Common mistakes

Opening the profile (-o) in DS9. The profile written by -o (traditionally n1234.dat) is a table of numbers, not an image. Use --reg for the ellipses on the image, and -m / --residual for images. See The profile. (In the traditional naming, n1234.prf is the model image from -m, which DS9 does open.)

Adding MODEL or MODEL=file to get a model. Since 0.1.4, -m FILE alone computes and writes the model. A bare MODEL is ignored; MODEL=file is an error. Name the file with -m.

A centre off by half a pixel. ELLIPROF coordinates are FITS/DS9 pixel numbers minus 0.5. A galaxy at DS9 pixel (567.5, 562.5) has X0=567 Y0=562. See How ELLIPROF fits a galaxy.

Expecting ELLIPROF to find the galaxy. It does not. X0, Y0 are required. Give the galaxy's peak; a centre on a star or several pixels off makes the inner isophotes wander.

Using SKY= to subtract the sky. SKY= only affects the de Vaucouleurs fit ELLIPROF prints. To subtract a sky, use --sky or --sky-image. See Sky and masks.

Forgetting the sky. Without --sky or --sky-image, nothing is subtracted. The outer profile is then wrong, and masked pixels (set to 0) are no longer distinguishable from sky-level pixels.

An inverted mask. By default, in elliprof masks 0 = bad. If your mask has 0 for good pixels and nonzero values for bad ones (common for bad-pixel maps and segmentation maps), add --mask-convention zero-good instead of inverting the file. elliprof reports how many pixels are masked. If that number is most of the image, the convention is probably the wrong one.

Treating mask values as weights. The mask is logical. 0.5 means "good", not "half weight".

A mask or sky image of a different size. They must match the science image exactly. elliprof never resizes or reprojects. Cut them to the same pixels first.

Forgetting quotes around an extension. elliprof galaxy.fits[SCI] ... fails in many shells, because the brackets are a shell pattern. Write 'galaxy.fits[SCI]'.

Reading A4 as a position angle. A3/A4 are phases around the ellipse, measured from the major axis. The position angle is alpha. See Harmonic analysis.

Comparing I4 with a4/a from another program. I4 is an intensity amplitude. Convert: a4/a ≈ I4 cos(4 A4) / (−slope).

Reading alpha as a sky position angle. alpha is counter-clockwise from the image +y axis. Use the WCS to convert to east of north. Also, 2° and 178° are nearly the same orientation; unwrap before plotting.

Believing the outer isophotes. Far out, the sky and the mask dominate. Test the sky's effect and watch for jumps. See Galaxy surface photometry.

Believing the model inside R0 or beyond R1. It is extrapolated there, not fitted.

Treating --model-harmonics as a fitting option. It changes the model image, not the profile. Only --sixth-order changes the fit.

Calling the residual an SBF measurement. The residual is an input to an SBF analysis. See ELLIPROF's role in SBF.

Using a residual that is 0 on masked pixels as data. Masked pixels in the residual are exactly 0, not measurements. Exclude them with the mask in any statistics.