[fitswcs] Complex values in FITS {External}

[egreisen]

On 2023-02-15 07:46, Patrick Leahy via fitswcs wrote:
> Hi Folks,
> 
>                 Not sure if this is the right list, but there is
> currently a discussion going on about the right way to store Faraday
> synthesis spectra and spectral cubes, which are complex-valued. One
> way is to use a COMPLEX axis as discussed in WCS paper I (and the
> original FITS papers).  But there seems to be a long-standing
> inconsistency in the FITS standard on this point. All references agree
> that COMPLEX value 1 is real and 2 is complex. But the example of this
> in the standard, Table 28, is:
> 
> SIMPLE=T
> 
> BITPIX=32
> 
> NAXIS=2
> 
> NAXIS1=2
> 
> NAXIS2=100
> 
> CTYPE1='COMPLEX'
> 
> CRVAL1=0.
> 
> CRPIX1=0.
> 
> CDELT1=1.
> 
> END
> 
> Here the complex axis runs from 0 to 1, not 1 to 2 as expected. Is
> this a mistake or am I misreading?
> 

Actually with CRPIX1 = 0, the two pixels have value 1 and 2.

Eric Greisen