Flats and 2x2 binning of data

[swag72]

You can tell I'm trying something new due to the dirth of silly questions!!!

Last night I took flats for the red data then went on and took a nights worth of red. But I decided to try binning and binned the red channel 2x2. So I have 1x1 binned flats and 2x2 bin data. Should I have taken 2x2 flats? Or can I upsize the light frames and THEN use the calibration files?

Just as an aside as well - bias frames - They need to match camera orientation right? So the ones I took ages ago, the camera has been on and off many times since, are no good now?

I said they were silly questions!!!

[ollypenrice]

Your camera doesn't know which way up it is when you have stopped any light getting into it. The bias is an image of the read noise, not the outside world, so orientation doesn't matter.

As for resizing flats or lights to match each other: the binned signal will be stronger so you probably won't get a decent match with unbinned. I'd expect some non linearity to differ in each. Interesting to try it though!

You do need binned flats for binned images really. Also binned bias and binned darks.

Olly

[swag72]

Cheers for that Olly - I'll retake the flats tonight and remember to bin them and the bias!! Blimey, nothing in this game is ever easy, is it?!!

[martin_h]

Interesting, I read somewhere that you could re-size a flat to the binned image size and that would do the job, but of course I cant find where I read it now. It went along the lines of darks and bias need to be the same pixel scale to work, but a flat is just a representation of the gradient/crud in the image train.

[martin_h]

Found it!

Down near the bottom of the page in the summary........................ http://www.dens-astropics.org.uk/page%2024.htm

[ollypenrice]

Aha, if Dennis says you can do then you can! I'm still surprised though. The professional observatories consider that flats contain pixel by pixel information so they would not do it. I interpret flats as you do, though, Martin.

Give it a go, Sara, and report back in finest detail!

Olly

PS Sara, you are not binning in the Pentax? Bit short on FL, I'd say.

[swag72]

Cheers for that Martin - I suppose I could try it and report back!! I'm not very good at spotting that they've worked though!!

I will have a go Olly, but you'll have to look at whether it's made a difference!! Yes binning on the Pentax - Should I not be?

Also, if my flats are at 1x1 and my lights at 2x2 - Do I resize the lights then use the flats, or scale the flats down, use them then scale the calibrated lights back up?

[martin_h]

I have actualy tried it...a long time ago and seem to remember it did the trick - ie. it got rid of dust bunnies etc. what it did at pixel scale I wouldnt like to say tho.

[martin_h]

Cheers for that Martin - I suppose I could try it and report back!! I'm not very good at spotting that they've worked though!!

I will have a go Olly, but you'll have to look at whether it's made a difference!! Yes binning on the Pentax - Should I not be?

Also, if my flats are at 1x1 and my lights at 2x2 - Do I resize the lights then use the flats, or scale the flats down, use them then scale the calibrated lights back up?

I rescaled the flat to suit the light, ie in software I binned the flat.

[swag72]

Yes I rescaled the flat and then created a master and used that to calibrate the red channel. Came up OK by all accounts!!

[ollypenrice]

Yup, Sara sent me a copy and it looked Very Nice Thank You Very Much!!

Olly

[swag72]

Yup, Sara sent me a copy and it looked Very Nice Thank You Very Much!!

Olly

Can I ask Olly - Were you surprised at that given I was binning 2x2 with the little Pentax? When you said I hope you're not binning, what would you have expected? .............. I know I've not got the tell tale blue yet!

It's interesting to note the times when the figures and what should or shouldn't work is different from reality.

[dph1nm]

The professional observatories consider that flats contain pixel by pixel information so they would not do it.

Oh, with my professional hat on I think one would happily do this. The pixel to pixel variations will still be correct for the binned pixels I think. You would have to downsize the flats though, not upsize the data.

NigelM

[swag72]

I have decided that as I've not managed any more data since the red, I shall retake the flats at 2x2 and the bias at 2x2 - Calibrate the files and stack them, then upsize to match the 1x1 luminance.

It got so complicated that my head almost exploded!!!!

[martin_h]

I have decided that as I've not managed any more data since the red, I shall retake the flats at 2x2 and the bias at 2x2 - Calibrate the files and stack them, then upsize to match the 1x1 luminance.

It got so complicated that my head almost exploded!!!!

Thats what I would do - upsize the colour to match the Lum......never mess with the luminance, thats the one that holds the detail.