Astrofriend's homepage


Ad / Annons:

Project News

Ad / Annons:

Dark image calibration and matching temperatures between darks and images

Dark calibration

  1. Influence from bad temperature match
  2. How can we extract temperature information from Canon EOS DSLRs?
  3. Automatic pairing process
  4. How I'm do it in practice

1: Influence from bad temperature match

Note, in modern low noise cameras it could be a good idea to replace dark calibration with dithering technique, especially the static pattern "noise" must be low.

I write this to you who are not aware that some cameras of DSLR type has a built-in temperature sensor.

To get the maximum quality from a dark calibration, among other things, the temperature at which dark frames and light images are taken has to be as similar as possible. It is not so easy when the temperature varies over night. Cameras that have built-in temperature sensor opens up possibilities for an easier handling of your dark frames, more precise calibration and use your observation time better.

Here is a link to an example of how it could look with good or bad temperature match: (Astronomy forum)

Back to contents

2: How can we extract temperature information from Canon EOS DSLRs?

Canon EOS with liveview and the old 5D (mark I) has an temperature sensor built in that can be read through Exif data, it has the tag $ E47. An easier way to get a overview what temperatures our images and darks has been taken at is to have the temperature in the filename, then you can sort them out.

A program that can rename the file name and add the temperature is IrfanView:

I talked with the developer about this and it turned out that the function is already there but he had never thought that someone could be interested to know the temperature and thus he didn't give details about it. Now I have told him that astronomers can increase the quality of their photos if they know about the temperature. He will put information about this in the future versions. But already now it is possible to insert the temperature in the filename and the only thing you need to know is the tag for the temperature, as told earlier, $ E47 (Canon).

I rename my files with batch-conversion / rename and use this code:

For dark frames:

  • Code: DA tv$E33434 iso$E34855 tc$E47 date$E36867.cr2

Filename will look like this: "DA tv60 iso800 tc-1 C date20130420 230932.cr2"

For image frames:

  • Code: LT date$E36867 tv$E33434 av$E33437 iso$E34855 tc$E47.cr2

Filename will look like this: "LT date20130406 010536 tv60 av5_6 iso800 TC1 C.cr2"

DA=dark, LT=light, av=aperature value (opening), tv=time value (exposure)

If you use the tag $ E36867 as in example above that provides date and time of the file, you must also remove the ":" character as part of the time. File names can not have this character. Use the replace ":" to some characters, such as spaces. Made during batch conversion / rename> options. It is advisable to add the date / time so that no more than one  file have the same name.

Now you have a file-list which has the temperature as a part of filename and it's easy to arrange which to use in calibration process.

Back to contents

Ad / Annons:

3: Automatic pairing process

There are programs that can manage to pair dark frames and light frame with respect to temperature, among other things.

Here is one example of software, they work together:

There are certainly many more.

Back to contents

4: How I'm do it in practice

I usually use cloudless nights only for astronomy photos. On cloudy nights, I take dark frames or during the day when the temperature is right. One should not waste the rare cloudless nights on unnecessary things. When you start a exposure sequence you will get a whole range of temperatures because the outside temperature varies under the night. The goal is to have at least 25 sub darks for every master dark temperature. Delete the first 3 to 5 darks you get from the camera because the camera inside temperature need some time to come in thermal equilibrium.

In general I run through all exposure times that I need to build a database with dark frames. Have normally the same ISO setting, but nothing prevents you to have multiple ISOs, you just get a little bit more files to manage. It shouldn't be more than 3 degrees between the temperature for every master-dark. All sub-darks in my files have the temperature +/- 1 degree i.e.

  • MasterDarks with +3 Celsius has subdarks of +2, +3, +4 Celsius
  • MasterDarks with 0 Celsius has subdarks of -1, 0, +1 Celsius
  • MasterDarks with -3 Celsius has subdarks of -4, -3, -2 Celsius

And for each temperature exposures 30, 60, 120, 240 seconds, all iso800.

Here are two examples of software that handle DSLR cameras:

I usually put it up so that the program toggles between the exposure time I want, for example: 30, 60, 120, 240seconds etc. as required. I therefore do not take the same exposure all the time until it is done and then move on to the next exposure. The advantage is that when the temperature changes, I get a better distribution of the temperature in my database.

When you do calibration the images have to divide it into several stages, each part has an acceptable temperature range. Combination master dark / DSS does it automatically, no need of masterdark manually.

The temperature sensor in the camera (Canon EOS) measure probably not the temperature of the sensor without the camera body, which will not be quite so good really, becomes a problem when the temperature changes rapidly. Those of you who have not had an eye on the camera temperature earlier will be surprised how much it varies over night, 10 degrees or more is not unusual and 7 degrees doubles the dark current!

I have an old Canon EOS 5D (MK1), which surprisingly had a temperature sensor built in. Otherwise it is the more modern Canon EOS cameras with Live View that have it. Probably also the other manufacturers of DSLR cameras has a temperature sensor in their cameras, check in exif data if the temperature is there.

Back to contents