Help with Wipe, Develop

[N_DD]

Hi,

I am having a bit of troubles removing vignetting from my images: I am using a modded Canon 1000D, and before stacking (DSS) I am converting the subs, flats and bias to linear TIFF by using DCRAW. However, when I try to remove vignetting I end up either with either too few removal, or if I increase the strength some strange circular artifacts pop up (concentric to the image). Gradient removal will only introduce some "butterfly shaped" gradient... I know my data is not so deep (just 26x4-min exposures from a red zone using a C8 @f/6.3) but am I doing anything wrong? Moreover, if I let develop "homing in" I end up usually with a lot of noise, which can't be addressed by the final noise reduction step: I suppose this depends on the quality of my data, but maybe some guru could give me a few hints!

For who would like to check my data, here is the stacked TIFF
https://www.dropbox.com/s/uhqa0p8vpohi2 ... 3.TIF?dl=0

while here is my best attempt at processing with Startools
http://www.astrobin.com/full/177256/0/

Thanks in advance!

Nico

[Rowland]

Given what you have you've done a pretty good job. The uncropped image looks like the flats didn't work. You didn't mention darks? Personally, I would retain the CR2 format, but I'm not sure whether DSS requires that for images to be imported into ST.

Not some guru!

[admin]

Seems like Rowland is right - the flats don't seem like they worked and may have actually introduced problems.

The concentric rings can be ameliorated somewhat by increasing precision. This is typically only necessary when dealing with severe undulating gradients or if the remaining data (after Wipe) is low-bit depth.

Also, what algorithm did you use to stack if I may ask? Median, Average, etc.?

[N_DD]

Hi,

thanks a lot for the feedback!

Rowland: I tried stacking the original CR2 files, but it seems that DSS applies some kind of color-balance to the results, making it very difficult to process the stacked image (there was a thread here on the forum some time ago, and Ivo suggested to convert CR2 to TIFF using DCRAW before stacking them in DSS). I am not using darks, because I find it too difficult to precisely temperature match them using a non-cooled DSLR. Instead, I am dithering between subs.

Ivo: I will try to troubleshoot my flats. I acquired them using the "t-shirt" method, and they seems ok. My camera is 12-bit (Canon 1000D), might this explain in part what am I seeing? I was using median stacking on dithered subs (around 10 pixel dithering) to remove hot and cold pixels.

Thanks!

Nico

[admin]

My camera is 12-bit (Canon 1000D), might this explain in part what am I seeing? I was using median stacking on dithered subs (around 10 pixel dithering) to remove hot and cold pixels.

Ah, good to know! I suspect you'll be able to get slightly more useful data if you stack using averaging. Good thing you're dithering too! It should have helped with the vertical bias and read noise lines.

Stacking with median will always retain the bit-depth (12-bit in your case). If a lot of those 12-bits are taken up by light pollution or some other bias, you will not have much to work with once that part of the dynamic range has been removed. Let us know how you go!

[Rowland]

I remember now - tiff.

Take darks any way. Keep imaging at the same rate and time as your lights. Dithering is not really a substitute for dark noise reduction and without dark reduction you will probably find that images are more difficult to process. 10 darks are better than nothing. And dithering will take care of calibration artifacts due to temperature mismatches. As I see it, the problem with DSS is it's inflexible approach to reduction.

I had a look at Regim today. It does a pretty good job of calibration. Couldn't get images to align, but ideal for DSLR reduction. It lets you create a master flat and master dark separately - just substitute bias frames for flat darks when calibrating flat frames. Should be possible to import into DSS.

[N_DD]

Hi!

Ivo: thanks for your suggestions! I will try averaging and see if I can improve the results. I'll let you know.

Rowland: I still have my dark library around and can see if including darks in my workflow will improve the final stacked image. I was also playing around with regim, but I could also not manage to align my subs. I'll try again.

Thanks for the comments, I really appreciate the way more expert people help beginners on this forum!

Nico

[Rowland]

I like the flexibility of Regim for master dark and flat creation. Save as tiff and use in DSS to calibrate, align and stack the images. That way you can avoid the double bias subtraction thing, which I believe DSS does when loading all your files at once.
I didn't like calibration in Regim, but that may be a settings issue and will improve with experience.

[N_DD]

Hi,

it seems that my flats were not checked in DSS . Wipe did a much better job if flats were taken into consideration.
Using both flats and darks and averaging stacking leaves unfortunately traces of the hot pixels in the final image. Switching to kappa-sigma improves the outliers rejection, but as Ivo suggested averaging seems to give the best overall image quality.
@Rowland: I definitely need to play more with regim. I like the automatic color balance (matching of stars color temperature from a catalogue) but I still can't align my subs. Have you had any success?

Thanks,

Nico

[ChrisLX200]

Hi,

it seems that my flats were not checked in DSS . Wipe did a much better job if flats were taken into consideration.
Using both flats and darks and averaging stacking leaves unfortunately traces of the hot pixels in the final image. Switching to kappa-sigma improves the outliers rejection, but as Ivo suggested averaging seems to give the best overall image quality.
@Rowland: I definitely need to play more with regim. I like the automatic color balance (matching of stars color temperature from a catalogue) but I still can't align my subs. Have you had any success?

Thanks,

Nico

Under stacking preferences (in DSS) , click on the 'Cosmetic' tab and check out the 'Remove hot and cold pixels' function. You have a couple of sliders which adjust the aggressiveness and also a ''Test on First Sub' so you can see how well the settings work (it counts the numbers of hot/cold pixels removed). The method is not as good as using well calibrated darks but it can remove a lot the remaining hot pixels. With a Sony sensor working at -10deg I find it's all I need so don't use darks.

ChrisH