1 COMMENT ANALYSE\COMMENT [TYPE|READ|WRITE|APPEND|EDIT|DELETE] ["comment"] This command READs, TYPEs or DELETEs an existing comment section (max. length: 1020 characters). The options WRITE, APPEND and EDIT will create a comment section if none is present; the options WRITE or APPEND expect as argument the comment line (eventually inclosed in double quotes (") when the comment is more than one word long). 1 CUBE ANALYSE\CUBE Filename [MATH_EXPRESSION] [/CHANNEL Cmin Cmax] [/STEP Stepx Stepy] This command builds either a data-cube or a 2-dimensional map at the GILDAS format from the current index of scans. It is up to the user to ensure that the current index concerns a single source and a single line, observed with a unique spectral resolution, with all spectra centered at the same velocity. If 'MATH_EXPRESSION' is present in the command line, ANALYSE\CUBE will make a 2-D map of (the result of Math_expression applied for each spectrum), for each spectra position. This is particulary handy when one use the SIC variables and functions (see HELP CUBE Functions, HELP LAS\SET VARIABLE and try the command SIC\EXAMINE /FUNCTION) in the Math_expression. In that case, The data cube axis types will be RA or L DEC or B, depending on the type of offsets In the other case (i.e., a real data cube of channels values), The data cube axis types will be Velocity or Channels or Rest or Image frequencies, depending on the current X unit. RA or L DEC or B, depending on the type of offsets Option /CHANNEL can be used to limit the part of the spectra that will be included in the cube (no use if you plot a Math_expression). Option /STEP can always be used to force the step size of the cube. The resulting 2 or 3D data cube can then be processed as a whole using the GILDAS (Grenoble Image and Line Data Analysis System) software. GILDAS includes tasks to transpose, reinterpolate, change the projection systems, make overlays with the IRAS point sources, contours and so on from data cubes. Call GRAPHIC at system level and then type EXPLAIN for more information on GILDAS tasks. 2 FUNCTIONS The Following Functions have been defined: STDV(dummy) returns the integrated area (K.Km/s) of the current spectrum TPEAK(dummy) returns the Peak temperature (K) of the current spectrum VPEAK(dummy) returns the Velocity at Peak (Km/s) of the current spectrum TDV(v1,v2) returns the integrated area (K.Km/s) of the current spectrum between velocities V1 and V2 TPEAKV(v1,v2) returns the Peak temperature (K) of the current spectrum between velocities V1 and V2 VPEAKV(v1,v2) returns the Velocity at Peak (Km/s) of the current spectrum between velocities V1 and V2 To create a map of The integrated area of your spectra, type: CUBE filename STDV(0) 1 DISPLAY ANALYSE\DISPLAY Prints the results of the profile fit of R for the current method, if any result of course. For the Gauss (default) Method (see HELP METHOD), the command displays (1) the Line number, (2) the Area, (3) the Position, (4) the Width and (5) the Intensity , each with their associated errors, and resulting RMS in the Base and Line windows. 1 DIVIDE ANALYSE\DIVIDE Threshold This command divides the spectrum in R by the spectrum in T, which is left unchanged. The two spectra must be on the same velocity scale (Use command RESAMPLE to do so). Threshold is a value to avoid zero divide. The output spectrum is blanked for channels where abs(T) < Threshold. 1 DRAW ANALYSE\DRAW [Action [Parameters]] This command executes the specified drawing action, or calls the graphic cursor and uses the returned character from the cursor to execute the corresponding action. The mnemonic for the cursor characters are : .nf F for "FILL" To interpolate a bad channel K for "KILL" To flag a bad channel L for "LOWER" To flag line identifications, ... M for "MASK" To show active masks T for "TEXT" To annotate your spectrum U for "UPPER" Similar to LOWER W for "WINDOW" To show current line windows .fi any other to display the channel number, frequency, velocity and antenna temperature at the cursor position. The explicit form does not call the cursor and can be used in procedures. 2 FILL ANALYSE\DRAW FILL Nchan Interpolates the value of the specified bad channel using the nearest valid channels. Channel Nchan must be blanked before. In cursor mode, the channel number is retrieved from the cursor position. 2 KILL ANALYSE\DRAW KILL Nchan Deletes the specified channel, i.e. attributes the blanking value to it. In cursor mode, the channel number is retrieved from the cursor position. 2 LOWER ANALYSE\DRAW LOWER Value "text" Draws a vertical line followed by the text "text" (written vertically also) at abscissa Value on lower axis. The line length is adjusted to approach closely the spectrum value at the corresponding abscissa. 2 MASK ANALYSE\DRAW MASK [level] Plots the active masks, with bars at the mask positions and Y=Level (Default 0). 2 UPPER ANALYSE\DRAW UPPER Value "text" Draws a vertical line followed by the text "text" (written vertically also) at abscissa Value on the upper axis. The line length is adjusted to approach closely the spectrum value at the corresponding abscissa. 2 TEXT ANALYSE\DRAW TEXT X Y "text" I Writes (horizontally) the string "text" at plot position (X,Y) in paper units, with the centering code I for the string. 2 WINDOW ANALYSE\DRAW WINDOW [level] Plots the current line windows, with bars at the mask positions and Y=Level (Default 0). 1 FFT ANALYSE\FFT [/REMOVE] FFT computes the Fourier Transform of R and plots it. R and T observations are unaffected. There is no control over the plot units, and no associated coordinate system (i.e. you cannot measure periods with the cursor). You can edit the Fourier Transform using the cursor. Strike key K to mark the beginning of an area to delete, move the cursor to the end of the area to delete and strike key K again. Then strike E to exit, and Fourier components are interpolated in module and phase from the boundaries of deleted areas. Non interactive editing is also possible and can be useful for repetitive applications. The syntax is then FFT Start End [/REMOVE] 2 /REMOVE ANALYSE\FFT /REMOVE Remove existing fits of the current method before computing the Fourier transform. If the Fourier Transform is edited, fits are added back upon exit. 1 FILL LAS\FILL start end Linearly interpolates the spectrum between start and end values (current units, not necessary channels). Can be used to remove mesospheric lines, several adjacent bad channels... 1 FIT ANALYSE\FIT [N] Plot the current fitted profile. The Nth line obtained the last time GAUSS was executed, or recovered by DISPLAY, is plotted on the screen. If N is 0, the sum of all lines is plotted, and if N is not given, the last value of N is used. 1 GAUSS ANALYSE\GAUSS GAUSS performs a fit of a theoretical profile in the R spectrum using initial values specified by the command LINES. The profile is determined by the METHOD command (GAUSS, SHELL or NH3 lines are currently implemented, and a special method, CONTINUUM, may be used for continuum data). If LINES is 0, an attempt to guess initial values for a single line is done by computing the moments of the spectrum. A first minimisation is performed using the Simplex technique; the results are improved and the uncertainties computed using the Gradient technique. GAUSS uses the masks defined by SET MASK. If GAUSS did not converge, the errors quoted in the results are meaningless, and you should use command ITERATE to try to do better. For CONTINUUM method, although a linear baseline is fitted with the gaussian, only the gaussian parameters are displayed. If the observation was made using beam-switching method and the reference beam appears in the drifts, a two-component gauss fit is done using separation specified in the data. 1 GREG ANALYSE\GREG Name [/FORMATTED] [/READ [X n] [Y m] Writes in file Name the contents of current scan as a GILDAS table, or as a formatted file if the /FORMATTED option is present. If the /READ option is present, performs the reverse operation: load the file content in the current scan. Only available for formatted files; the equivalent capability for GILDAS tables is available through a combination of DEFINE IMAGE and MODEL. For the table, the information stored is : 1. Intensity 2. Channel number 3. Velocity 4. Offset frequency 5. Rest frequency 6. Image frequency 7. Fitted profiles if any - fit(i),i=0,nline - in column 7+i The output table can be put later in a formatted way using GILDAS task LIST in you need. The table may be used as input to GreG to produce fancy plots, or by the GILDAS software for other applications. In particular, SIC variables can be used to subtract any of the fits from the spectrum to produce residuals if needed. For plots, note that you can produce essentially the same within CLASS, except under very special circumstances, and moreover that only CLASS can handle the different X-unit simultaneously. For the formatted list, less information is written (and it takes longer to write it of course) : 1. Current X unit in column 1 2. Intensity in column 2 3. Fitted profiles - (fit(i),i=0,nline) - in column 3+i For the /READ option, columns are X 1 Y 2 by default but can be changed. If X 0 is given as zero, then only Y is read and X is filled with channel numbers. Only very coarse header information is filled in. A lot of complementary header editing will often be needed. 1 GRID ANALYSE\GRID [Filename] [NEW|APPEND] [MATH_EXPRESSION] [/CHANNEL [C|V|F|I] min max] [/IMAGE [BEAM beam] [LAM nx rx vx ix] [BETA ny ry vy iy] [LIKE tempfile]] [/PLOT [where]] [/SMOOTH fx fy] This command builds a GILDAS table containing the following values Column 1 X offset (in radians) Column 2 Y Offset (in radians) Column 3 Weight of the data and Columns 4 to... : values of the specified Spectral channel(s) or Column 4: Result of the mathematical expression from the current index of scans. IT IS UP TO THE USER to ENSURE that the current index concerns a SINGLE SOURCE and a SINGLE LINE, observed with a UNIQUE SPECTRAL RESOLUTION, with ALL spectra centered at the SAME VELOCITY. Option /IMAGE is used to grid the data into a (2-d or) 3-d cube, option /PLOT to plot a map of channel-averaged data. If "Filename" is not present the data is appended to the current table (the one opened by the last GRID command). If "APPEND" is given as the second argument the table Filename is opened and data is appended to it; otherwise "NEW" may be specified to open a new table; this is the default. If "MATH_EXPRESSION" is present in the command line, Column 4 will contain the result of Math_expression applied for each spectrum. This is particulary handy when one use the SIC variables and functions (see HELP CUBE Functions, HELP LAS\SET VARIABLE and try the command SIC\EXAMINE /FUNCTION) in the Math_expression. Option /CHANNEL can be used to limit the part of the spectra that will be included in the table (not used if you treat a Math_expression). The resulting table can then be processed as a whole using the GILDAS (Grenoble Image and Line Data Analysis System) software. With /IMAGE a data cube is created from this table. GILDAS includes tasks to do the same job with more flexibility (GRID_SG and GRID_EXTEND), and transpose, reinterpolate, change the projection systems, make overlays with the IRAS point sources, contours and so on from data cubes. Call GRAPHIC at system level and then type EXPLAIN for more information on GILDAS tasks. 2 /CHANNEL GRID /CHANNEL [CHANNEL|VELO|FREQ|IMAGE] min max Select the channels to be used for the tabulation and gridding process. This fixes the third dimension of the final data cube. Use CHANNEL (the default) to give min and max in channels numbers, or one of the other units: Velocity, Frequency, ir Image frequency. 2 /IMAGE GRID /IMAGE [BEAM beam] [LAM nx rx vx ix] [BETA ny ry vy iy] [LIKE tempfile] Grid the tabulated data in a data cube (or a 2-d image if only one channel is processed, or a mathematical expression). The cube coordinate axes are normally defaulted from the position of the spectra, assuming that the angular resolution was 'beam' (in current angle units; a recommended default is taken for 30m telescope data); otherwise the LAMBDA and/or BETA axes may be forced by giving the number of pixels (nx,ny), the reference pixel (rx,ry), the coordinate value at the reference pixel (vx,vy) , and the pixel size (ix,iy, all in current angle units). Finally both coordinate axes may be taken from a template image (LIKE tempfile). In the APPEND mode, none of the /IMAGE arguments should be given, the current image (from the previous GRID/IMAGE command) is incrementally recomputed, adding the new data. This is intended for a real time display while observing. 2 /PLOT GRID /PLOT [WHERE] To be used only with /IMAGE. An image of the average of all the channels is plotted. WHERE plots GreG markers at the observed positions. 2 /SMOOTH GRID /SMOOTH fx [fy] To be used only with /IMAGE. Smoothes the image by 'fx' and 'fy'. 1 ITERATE ANALYSE\ITERATE ITERATE tries the GRADIENT technique to improve the last results of the profile fit. Do not expect spectacular results if the absolute minimum is far away ! Typically, this command should be used if the previous fit did not converge, but seems close enough. If you are using the GAUSS method, be cautious not to change the X unit in the mean time. 1 KEEP ANALYSE\KEEP Save profile fit results in the output file. KEEP is equivalent to UPDATE, but only profile fit results are saved; KEEP is subject to the same restrictions as UPDATE. 1 Language ANALYSE\ Language Summary COMMENT : Manipulates the COMMENT section of spectra CUBE file [fnct]: Creates a 2 or 3-D data cube for processing by GILDAS. DISPLAY : Prints the results of the profile fit of R. DIVIDE : Divides R spectrum by T spectrum. DRAW : Calls the cursor or plot comments. FILL start end : Linearly interpolates the spectrum between start and end. FFT : Computes and edit Fourier Transform of R. FIT [N] : Plots the current fitted profile. GAUSS : Performs a profile fit. GREG [name] : Makes a TABLE or formatted file from current scan GRID file [fnct]: Creates a data table or image for processing by GILDAS. ITERATE : Iterates the profile fit. KEEP : Saves Gaussian fit results in the output file. LINES [N] : Enters the initial values for the profile fit. METHOD Arg : Selects the line profile for fits. MAP [M|W] : Plots a map of spectra (or only their location). MEMORIZE Arg : Memorize the current observation. MODEL Var : Generates a CLASS spectrum from a 1D variable. NOISE [S [NEW]] : Generates gaussian noise PRINT [Arg] : Prints values in a formatted file, or a table. POPUP : Zoom a spectrum from a STAMP or MAP plot. RESIDUAL [n] : Computes the residuals of the profile fit. RESAMPLE Arg : Resamples a spectrum on a specified grid REDUCE : Reduces a SKYDIP RETRIEVE Arg : Retrieve an observation from the memories. SMOOTH [Arg...] : Smoothes the spectrum in R STRIP File : Creates an image for Position-Velocity plots. STAMP : Displays all spectra in index on one plot. 2 NEWS Version 4.1 - Commands MODEL to generate spectra from SIC variables. 06-Nov-1989 - Commands MEMORIZE and RETRIEVE to temporarily save spectra. 8-Apr- 1988 Version 4.0 - Support of method HFS in fits. 1-Feb-1988 - Support of table format in command PRINT. 10-Jan-1988 - Command REDUCE for skydips. 1-Dec-1987 - Support for continuum in GAUSS, ITERATE, FIT, DISPLAY, PRINT and GREG. 1-Oct-1987 - STRIP command now uses the GILDAS format. 25-Aug-1986 - Commands RESAMPLE and DIVIDE added. - TABLE format in command GREG. 1-Jul-1986 - CUBE command for general processing (contour maps, velocity-position plots and so on) via the GILDAS software. 20-Jun-1986 1 LINES ANALYSE\LINES [N] [/NOCURSOR] [/INPUT File_Name] Enter the initial values for the profile fit. If argument "AUTO" is given instead of a number, then the number of lines found in the last read spectrum will be used for ITERATE command. If N is not specified, then the last value for N is used. If N is 0, GAUSS will attempt to guess initial values for a single line by computing the moments of the spectrum. If N is not 0, LINES prompts for the initial values for line parameters. These values should be separated by spaces and can be entered as SIC expressions as follows: Code, Value, Code, Value, Code, Value, Code, Value for GAUSS method -Temperature- -Position- -Width- for SHELL method -Area (K.MHz)- -Position- -Width (MHz)- -Horn/Center- for NH3 method -Temperature- -Position- -Width (km/s)- -Opacity- The code is an integer interpreted as follows : 0 adjustable parameter 1 fixed parameter 2 adjustable parameter (head of group) 3 parameter fixed with respect to parameter coded 2 or 4 4 fixed parameter (head of group) Codes 2 3 and 4 are used to fit dependent lines (e.g. HCN, for which the displacements are 4.842 and -7.064 km/s, or -1.431 and 2.088 MHz, and line ratios 1:0.6:0.2). Codes 0 and 1 only are allowed for NH3 method. The value for a parameter with code 3 should be either the ratio to the head of group value (for both intensity and width), or the offset from head of group (for the position). If the cursor is on, you will have to define the boundaries of the N lines. The program then computes the moments of the spectra between these boundaries as input values. You cannot change the codes in this way which is meaningfull ONLY FOR FREE and INDEPENDENT lines. This cursor selection is not used if you specify the /NOCURSOR option. Guesses can be read from an input file using the /INPUT File_Name option. "Free" format is used to read in this input file. The first line must contain the number of lines, others are as the input at the keyboard. Command LINES is not supported for method CONTINUUM. 1 MAP ANALYSE\MAP [Match|Where] [/CELL Size_X [Size_Y]] [/GRID] [/NOLABEL] [/NUMBER] Makes a map of spectra. Used in conjonction with the keyword W (Where), MAP will plot the location of spectra in the current index. Otherwise, MAP computes the offset limits for the spectra in the current index, automatically adjusts the size of a spectrum depending in the current box size, then draws the spectra at the position corresponding to the offsets, and finally draws the axes with labels. The only constraint of MAP in this mode is that the X and Y limits must be fixed so that the spectra share a common scale. Argument MATCH can be given to force matching the X to Y ratio, which is otherwise adjusted so as to make the largest possible drawing. The options control presentation details. The actual graphic output is delayed until the end of the index has been reached. This process may be slow on some inplementations and graphic terminals when the number of spectra is large. The MAP command can be interrupted by typing Control-C. 2 /GRID ANALYSE\MAP/GRID A grid is plotted to separate the spectra. 2 /CELL ANALYSE\MAP/CELL Size_X [Size_Y] If the option /CELL is given, the cell size of the grid is taken to be Size_X by Size_Y (default Size_X) current angular units. Spectra always fill the cells. 2 /NOLABEL ANALYSE\MAP/NOLABEL If the option /NOLABEL is present, the coordinates are not written but ticks are still present. Tick can be suppressed by specifying a tick size of zero in command SET TICK. 2 /NUMBER ANALYSE\MAP/NUMBER If the option /NUMBER is present, the observation number is written for each spectrum in the upper left corner. 1 MEMORIZE ANALYSE\MEMORIZE [Name] If a Name is specified, put a copy of the current observation in a "memory" called Name. The observation can later be retrieved using command RETRIEVE. If no Name is specified, list the currently defined memories. Up to ten memories can be defined at a time. MEMORY Name/DELETE can be used to delete the "memory" area. 1 METHOD ANALYSE\METHOD Arg [Parameters ...] Select the line profile for fits. Five type of profiles are available. - GAUSS Select gaussian profiles. Up to 5 Gaussian, dependent or independent can be fitted in a spectrum, according to command lines. Variables are Area Velocity and Width (FWHM). - SHELL Horn-type profiles for Circumstellar Shells. Variables are 1) Area, 2) Position, 3) Full Width at zero level, and 4) Horn-to-center ratio (-1 for parabolic optically thick lines, 0 for flat-topped lines, Infinity for perfectly double peaked lines). Profiles are symmetric. Up to 5 dependent or independent lines may be fitted. Area units are K.MHz. - NH3(1,1) or NH3(2,2) or NH3(3,3) Compute ammonia line profiles, with the assumptions of gaussian velocity distribution and equal excitation temperatures. Variables are 1) Main group opacity times ( Excitation temperature minus Background temperature), 2) Velocity, 3) Width, and 4) Main group opacity. Up to 3 independent lines may be fitted, but it may be slow... - HFS FileName for HyperFine Structure This is very similar to the NH3 method, but the coefficient to describe the hyperfine structure are read in the file FileName. The first line of this file must contain the number of hyperfine components (< 40). The other lines must contain, for each component, the velocity offset and the relative intensity. The parameters are the same as for NH3 method. - CONTINUUM for Continuum drifts. Fit a single gaussian and a linear baseline at the same time, with automatic determination of the parameters. Uses beam-switched information for two-component fitting if required. The command METHOD CONTINUUM accept additional arguments, which can be used to customize the fit. METHOD CONTINUUM [Width [Area_Ratio [Width_Ratio]]] where - Width is the width of the beam in arc seconds. By default, Width is a free parameter. - Area_Ratio is the reference to main beam area ratio, used for beam- switch observing. By default, Area_Ratio is 1. - Width_Ratio is the reference to main beam width ratio, used for beam-switch observing. By default, Width_Ratio is 1. A * can be used instead of a numerical value for each parameter, and indicates that the parameter is free. Note that in command LINES, one enters the peak line antenna temperature instead of the first variable in the guesses, except for SHELL method. Command LINES has no effect for CONTINUUM. CONTINUUM and GAUSS methods only are relevant for continuum data, while GAUSS, SHELL, NH3 and HFS methods can be used for spectroscopic data. 1 MODEL ANALYSE\MODEL variable [X_variable] Generates a CLASS spectrum from a 1D SIC variable. Note that the 1D variable can be (and frequently is) a subset of a higher dimensionality variable. This command is typically used either to generate test data, or more frequently to transfer data from a GILDAS image (or cube) to CLASS. A second variable, if present, is interpreted as an associated axis scale for a presumably non linear frequency axis, as can occur for non resampled data from AOS backends (or optical spectra). Please note however that non linear axes are a largely untested feature in CLASS. 1 NOISE ANALYSE\NOISE [S [NEW]] Generates and draws a spectrum with only gaussian noise of rms S, to be compared with actual receiver noise. S defaults to the rms of the current spectrum. If argument NEW is present, the spectrum is not plotted but the R buffer is copied in T, and the generated noise put into the R buffer. 1 POPUP ANALYSE\POPUP Number or ANALYSE\POPUP OffX OffY POPUP Number will display a separate window containing a complete plot of an observation specified by number, if it has been previously displayed by a command STAMP or MAP. POPUP OffX OffY will display a separate window containing a complete plot of an observation with offsets matching OffX and OffY, which has been previously displayed by a command MAP. If more than one observations have the same offsets, only the first one will be displayed. In both cases, POPUP assumes that the index has not been modified since the last STAMP or MAP command. POP with argument will active the cursor to select interactively th eobservation to be displayed. Point the cursor to its spectrum in the MAP or STAMP display and press P or click the mouse second button. Use the third button or E to exit the cursor mode. The first button or any key will display the offsets or observation number. 1 PRINT ANALYSE\PRINT [Arg] [/OUTPUT File] [/TABLE File] PRINT enables one to list some parameters in a GreG compatible format. Arg is an optional argument to precise what to print : - FIT Results of profile fits in current method (default) The information written is (1) the observation number, (2,3) the offsets in current angle units, (4,5,...) fit results of the current method in the same order as in DISPLAY command. For the Gauss Method, this order is: (4) Line Area, (5) its incertainty (6) Position (7) incert. (8) Width (9) incert. (10) Intensity (11) Base RMS (12) Line RMS. - AREA Integrated area computed by BASE The information written is (1) the observation number, (2,3) the offsets in current angle units, (4,5) Area and rms noise. - AREA v1 v2 v3 ... Area in velocity ranges v1 to v2 and v2 to v3, ... etc. The information written is (1) the observation number, (2,3) the offsets in current angle units, (4,5,...) the areas. - CHANNEL List list of channels (in the same format as the FOR-NEXT loop). Up to 15 channels at a time may be listed. The information written is (1) the observation number, (2,3) the offsets in current angle units, (4,5,...) the channels value. - MOMENT m1 m2 ... Moments of the observations (area, mean velocity, width) in the intervals [m1,m2] [m3,m4] etc... Current units is used. The information written is (1) the observation number, (2,3) the offsets in current angle units, (4,5,6) area, position, width, (7,8,9) etc... - POINTING Results of CONTINUUM method fits printed in a format adapted to pointing constants measurements. The information written is (1,2) The observation and scan number, (3) a code for scan direction, (4,5) Azimuth and Elevation in degrees, (6) Time in hours, (7,8) Collimation and error along drift direction, (9,10) the antenna number and station code, (11,12) Width and error, (13,14) Intensity and rms noise, (15) Source name. - FLUX Results of CONTINUUM method fits printed in a format adapted to flux measurements. The information written is (1,2) The observation and scan number, (3) a code for scan direction, (4,5) Azimuth and Elevation in degrees, (6) Time in hours, (7,8) Position and error along drift direction, (9) the antenna number, (10,11) Width and error, (12,13) Intensity and rms noise, (14) Image gain ratio, (15,16) Signal and Image frequencies, (17) Source name, (18) Observing date. /OUTPUT File Write the results in a formatted file named File instead of on the screen. The file is suited for later processing by GreG, in particular for contour maps. /TABLE File Write the results in a table named File. This table can also be accessed by GreG, in a much faster (50 times) way than a formatted file. Tables can also handle many more columns than formatted files, and they are not limited in precision by the formatting. Mathematical operations can be done directly on the table columns. However the FIT results cannot be written to a table. 1 REDUCE ANALYSE\REDUCE This command reduces a SKYDIP observation, by fitting the sky emission using atmospheric parameters written in the observation, and displays the results : Water vapor content and Telescope Forward Efficiency 1 RESAMPLE ANALYSE\RESAMPLE NX Xref Xval Xinc UNIT [shape] [width] [/NOFFT] This command resamples a given spectrum on a specified grid. The arguments are: NX the number of channels desired Xref the new reference channel Xval the value at the reference channel Xinc the new channel separation UNIT either VELOCITY or FREQUENCY to indicate the units used for Xval and Xinc. SHAPE (optional) a keyword for the frequency response of the synthesized channels: - TBOX means a rectangular function in the delay domain, as for unsmoothed correlator channels (sin(pi.x)/(pi.x) in the frequency domain) - TPAR means a parabolic function in the delay domain, as for smoothed correlator channels (smoothing function used at Plateau de Bure) - FBOX means a rectangular function in the frequency domain (as for filterbank channels) - FTRIANGLE means a triangular function in the frequency domain (as for filterbank channels, Hanning smoothed) By default the shape of the output channels is the same as the shape of the input channels. WIDTH (optional) the width of the output channels in units of the channel separation. The default (and minimum value) is 1. The resampling is done in the delay domain, and the data are Fourier transformed, unless the option /NOFFT is used. The Fourier transform is divided by the transform of the input channel shape, then extrapolated by zeroes (if interpolation in the frequency domain is required), multiplied by the transform of the desired channel shape, and finally transformed back to frequency domain. The output channel separation usually needs to be slightly rounded (to enable use of the FFT, the input spectrum must contain a round number of output channels). To obtain an exact channel separation, the NOFFT option is necessary. The WIDTH parameter is used to provide an oversampled output spectrum if needed. The shape of the input channels is derived from the backend number read in the telescope name for IRAM-30m spectra: TBox for autocorrelator, FBOX for other (filter) backends. Otherwise, rectangular filter channels are assumed. Resample may be used on on-the-fly scans too. 1 RESIDUAL ANALYSE\RESIDUAL [n] Compute and draw the residuals of the profile fit. Copies R into T, then subtract the line profile number N from the spectrum in R. If N is 0, the sum of all components is subtracted. 1 RETRIEVE ANALYSE\RETRIEVE Name Retrieve the observation in memory "Name" and copy it in the R observation. 1 SMOOTH ANALYSE\SMOOTH [Arg...] SMOOTH copies R into T, then degrades the frequency/velocity resolution of R. The arguments are used to specify the method : - HANNING HANNING smoothing is used (default), in which the new spectrum has twice less channels and twice less resolution. - AUTO uses a cross-validation smoothing with automatic choice of smoothing parameter. To be used with caution. Noise in adjacent channels must be independant, otherwise the algorithm flatly refuses to do any smoothing at all (but needs a long computing time to realise that...). - GAUSS Width convolves (by multiplication in the Fourier plane) by a gaussian of specified width in current units. - BOX Ncan averages Ncan adjacent channels to produce a spectrum with Ncan less resolution and channels. - NOISE Flux Nc for each channel, sums at most Nc neighbouring channels until a total flux Flux is reached in the sum. Then attributes the average value (sum divided by number of channels added) to the channel. This smoothing is non-uniform, strictly positive, and has an obvious tendency to produce wings... 1 STAMP ANALYSE\STAMP NX [NY] [/NUMBER] This command is intended to have a quick look at all the observations in the index. It makes a plot of all the observations in the index, arranged in a "map" of NX by NY observations. (If the index contains many observations, they will be very small, hence the name of the command). The option /NUMBER writes the number of each observation in the upper left part of its box. The command does not require that X and Y scales be fixed. Command POPUP allows a selective zooming of any observation displayed by command STAMP. 1 STRIP ANALYSE\STRIP File STRIP creates a GILDAS 2-D image which can be used later by GILDAS to produce Velocity-Position plots using command RGMAP. STRIP works on the current index, and checks that it defines a true strip along one of the main directions ; that means you should have build the current index with FIND/OFFSET Value * or FIND/OFFSET * Value. The current X unit is used. The name of the output file must be given. If you want to produce a strip not along one of the main directions, you will have to change the offsets of your spectra using command MODIFY OFFSETS, and writing the modified spectra in a new output file. STRIP can also produce a GILDAS 2-D image from a set of parallel continuum drifts. The drifts must have the same steps, and be regularly spaced. The index must define a coherent map. 1 END