus a flux table - previously created with STANDARD/IRSPEC - containing the intrinsic star fluxes. It automatically takes into account, and correct for, the on chip integration time so that the output response frame will contain the counts/sec given by a given flux unit. \up\sy Syntax: RESPONSE/IRSPEC in_ima flux_table out_response_ima [yrows=y1,y2,y3,y4] [obs=observation_mode] [norm=normalize_option] [rect=rectify_option] \ys\pa in_ima = Input 2D frame containing the observed spectrum(a) of the reference star. It must be cleaned, flat fielded "sky-subtracted" (e.g. the result of SKYSUB/IRSPEC, see also the observing_mode below) and wavelength calibrated (see CALIBRATE/IRSPEC). It must not be normalized unless you use the normalize_option (see below). The command also rectify the image if you have forgotten it, unless you used rect=0 (see below). N.B. the input frame is untouched on exit. \ap\pa flux_table = Input table - created using STANDARD/IRSPEC - containing wavelengths and star fluxes (columns :wl :flux). In case you are only interested in equivalent widths or normalized spectra and you do not want to be bothered with flux calibration you can set this parameter to - 0 - (see last example). Note that in such a case the average value of your response frame will be forced to unity. \ap\pa out_response_ima =\\ Output 1D image containing the instrumental response (counts/sec per unit flux) or just the 1D star spectrum, if you set the flux_table to 0. It will be used in FLUX/IRSPEC.\\ \\ Additional parameters and options \ap\pa yrows = Y-positions (rows) which has to be summed to obtain the 1D stellar spectrum. You need 4 numbers because the program expects to have 2 star spectra - one positive and one negative - in the 2D frame. If you do not specify anything you will be asked to define the four positions with the display cursor. In case your star observations were of the type "object-sky" (see parameter obs=... below) you still need 4 row (cursor) positions, 2 to define the star and 2 to define the sky. \ap\pa observing_mode = \\ "ab" or "os", default=ab. The most usual and recommended observing procedure is to take "object" and "sky" frames with reference to the star set at two different positions along the slit. In this case you will have 2 spectra, one positive and one negative, in your frame. For these kind of frames use obs=ab (default). If you have just one spectrum in your 2D image you must use obs=os (o=o is enough) otherwise the response values will be a factor of 2 too low. \ap\pa normalize_option = \\ 0,1, default=1. The command automatically takes into account, and corrects for, the on-chip integration time so that the output response frame will contain the counts/sec given by a given flux unit (in short it "normalizes" the observed counts). Use norm=0 (or just n=0) to force the comma to skip the automatic, normalization procedure. \ap\pa rectify_option = \\ 0,1, default=1. The command automatically checks if the input 2D image was rectified - using RECTIFY/IRSPEC - and performs this operation in case this was not previously done. Use rect=0 (or just r=0) to skip this control. \ap\no Note: To normalize the frame to counts/sec the command reads the on-chip integration time which is stored in the original frames as a descriptor. If you get strange error messages (e.g. descriptor not present) it probably means that your input image is the result of a series of non standard operations in which the descriptor got lost. Use\\ COPY/DD original_ima *,3 your_image \\ to recover it. A similar argument applies to the "rectify-control" which may not recognize that the image was already rectified if the proper descriptor got lost into non standard operations after the the RECTIFY/IRSPEC command. Use COPY/DD ... *,3 .... in this case too. \\ The Y-positions of the rows that you have defined with the cursor are stored in the output frame in the descriptor IRS_YROWS, to look at them use\\ READ/DESC frame_name IRS_YROWS\\ In case the reference star has obvious, intrinsic features you should correct them in the response file. There is no - or at least I could not find - a standard approach to it. You may fit gaussian to the absorption line but this works easily only for strong and isolated features in regions of good atmospheric transmission, in which case modifying the data manually using MOD/GCURS may give similarly good results. A convenient approach could be that of having two response files from stars which you can assume not to have lines in common (e.g. a very hot and a fairly cool type) in which case by dividing the two frames you should be able to recognize and correct the star features as they appear in absorption or in emission depending to which of the two stars they belong. But this means that you should have already planned your observations in advance and, indeed, "lost" quite a bit of time just looking at the stars... \on\exs Examples: \ex RESP/IRSP bgdiff hr4138 bgresp Create response frame "bgresp" from 2D image "bgdiff" and flux table "hr4138" using the cursor to specify the Y-positions of the star spectra \xe\ex RESP/IRSP bgdiff hr4138 bgresp y=23,28,44,49 Same as above but with the explicit specification of the Y-positions (note that y=... is enough). \xe\ex RESP/IRSP bgdiff hr4138 bgresp y=23,28,44,49 obs=o As above but with the original frame containing just one star spectrum (i.e. with the star out of the slit in the original sky frames). \xe\ex RESP/IRSP pbstd 0 pbrnoflux y=19,25,48,53 Create a "non-flux-calibrated" response frame called "pbrnoflux" which just contains the observed star spectrum (in counts/sec). \xe \sxe