Proper Use of PI Darks, Flats, and Bias Frames Using IC

[sreilly]

Continued..

[tsaban]

These are the calibrated as noted images.....

This is the answer you were looking for sreilly  , your flat frame causes the trouble! Something is wrong with that image.

If I were you I would delete all intermediate step frames and integrate the flat again.

Clear skies
Tahir

[pfile]

how was the master flat calibrated? i'd suggest using separate flat darks (which match the flat duration) if you have not done so.

[sreilly]

This is the answer you were looking for sreilly  , your flat frame causes the trouble! Something is wrong with that image.

Actually, the flat has some issues but it's not the real problem. Using these same calibration frames in MaxIm, the resulting image is here http://www.astral-imaging.com/LBN999.htm and this is only the one night's data. DBE does a good job of helping correct the flat issue.

[sreilly]

how was the master flat calibrated? i'd suggest using separate flat darks (which match the flat duration) if you have not done so.

I don't typically use flat darks at all as they are under 10 secs each and I've never seen an improvement. I usually use the master bias though when making the master flats. I should also mention that I use ACP for automation of imaging and the sky flat script is constantly changing each exposure as the sky darkens/lightens. My minimum exposure for flats is .1 second and ranges to 10 seconds before shutting down. I then use a Sigma Reject Mean method with a sigma multiplier of 3 to create the master flats.

[Cleon_Wells]

Steve, looking over your Mean Stats on the MasterBias (971) and MasterDark (931.5) , I assume your Dark subs were not Caled before Integration. If you subtract your MasterBias from one of you Dark subs the Mean Stat of this one Dark sub should be close to 0; I wonder if you could do a quick test with one Dark sub?
Cleon

[pfile]

how was the master flat calibrated? i'd suggest using separate flat darks (which match the flat duration) if you have not done so.

I don't typically use flat darks at all as they are under 10 secs each and I've never seen an improvement. I usually use the master bias though when making the master flats. I should also mention that I use ACP for automation of imaging and the sky flat script is constantly changing each exposure as the sky darkens/lightens. My minimum exposure for flats is .1 second and ranges to 10 seconds before shutting down. I then use a Sigma Reject Mean method with a sigma multiplier of 3 to create the master flats.

that's okay too, i was going to suggest at the very least one needs to bias subtract the master flat (or the flat subs) but since the flat dark contains the bias noise...

[pfile]

by the way, i did not realize until right now that the source of the images was maxim. when the images come into pixinsight, are they still in i16 format? or has maxim save them as f32?

maxim does some silly stuff with f32, sometimes even representing i16 data as f32 data in the range [0.0 .. 65535.0]. this could be the source of your troubles as by default PI wants f32 files to be in the range [0.0 - 1.0]. do you get an out of range warning when you open the subs?

[sreilly]

No problem with the MaxIm data and it's saved as IEEE Float when I calibrate in MaxIm. Raw data is in 16 bit depth. PI doesn't have a problem reading these files but anything I do in PI and save the resulting images they are saved as 0,1 and no longer read properly in MaxIm. The images I save in PI are always saved as 32 bit. I calibrated the raw images using PI using only the master flat and master darks which appear normal. The use of the master Bias made in PI seems to be causing the problem.

Not sure I've replied to all the replies as I've been running around all day getting stuff done.

Thanks,

[pfile]

PI doesnt have a problem reading the files but it is important to configure the fits reader properly...

at any rate i know i had an impossible time using master flat files that had been created in maxim. of course, that's not your problem. your flats look like they are properly exposed. but for me the flats were just weird - the scaling factors were too high and i had to adjust the histogram of the flats to make them work right. like until the data was all in the range of 0.98-1.0.

i'm lost though, i thought it was the flat that was causing the problem... or is it the bias?

[sreilly]

Definitely the bias frame.

Thanks,