“ERROR: arithmetic exception” in imexamine

Adam Solomon wrote on Oct 26, 2006

Hello! Having a slight problem with IRAF here (running X11 port on Cygwin for Windows XP). I'm analyzing a star field with imexamine, and when I try to do a , on certain stars, I get "ERROR: arithmetic exception", and the program leaves the package. I can measure some objects in the field, mostly dimmer ones. Many of the brighter objects give me this error. Does anyone know why this is happening, and how to fix it (if that's possible)? Thanks!

---

Mike Fitzpatrick wrote on Oct 26, 2006

Without seeing the data or your parameters it's hard to say more than an 'arithmetic exception' is something like a divide-by-zero, log() of a negative number, etc. It is sometimes a coding error but in a heavily used command/task more often it is data or parameter related. Try unlearning the task and resetting the params slowly to see whether it makes a difference. You might also look at the pixel values with the 'z' key to see if they "should work". Otherwise, please post a url to the image and your parameters (i.e. "cl> dpar imexam").

Cheers,
-Mike

---

Adam Solomon wrote on Oct 26, 2006

Thanks! I used "z" and here are the pixels on one of the "bad" stars, taken as close to its center as I could get it:

      390   391   392   393   394   395   396   397   398   399   400 
 442 55.99 70.72 89.43 112.7  128. 159.9 162.1 173.8 175.7 146.9 122.1
 441 63.19 79.76 102.3 127.6 169.7 192.6 216.3 220.1 214.1  189. 142.7
 440 69.85 83.83 114.9 151.7 196.2 221.7 249.9  253. 240.5 221.1 183.6
 439 73.39 100.2 134.8 162.5 198.1 247.6 263.1 276.8 267.9 222.9 199.4
 438 74.57 104.6 133.4 159.8 213.1 230.8 272.8 275.3 263.3 225.2 178.1
 437 76.65 95.43 123.3 169.1 195.7 214.4 250.8  249. 230.2  205. 166.7
 436 69.11  93.4 112.9 141.6 175.3 199.2 209.5 226.2 215.2 174.5 151.9
 435 68.15 80.59 93.79 123.1 149.3 165.7 175.3 170.3 167.7 148.4 128.1
 434 56.72 62.53 89.11 96.17 118.7 126.2 141.9 135.1 126.9 114.9 101.1
 433   45. 46.08 62.15 80.17 88.81 99.65 107.7 110.8 96.94 81.88  73.7
 432  45.2 45.09 52.54 61.61 60.27 73.38 76.12 87.75  66.8   65. 54.75

So I see nothing wrong but maybe I'm missing something. My parameters are:

imexamine.input = "*10-25.fits"
imexamine.frame = 1
imexamine.image =
imexamine.output = ""
imexamine.ncoutput = 101
imexamine.nloutput = 101
imexamine.logfile = "log.txt"
imexamine.keeplog = yes
imexamine.defkey = "a"
imexamine.autoredraw = yes
imexamine.allframes = yes
imexamine.nframes = 0
imexamine.ncstat = 5
imexamine.nlstat = 5
imexamine.graphcur = ""
imexamine.imagecur = ""
imexamine.wcs = "logical"
imexamine.xformat = ""
imexamine.yformat = ""
imexamine.graphics = "stdgraph"
imexamine.display = "display(image='$1',frame=$2)"
imexamine.use_display = yes
imexamine.mode = "ql"
rimexam.banner = yes
rimexam.title = ""
rimexam.xlabel = "Radius"
rimexam.ylabel = "Pixel Value"
rimexam.fitplot = yes
rimexam.fittype = "moffat"
rimexam.center = yes
rimexam.background = yes
rimexam.radius = 3.
rimexam.buffer = 12.
rimexam.width = 12.
rimexam.iterations = 1
rimexam.xorder = 0
rimexam.yorder = 0
rimexam.magzero = 25.
rimexam.beta = INDEF
rimexam.rplot = 8.
rimexam.x1 = INDEF
rimexam.x2 = INDEF
rimexam.y1 = INDEF
rimexam.y2 = INDEF
rimexam.pointmode = yes
rimexam.marker = "plus"
rimexam.szmarker = 1.
rimexam.logx = no
rimexam.logy = no
rimexam.box = yes
rimexam.ticklabels = yes
rimexam.majrx = 5
rimexam.minrx = 5
rimexam.majry = 5
rimexam.minry = 5
rimexam.round = no
rimexam.mode = "ql"

(Included rimexam since I changed its parameters a bit, too).

For some reason, pressing "a" works fine, I don't get the same error. Any thoughts are appreciated, thanks!

---

Francisco Valdes wrote on Oct 26, 2006

The profile fitting algorithm is complex and is known to occasionally have problems with certain data. If you can extract a subraster (imcopy with an image sections) that continues to reproduce the error I would be glad to diagnose it. Send me an email with the small image attached or else place it at ftp://iraf.noao.edu/pub and let me know about it.

Yours,
Frank Valdes

---

Last post on Oct 26, 2006