shooting dark and flat frames for stacking

[Carl]

Hi folks
deepsky stacker allows for stacking of dark frames. I have a canon350eosdslr how do i shoot dark frames. I can find no reference in my user guide

Regards
Carl

[Striker (Tony)]

HI Carl,

Dark frames are done by placing the camera cap on not allowing any light getting to the sensor..these are done for the same length exposure as your images...for example 5minute exposure are followed up with 5minute dark frames.

[[1ponders] (Paul)]

Hi Carl, ditto to Tony's info, plus for your flat frames you image a flat light coloured evenly illuminated wall or similar. These should be long enought to show any vignetting. I find if I select my ISO and then allow the camera to automatically adjust for exposure etc they turn out ok.

The really really really important thing here is, you have to leave your camera attached to your telescope to take flats. Don't move it, rotate it, adjust focus, anything. When you take a flat you are trying to get an image of your systems optical "faults" like vignetting, dust on your chip, the "shape" of optical light path etc and this will change with changes in camera orientation and focus. So don't adjust anything. I usually leave taking my flats til the last thing when I take my scope inside. I set it up on a table, pointing at a wall about 600mm away from the scope. The wall needs to be very evenly illumninated otherwise you will get false flats. Just one light coming down from above will give a poor flat. There is a good article by Eddie_T in the Resources section on the how and why of taking flats.

When it comes to taking darks, as long as your temperature hasn't fluctuated more than 2 or 3 degrees during a session you can take your darks with the camera off the scope and just the lens and cap on when you are tidying up. BUT only do this if you have already taken your flats and the temperature hasn't changed, otherwise take them during your imaging session after your lights as Tony suggests with a cover over the end of your scope.

[5ash (Philip)]

Hi carl ,
I have problems understanding the technical language used in Eddie_Ts article . Can anyone interpret the following qoute from the article in a way that i can step by step take a flat to achieve the quoted parameters "The main aim is to get a flat field that has an average pixel value of about 30% of the maximum pixel value that your camera is capable of. For a 16-bit camera, this would be approximately 20,000. Most image processing software will give you the average pixel value in an image so this value is generally easily obtained."
philip

[[1ponders] (Paul)]

What I think he is saying is that when you take your flat you want to have the average pixel well one third full. In other words don't overexpose your flats so they become saturated. That is why I use the autoexposure built in, it seems to get close the the required value automatically.

[avandonk]

Phillip the easiest way is to set the camera to Av (aperture priority) and let the camera take care of it. The best even light source is the twilight sky at about the zenith before any stars are visible and after the sun is below horizon.

Bert

[Striker (Tony)]

Here is what is what a common Flat frame looks like...notice the dust donuts and the difference in light in the frame...all this will be subtracted from your image in calibration.

[Dennis]

Originally Posted by 5ash

Hi carl ,
I have problems understanding the technical language used in Eddie_Ts article . Can anyone interpret the following qoute from the article in a way that i can step by step take a flat to achieve the quoted parameters "The main aim is to get a flat field that has an average pixel value of about 30% of the maximum pixel value that your camera is capable of. For a 16-bit camera, this would be approximately 20,000. Most image processing software will give you the average pixel value in an image so this value is generally easily obtained."
philip

Hi Philip

The ccd chip is made up of an array (rows/columns) of charge wells that accumulate photons. A good analogy is to think of a paddock, filled with rows of buckets that collect droplets of rain. Eddie is saying that a single 16-bit bucket can hold 2 to the power 16 rain drops before it becomes full and overflows into adjacent buckets.

2 to the power of 16 is 2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2= 65,536. That is, the depth of the charge well on a 16-bit camera is 65,536 photons. If you continue to expose the ccd after this saturation level has been reached, it cannot register any more photons in the saturated regions and the corresponding pixels on the screen become totally whited out with no detail in them - a white blob.

So, if you set the exposure time to approx 30% of the time required to saturate the ccd, then the charge well would only accumulate 30% of 65536, which is approximately the 20,000 that Eddie is describing.

My ccd camera, an SBIG ST7E displays the photon count of the charge well (pixel) where the greatest number of photons have been registered. Typically this happens in the centre of very bright stars or in the core of bright galaxies.

Using the camera control software, called CCDSoft, I can hover my mouse over any pixel in the image on my screen and the x-y coordinates of that pixel and the number of photons captured in there will be displayed.

Typically, DSLR’s do not have the same control software so I’m not sure how you can measure the number of photons in a pixel in real time, other than to take an image and then open that image in a photo application that allows you to see the maximum number of photons registered.

Hope that makes it a little clearer.

Cheers

Dennis

[RB (Andrew)]

Originally Posted by Dennis

Typically, DSLR’s do not have the same control software so I’m not sure how you can measure the number of photons in a pixel in real time, other than to take an image and then open that image in a photo application that allows you to see the maximum number of photons registered.

Here's where the DSLR's in-camera histogram can help.
Just check it after you take the flat and see if it's correct.

[[1ponders] (Paul)]

Andrew, you wouldn't happen to have an example handy of what the histogram should look like would you

[Striker (Tony)]

Paul,

Just select push the info button on your camera and it will bring up the histogram....you want a nice sharp peek approx 1/3 to 1/2 way along the histogram.
If your looking at pixel count it goes from 0 = Black to 65000 = white some where between 20000-30000 is good place to start.

[[1ponders] (Paul)]

Thanks Tony, I was thinking more of showing Carl what the histogram might looks like, but your description covers that.