creating Darks.

[ribuck]

Hi All,

I'm using Maxim Dl, and and a Starlight Xpress MX916. I started creating Darks for different exposure lenghts so that i dont have to keep doing darks everytime i change the exposure lenght.

I also have a manual filter wheel, with LRGB filters.

If i also want to avoid doing darks every time i change the filter, do i need to create darks for each filter colour as well or do i generally just need to do them for the Lumenance only ?

Regards,

Rich.

[narrowbandpaul]

nope...

a dark has the shutter closed anyway, so it doesnt matter whether there are filters in front of the ccd.

a 5 min dark can be applied to a 5minute luminance,or 5 minute Red or Green or Blue filtered image.

Dark current does not depend on the radiation incoming...since none hits the sensor.

flats however...yes. One set for each filter, since there could be dust placed arbitrarily on the filter

[ribuck]

What exactly are the flat's and Bias used for ? I thought you just used them if you suffered from dust donuts.

How do you take Bias and Flats images ?

[narrowbandpaul]

flats or flat fields are a picture of a evenly illuminated surface, taken in the same optical configuration as your normal images...ie a picture taken through the same telescope, with the focusser in the same position.

This setup helps remove vignetting...the darkening experience near the corners of an image. it also removes dust that may be on the filter, mirror/lens or the camera window.

it also removes Fixed pattern noise (for an explanation see and for other ccd noise sources here... http://stargazerslounge.com/index.php/topic,33629.msg343158.html#msg343158)

important to make sure that you take your flat at a high enough signal level....that will the flat will show the FPN and be removed....make sure your flat field image is bright enough!

believe it or not, the FPN limits your image quality more than vignetting (in terms of signal to noise ratio.)

You need to find an even illumination...the blue sky can work, but i like the use of a white cloth or t shirt over the end of the scope, and then point at a roughly even patch of ground, or sky...but make sure sunlight does not hit the t shirt directly, or your flat will have a gradient. And make sure the t shirt is pulled tight over the front of the scope.

Once you find a way that removes the vignetting, take loads of flats and average them to reduce the overall noise

Bias are easy...

just take the shortest exposure you can....with the cover over the scope or camera...ie its the shortest possible dark. again take a load

The bias frame contains offset and read noise...

offset is a constant value added to all pixels, and the read noise is the noise created by digitizing and amplifying charge. by stacking lots of frames, the read noise is cut down, and so the averaged frames represent pretty well the offset level used in the camera.

a bias is only really appropriate if your dark exposure length doesnt match the length of your light frames...

so try and keep the darks the same length as the light frames for best noise reduction.

Hope that helps

paul

[ribuck]

Paul, becuase the object i'm focusing on will be different each tiime, should i be doing flats for each different set of exposures.

I.e. Do Luminance + Flats, Red + Flats, Green + FLats, Blue + flats - Then do this all again every time i image a different object ?

I read your other post, but it was all over my head, lots of technical jargon that i dont understand yet seeing as i'm new to all of this imaging.

at the moment, I guess i just need to know the basical principals until i get more experienced with imaging, then i can start delving into the nitty gritty of it all.

[narrowbandpaul]

if you only image astronomical objects, then the focus is in the same place each time (assuming temperature effects are ignored, which they are), so one set of flats for each filter is enough...

ie...

take a set of say 20 flats for each of lum, RGB, and thats it...

go and image whatever astro object, and apply the same flats each time...

each time you take a lum image, apply the lum flats, and so on. No need to take new flats for each new object. the one set will be fine.

You should (but dont have to) take darks every imaging session, matching the exposure of your light frames. This is especially important if your camera doesnt have set point cooling, because then you have no way of knowing the temp of the chip, and each nights images may have been taken at a slightly different temp. Your starlight doesnt have set point, for the life of me I dont know why, so i suggest darks every imaging session, matching the exposure of the lights.

If you use a different telescope, you will have to do your flats again.

keeping a library of darks is OK, but the best results will come if they darks are taken on the same night as the lights. Also because the chip alters over time, so a dark library may only be valid for a few months!

but thats a different issue.

[ribuck]

Paul,

when you say take 20 flats for each channel, should they all be the same exposure lenght i.e. 20 x 1min or should i take the flats at different exposure lenthgs, 1m, 2m 3, 4m etc.

I'm assuming 20 at the same lenght exposure. What sort of exposure length should i use or do they have to match the darks ?

[Psychobilly]

flats should be realtively short exposures... as they are of a bright but uniform light source...

Billy..

[ribuck]

Ok, so they need to be short, so basically 20 x 5 sec exposure, would that do the trick.

Ok one last question about the flats, should the illumination of the flats be of varying stength from very dim -> to very bright ? or do they have to be created with a constant illuminated source ?

[Psychobilly]

Ok, so they need to be short, so basically 20 x 5 sec exposure, would that do the trick.

Ok one last question about the flats, should the illumination of the flats be of varying stength from very dim -> to very bright ? or do they have to be created with a constant illuminated source ?

They need to be whatever they need to be to get a nice histogram with the right exposure level... could even use auto exposure which would try and achieve a 50% gey level ... I am sure Paul will tell us exactly what level the exposure needs to be...

[narrowbandpaul]

i thought I posted this already...i dont see my post...

most SLR's I would think would exhibit more than 5% PRNU, so any exposure showing a histogram around the half full level is ample....

dont need to be that exact...

normally around 20*1/2000th of a second, if you are pointing at sky...

just enough to get to half full well...

hope that helps

[Psychobilly]

i thought I posted this already...i dont see my post...

most SLR's I would think would exhibit more than 5% PRNU, so any exposure showing a histogram around the half full level is ample....

dont need to be that exact...

normally around 20*1/2000th of a second, if you are pointing at sky...

just enough to get to half full well...

hope that helps

You said soemthing along the lines ...

[George]

Your dark frames also need to be taken at the same ambient temperature as your images, temperature does not matter for bias or flats.