Page 1 of 1

Suggestion: Flats to include camera name

Posted: Thu Oct 03, 2019 3:11 am
by celkins
Hi, Robin,
another suggestion to come out of my session tonight is to have the camera name included [at the start] of the saved flat field name - it's a pain with 2 cameras, having the flats in the same folder, but only named for the time taken - I have to remember which I did first, since there's not even an INSTRUMEnt header in the FITS...
This shouldn't cause any issues for single camera users, though you might have to ripple it through your code...


Re: Suggestion: Flats to include camera name

Posted: Thu Oct 03, 2019 7:01 pm
by admin

A workaround should be to use the manual filename template option and include the '{Camera}' tag in the directory name which should separate out all the files into different folders by camera.

Cheers, Robin

Re: Suggestion: Flats to include camera name

Posted: Sat Nov 16, 2019 11:54 am
by Tommohawk
It wold certainly be useful if the flat FITS header had the exposure time and gain - so you can see those details when calibrating/stacking or if doing dark flats subsequently which need to have the same settings.

Re: Suggestion: Flats to include camera name

Posted: Sat Nov 16, 2019 9:51 pm
by admin

gain and exposure are unimportant when using flats (you can take them with completely different values to the lights without ill effect). The important thing is that the optical train (scope, reducer/flattener/barlow, filter, camera) is unchanged between flats and lights which means you tend not to re-use them as much as darks.

Also, as I have mentioned before, there is no measurable benefit in subtracting a true dark flat over subtracting a flat bias frame (very short exposure flat) as SharpCap does at the moment. The advantage of using the bias frame option is that it doesn't require operator intervention and can be done entirely in software by varying the exposure time.

Cheers, Robin

Re: Suggestion: Flats to include camera name

Posted: Sun Nov 17, 2019 6:52 pm
by Tommohawk
Hi Robin
Take your point about the setting for flats being immaterial.. except that if going the dark flat route the settings do need to match the light flat. If I understand thins right. (Maybe not!)

I've read a few thread about using bias frames and one view (from a guy who seems to know his stuff) is that the offsets can be variable and don't work as well as dark flats - if I understand it correctly. That said, you clearly also know your stuff, so I'm happy to take your word for it.

However, are you saying that the offset is automatically added to the (light) flat when doing the capture using the flats capture tool?

That would mean when doing calibration and stacking, all thats needed is the lights , the (light) master flat created by SC, and the appropriate darks/masterdark, correct?

Edit: just realised there is a tick box for adding offset to the flat. But still not sure if adding offset to the light flat means no offset frame is needed when stacking? I though an offset was used to correct the lights, but a dark flat was an alternative. Maybe I have this in a muddle.

Also on another thread you said:

"SharpCap already has the option to capture and subtract bias frames when creating flat frames. The bias frames are just short exposure flats that help remove the offset level from the calculated flat field correction. In my opinion, providing your flat frame exposures are less than one second, there is no point in taking true dark flats when capturing flat frames – the amount of thermal noise building up in a CMOS camera in some one second exposure is really not worth worrying about, particularly at low gains."

Apologies - I should have read this first as it makes the point about the tick box for the offset in the flat capture. However when I try and take flats < 1 second, I get banding artifacts, so I've been using a diffuser and taking exposures more like 2-8 seconds, depending on the filter.

The reason I'm getting stressed about it is that 2 years after getting my ASI1600 I'm still really struggling with artifacts on my flats, but only on lower gain. I'm fine with higher gain and narrowband. But that's going to muddy this thread so I'll post a separate thread on that.

Thanks Tom

Re: Suggestion: Flats to include camera name

Posted: Mon Nov 18, 2019 7:34 pm
by admin
Hi Tom,

The goal of capturing a flat frame is to determine the relative intensity of light on each part of the sensor under even illumination. If you think that the signal that comes from the camera is basically made up of two parts – a constant offset and an image partner depends on brightness, illumination and other camera settings, it becomes clear that you can calculate the relative illumination for a flat either by

flat(1s exposure) - dark (1s exposure)


flat(1s exposure) - flat (0.001s exposure)

Both will correctly remove the constant offset portion of the signal. The former will deal with any dark noise accumulated during the flat exposure times but requires intervention. The latter will give a result that is 99.9% of the former providing that the dark noise over the flat exposure time can be ignored. Remember that the second option will be evenly reduced by 0.1% across all pixels and will therefore still contain data that will work properly as a flat frame.

Once you have your flat frame, you need to apply it. For this to work properly you must either (ideally) subtract a dark frame or as a second Best alternative subtract the offset from each image frame before correcting using the flat data. Without this subtraction, the flat frame correction will also be applying to the offset and/or dark frame level and therefore giving incorrect results.

My take on the situation is that unless you are taking flats that are long exposure (several seconds or more) there is no need for true dark flats, but if you're light frames are more than a few seconds then you should use dark frames as well as flat frames to get results. If you aren't using dark frames but did use the bias option when creating your flat then SharpCap will do its best – it records the offset value from the bias frame in the flat frame, and will use that to attempt to correct for the offset of the light data that flat frame.

Cheers, Robin