Starshoot Raw Frames

[Paul Mahoney]

Hi again,

Hope I'm not being a pain but I picked this up from the Maxim Group. It is a reply to a question regarding File saving from Douglas George of Maxim.

"The FITS standard does not include support for unsigned numbers. As defined by
the standard, 16-bit integers run from -32768 to + 32767. Not very convenient
for saving 16-bit CCD images that run from 0 to 65535! At first glance you
would need to save in 32-bit mode, which would double the disk space required.

The value of a pixel is defined to be (NOST FITS Standard Equation 5.3) as:

physical_value = BZERO + BSCALE * array_value

As you can see, choosing a value of 32768 results in a physical pixel data range
from 0 to 65535. So we set BZERO to 32768, and *subtract* that number from
every pixel as we write it to the file. When a program that correctly
implements the FITS standard reads the file, you get the correct numbers.

This is what you need to save a 16-bit CCD image in a 16-bit FITS file.

Some programs violate the FITS standard and save unsigned data. There are
compatibility features for this, see:
http://www.cyanogen.com/help/maximdl/SettingsFiles.htm"

Regards,

Paul
London, UK

[Jack Harvey]

Looks like there is an issue with Bzero which reads 0 in your fits header and

BZERO      = 32768.000000000000

in mine (the 32768 to my understanding is the correct value as stated in the thread you copied above from Maxim).

[Paul Mahoney]

Hi Jack,

Yes, I agree but I can't find a way of saving in Maxim that gives me a  BZERO of 32768.

Can you tell me which version of Maxim you are using and which setting you use to save your images.

I really appreciate it,

Paul
London, UK

[Jack Harvey]

At present I am using 5.08 but as I said have not had issues with Maxim 4 or any ver of 5.  From Maxim Help menu

BZERO – this value should be added to the data array values when reading the FITS file. For 16-bit integer files, MaxIm DL writes 32768 (unless overridden by the Settings dialog).

My guess is an issue with the settings dialog?  So your query may best be followed up on the Maxim Forum?
On my  Settings window under Files I have "Treat data as Unsigned" as the only button enabled.  Do you?

[Paul Mahoney]

Hi again Jack,

Yes I have "Treat data as Unsigned" checked in my settings. I am also using V5.08 BUT I cannot save a file and get the FITS Header BZERO to read 32768.

However, you are right. This is not really for this forum and I will pursue it in the Maxim Group if I can't get to the bottom of the problem. However, just one last thing and I hope it doesn't sound impertinent. Have you checked the FITS header of an image actually captured using V5.08 that it Reads BZERO = 32768

Thanks for your help, Jack, much appreciated.

Paul
London, UK

[Paul Mahoney]

A little more investigating and I find that it is only possible to save as 16-bit Int in Maxim. No other alternatives in the save as dialogue box for 16 bit.

In a general settings area there is an option to OPEN 16 bit images with negative numbers to be either 'Treated as unsigned' or to 'Add offset to make positive'

With 'Treat as unsigned' checked, I save the image in 16-bit Int and open in PixInsight and the unwanted pixels are in the core of the stars.

With 'Add offset to make positive' checked, I save the image in 16-bit Int and open in PI and the unwanted pixels are gone.

When I read the FITS Header of an image just captured but not saved the BITPIX reads -32/8 unsigned int, 16&32 int, -32&-64 real.

NOT in any version does the  FITS Header BZERO = 32768.

Juan, Niall & Jack, am I right in assuming that if I choose the setting 'Add offset to make positive' that my images will be correctly read by PI. (If so Jack, I am confused that you do get a BZERO reading of 32768 and haven't experienced any problems AND use the 'Treat as unsigned' option in Maxim.

I am really beginning to feel that I am becoming a nuisance with all my posts on this subject, particularly that it more concerns other software of which PI has no control. But I do not mean to be and I am very grateful to everyone for their input. I did ask PI before purchase if there were any issues with using Maxim captured images and was assured "You'll have no problems at all if you save your raw images as unsigned 16-bit integer FITS files in Maxim DL". AND I thought I was doing just that!


Kind regards to all,

Paul
London, UK

[Carlos Milovic]

Hi Paul
Don't worry about that! We are glad to follow this discussion There is always something new that we learn

I think that adding the pedestal should work, since it is converting the data to unsigned format. BUT, make sure that it is using 16 bits and not 15 (take a look at PI's histogram of the image, if there is data beyond 0.5. Also you may check with the readout (use 16bits instead of the normaliced range) for even and odd numbers. (Most probably this is not necessary, but who knows...)

[Jack Harvey]

No not a nuisance at all.  I am wondering if the camera firmware has any bearing on this.  I am using Apogee cameras and you are using a different model?