Modified the IRTF reading to handle properly the data type of the flux.

bretonr · Oct 27, 2015 · 6515d4d · 6515d4d
1 parent 2bd9b11
commit 6515d4d
Showing 1 changed file with 3 additions and 7 deletions.
diff --git a/Atmosphere/Atmo_spectro_IRTF.py b/Atmosphere/Atmo_spectro_IRTF.py
@@ -333,13 +333,9 @@ def Read_IRTF(fln, oversample=None, sigma=None, tophat=None, thin=None, wave_cut
     ## The wavelength is contained in the first column and the flux in the second.
     wav, grid = np.loadtxt(fln, usecols=(0,1), unpack=True)
 
-    ## Grid values are log10(F_lambda) [cgs]
-    #grid = 10**grid * 4/cts.PI**2 # Conversion to flux units (make sure that in the Get_flux routine does not re-correct again!)
-    #grid = (2 / cts.PI / np.sqrt(3)) * 10**grid
-    #grid = (2 / cts.PI / np.sqrt(3)) * grid
-    ## NOT COMPLETELY CERTAIN
-    ## BUT: if the values are stored in log10(F_lambda) [cgs], then we need to convert as np.log(10**grid) to be usable by our scripts
-    grid = grid * np.log(10)
+    ## Data file values are stored in F_lambda [cgs]
+    ## For the grid, we store values in log(F_lambda) [log(cgs)] -- natural log -- and then the interpolator interpolates and exponentiates them because summing over the surface elements.
+    grid = np.log(grid)
 
     ## Wavelengths are often not ordered so we re-order them
     inds = wav.argsort()