Help with Bias

[Ken Pendlebury]

Hi folks,

I am hoping somebody might be able to help me understand some image behavior that has baffled me...

I am attempting to take bias frames for my camera.  I have gathered the 16-bit frames with the Astroart application and inspected them.  All frames are at an acceptable level (below 1000 ADU).  To verify, I open the same frame in the Nebulosity capture application and it verifies that the images is below 1000 ADU.  When I open this very same image in PI, it shows a mean ADU of 33682!

Obviously integrating these frames as PI sees them creates a terrible bias frame. 

Strangely, when I use Nebulosity to capture the bias frames, PI does indicate that the ADU value is less than 1000.  Astroart uses an ASCOM driver for capture (QHY9 Camera) while Nebulosity uses a native driver.  I'm not sure if this is relevant at all, but though I would mention it.

Any thoughts are appreciated

Regards,
Ken

[Nocturnal]

Hi Ken,

the driver has zip to do with that. It's all about how the app saves the file. I use Neb 2 and tell it to use Maxim format and that seems to work well. Perhaps the FITS header and format in AA generated files isn't quite standard. I think PI can be told to interpret the data differently using hints. Do a search for 'fits hints' and see if anything comes up. If not then at least post the FITS headers for the file that doesn't load correctly. Without that I doubt Juan can offer much help.

[Ken Pendlebury]

Thanks Sander.  I will check the hints out and post info if I need to...

There seems to be some inconsistency with this issue...

When I open a "light" frame taken with Astroart in PI, it opens just fine, appears normal and reports all of the correct image statistics.  The error in opening seems to occur with 100% frequency when I am inspecting bias or flats (which baffles me since flats are actually lights).

Thanks,
Ken

[harist]

This might help:
http://pixinsight.com/forum/index.php?topic=1809.msg10976#msg10976

Have a look at Reply #42

Tasos

[Juan Conejero]

Hi Ken,

The problem is that those bias have been stored as signed images. You have two ways to load these images in PixInsight: modify global FITS settings to enable the "Signed integer images store physical data" option in the FITS Format Preferences dialog (go to the Format Explorer window and double-click the FITS item), or use the "signed-is-physical" input hint with the ImageIntegration tool.

More information here:

http://pixinsight.com/forum/index.php?topic=3344.msg23085#msg23085

[Ken Pendlebury]

Tasos and Juan,

Thanks so much for your help.  I feel a bit silly that I was unable to find these threads myself.

While I understand what the issue is, I have one more question about the flats I take in AA5.  As I understand it, AA5 will use 16 bit signed to save bias frames and 16 bit unsigned to save longer frames (darks, lights etc..).  My flat frames are about 6 seconds in length and in AA5 and Nebulosity show a mean of ~25K ADU. 

Making the assumptions outlined above (that this image is also 16 bit signed), I checked the box to enable "Signed integer images store physical data".  Opening this frame with this option on or off yields an image with a mean of ~63K.  I am almost certain this is user error, but I am not exactly sure what I am doing wrong.  Any thoughts on this?

I should also note that, checking this box and opening a bias frame works much better.  One slight difference is that the mean, as indicated by Nebulosity and AA5 is 914 ADU.  When PI opens it, the image shows a mean on ~1800 ADU.  This is probably not significant enough to affect a light image during calibration, but I though I would point it out...

To test this, I opened an AA5 bias frame without the "Signed integer images store physical data" checked and verified the  mean ADU was around 33k.  Using the RGB/K: $T-0.5, the image is manipulated such that the mean ADU is the expected 900.  Nor sure why checking the "Signed integer images store physical data" doubled the mean ADU in this case.

Regards,
Ken