To be compatible with what we
already have done with OMEGA and PFS data: leave the Nbins everywhere - For
NOMAD, it will be always =1 if Nbins = 1: no need to use 2D (3D) matrices; for example AOTFFrequency will be a 1-column matrix (vector) of length Nspectra |
All dimensions are of the following: [Number of spectra] x [Number of bins] x [Number of pixels] | (1 subdomain, 1 detector bin) =
1 spectrum !! For SO and LNO: Nbins = 1; and Nspectra = [Nb_observationsx Nb_Detector Bins] |
for UVIS : Nbins = 1; Nspectra = Nb_observations |
||
Dataset | Description | Dimension | SO/LNO Example | UVIS Example | Notes |
Bins | Start and end rows of each bin in the spectrum | [Nspectra x Nbins] x [2] | [116,119],[120,123],…. | [0,15] for UVIS full frame. [vstart,end] for binned spectrum | |
X | X values | [Nspectra x Nbins] x [YNb] | 2100.00, 2100.20, 2100.40cm-1 | 200,201,202nm | |
XUnitFlag | Unit of X axis. 0 = pixel, 1 = cm-1, 2 = nm, 3 = um, 4 = Relative cm-1 (used in SOIR) | [Nspectra x Nbins] | 0 | 0 | |
YError | Error on Y values | [Nspectra x Nbins] x [YNb] | Detector pixel DNs error | Detector pixel DNs error | |
YErrorFlag | Flag indicating if and what kind of error is supplied. =0 : no error given (YErrorNb =0), =1 : only one value of noise for all data points (YErrorNb =1), =2 : one value of noise for each Y value (YErrorNb = YNb) | [Nspectra x Nbins] | 0 | 0 | |
YNb | Number of data points | [Nspectra x Nbins] | 320 | 1048 | |
YTypeFlag | Type of Y values. 0 = raw DN (digital number), 1 = radiance, 2 = transmittance, 3 = radiance factor | [Nspectra x Nbins] | 0 | 0 | |
YUnitFlag | Units of Y axis. 0 = uncalibrated, 1 = none (but calibrated), 2 = W/cm2/Sr/cm-1, 3 = Brightness Temperature (K), 4=W/m2/Sr/nm | [Nspectra x Nbins] | 0 | 0 | |
Y | Y values for each bin | [Nspectra x Nbins] x [YNb] | Detector pixel DNs | Detector pixel DNs | |