Error 'Image is completely blanked' - linmos and regrid

[sarahr]

Hi,

I have come across the same error (see below) in two miriad tasks, linmos and regrid, which I assume is a bug.

1. I ran linmos on a single data cube, to do the primary beam correction. Following this, any tasks I run on the PB-corrected cube are met with the error that the image is completed blanked. Error messages pasted below.

Sigest: version 14-Apr-1999
### Fatal Error: Image is completely blanked

moment: Revision 1.17, 2011/10/25 09:43:44 UTC
Finding image minimum and maximum
### Fatal Error: All pixels are 0.00

2. I ran regrid on some images. Following this, I get the same 'Image is completely blanked' error with any task I try to run. Note that for some reason this only occurred with certain images, but not with others. James Allison says he has had the same problem with regrid.

Please note that I have the most up-to-date version of miriad, installed from source. Prior to that, I was running an old binary distribution (late 2011), in which both of these tasks worked just fine.

Any help sorting this out would be much appreciated!

Thanks,
Sarah Reeves.

[Mark.Wieringa]

Hi Sarah,

I just tried running linmos on a cube and sigest was happy with the results, so there must be something different about your data.
Could you let me know where the data is, or submit a bug report using mirbug?

Cheers,

Mark

[sarahr]

Hi Mark,

Thanks for your reply.

After you said that linmos worked fine for you, I tried a few more things.

I tried making a sub-cube (the original cube was very large!) - and when I ran linmos on the sub-cube I no longer got an error. From a few tests, it seems that the number of channels in the cube (and not the file size) is the relevant factor in whether it produces an error or not. You may want to check this for yourself though. Please note that I have also tried running linmos with and without the 'bw' parameter set, and both produce the same result.

I'm still having problems with regrid, and have no idea what's causing that.

I'll send you the location of the files in a separate email.

Thanks again!

Sarah.

[Mark.Wieringa]

Hi Sarah,

I've found the problem in linmos - it happened only for cubes with negative velocities, and you were right, it was introduced around Sep 2012. I've implemented a fix in version 1.22 which should appear tomorrow.

The regrid problem has now been fixed too, it must have appeared with coordinate routine changes last year.
Let me know if you notice any more strange behaviour from these tasks.

Cheers,

Mark

[sarahr]

Hi Mark,

Great, thanks very much!

Cheers,
Sarah.