DSLR: Save CR2 or FITS?


Hi, when using a (Canon 550D) DSLR, is there any advantage in saving in FITS instead of CR2? (the native Canon Raw format). I know FITS has the advantage of saving all settings in the header, but as far as processing/image quality is there any advantage/disadvantage?..Or maybe it’s best to save in BOTH formats?




Most programs have to use DirectDraw to convert the Canon/Nikon format to something they can use. I’d say it doesn’t really matter when, but at some point that process has to happen. It might as well happen in SGP.



I tried to save in Fit format once and I had huge issues trying to calibrate my lights. The files were randomly rotated. I had to reshoot my calibration frames with SGP and save them in Fit format. Since that time I always save in CR2. I use Pixinsight for everything maybe PI didn’t know how to handle the Fit files coming from SGP.
I never had any calibration issues with CR2 files off of SGP.


There might be slight inconsistencies between the 2 files so you should not use one type of file for calibration and then the other for lights… that will almost always get you in trouble. Most DLSR folks say they have better luck with RAW, but I have also seen fine results with DSLRs and FITS.

That’s odd… I have not seen this before. Not sure how it could happen either. In order to make the FITS file we use the CR2 as the “source”. Not saying it can’t happen, just that I have never seen it.


Thanks Andy and Ken for your replies. It therefore seems safer to save in CR2 format. It’s a pity because the FITS header is very useful…

I also use Pixinsight so I will do a test to see how it handles FITS files from SGP.

Thanks again!


I use SGPro, a Canon 60D and Pixinsight. All my files (bias, darks, flats, lights) are in FITS format. Works perfectly.

The only time I had an issue was when I saved my darks or flats or bias as CR2 and converted in Pixinsight to FITs (instead of SGPro) and generated my lights in SGPro. The two programs have a slightly different treatment of overscan, so the two set of files have different pixel dimensions. This causes “incompatible” error messages which can be fixed by careful cropping.




Scott, thanks for your input. Is there any image quality improvement in
using FITS instead of CR2? (When processing in Pixinsight).


El 30/11/2015 13:10, “scott_astroman” forums@mainsequencesoftware.com


I originally had my Nikons files saved as NEF only, but then the plate solving was not working because it was expecting a fit file and reported it cannot find it. Am I correct that if you want to use plate solving you have to have it set with fit files as well? So RAW+Fit or Just Fit.


Actually, I use CR2 files for plate solving and it *normally *works,
although frequently It has to blind solve with astrometry.net because
PinPoint cannot solve it…Maybe it is because it is not a FITS file? Any




I was getting the error when trying to use the “center” to feature. It was reporting that the “file.fit” was not found, however in the directory was a .NEF.


Hi. I’ve had this problem but making sure the hint “raw cfa” was used in PI solved it.


Hi, I cannot find the option you mention (raw cfa). By “PI” do you refer to Pixinsight or did you mean PinPoint? Thanks.




I mean Pixinsight - in the image calibration and integration processes you will see a drop down for “Format Hints” - add the input hint “raw cfa”. This tells PI that the files are raw DSLR files.


Just doing some calibration tests with my new Nikon D5300 and am definitely seeing that not having Lights and Calibration files (Bias/Flats) taken from the same source are causing calibration problems.

I had tried calibrating SGP Light FIT files with RAW Bias/Flats and it was not working out. Re-taking the calibration files in SGP seemed to solve this. I was trying to avoid using RAW Files but moving forward but I have to decide which way to go.


yes, this is because SGP is scaling the 14-bit DSLR data to 16 bits when saving the fits files.

i asked as a feature request that SGP not do this but it was rejected. from this thread it looks like PI calls DCRAW on NEF/CR2 with different overscan handling parameters than SGP does, so it may be a moot point - even if SGP did not scale, the geometry would be incompatible with PI.

still, on first principles i think an image acquisition program should not scale the data coming off the camera.

i also think there might be some funny stuff when the data becomes saturated (meaning saturated values don’t show up as 0xFFFF in the scaled values) but i have not yet hooked up my 50D to SGP to test that out.



With both Nikon D600 and Canon 6D I have always used strictly FIT files for all image and calibration frames, then process with PI. Works perfectly, and will save you a lot of grief.