Hi All
Well here we have my attempt at M101 following the M8 tutorial.
As you can see there is a lot of noise giving the whole thing a red cast and looks a bit like an HDR photo.
I am beginning to think I should use the unmodded 600D on galaxies and leave the modded 450D for the widefield nebula only. Any thoughts ?
This was taken with a modded Canon 450D (with CLS CCD clip filter) 11 x 5 min exposures at ISO 1600 stacked in DSS and output as 32 bit FITS Integer.
Mount is HEQ5 Pro - Skywatcher 200PDS scope - Guided with QHY5 on 70mm refractor - No moon - milky way faintly seen with Mk1 eyeball.
Had a go with adjusting the TIF file in lightroom and it did make things better but I should be able to get the same result in Startools I would think. File as JPG below.
Could someone have a go at my data and see if they can get a better result and explaining what steps were taken etc.
I have put it here. https://www.dropbox.com/s/11sfq1353oz2h ... new_32.FTS
Looking forward to see what others can do or if my data is just rubbish
Thanks in advance
Steve
M101 noise problem
Re: M101 noise problem
Gee, Steve. This is a tough one!
Here's what I did using the latest beta.
As always, I do an AutoDev to see what we got. The data seems to be pre-color balanced. It appears to be very, very noisy with the spiral arms just sitting above the noise floor. This is going to be a tough one! I further see a yellow bias and noise grain that looks like its coarser than 1 pixel - a sign that this data is from a DSLR and the noise was introduced by the debayering stage.
I bin to 50% to reduce noise and at the very least reduce noise grain to 1 pixel.
I also crop the image to frame M101 a bit better.
Next I do a Wipe, Dark Anomaly Filter at 5px (to be absolutely sure no anomalous dark pixels interfere with true background detection). The 'Temporary Autodev' setting lets us see what we're doing. Considering the apparent lack of flats, the result doesn't look to bad.
I choose to do a manual Develop (I don't usually), as I really want to keep that noise floor under control. I could've use an AutoDev with an ROI and turning Outside ROI way down. AutoDev can struggle a little with sever noise like this - it just wants to bring it out. [Digital Development] set to [87.08 %]
Not expecting too much in the face of so much noise, I try my luck at Decon, I choose a radius of 1.4px, Parameter [Regularization] set to [1.15 (smoother, less detail)].
I use the HDR 'Reveal' preset and set 'Detail Size Range' to 32px to make it a bit more aggressive.
I perform some wavelet sharpening, setting Amount 200%, setting scale 1 to 0% (as to not exacerbate any fine noise).
In the Color module I let it do its thing, and set Cap Green to Yellow. The muted colors are typical of an image that was shot with a CLS filter in place. Specific hues spedtral lines were irreversibly removed from the red and green channels and no amount of rebalancing won't bring them back.
In a last attempt to push back noise a little, I use the Life module's Isolate preset. I overdrive the Saturation by 200% and set Strength to 50%.
Finally I stop Tracking and perform final noise reduction;
Parameter [Color Detail Loss] set to [6 %]
Parameter [Brightness Detail Loss] set to [20 %]
Parameter [Redistribution Kernel] set to [5.6 pixels]
Parameter [Read Noise Compensation] set to [33.54 %]
Parameter [Smoothness] set to [51 %]
And this is what I ended up with;
Here's what I did using the latest beta.
As always, I do an AutoDev to see what we got. The data seems to be pre-color balanced. It appears to be very, very noisy with the spiral arms just sitting above the noise floor. This is going to be a tough one! I further see a yellow bias and noise grain that looks like its coarser than 1 pixel - a sign that this data is from a DSLR and the noise was introduced by the debayering stage.
I bin to 50% to reduce noise and at the very least reduce noise grain to 1 pixel.
I also crop the image to frame M101 a bit better.
Next I do a Wipe, Dark Anomaly Filter at 5px (to be absolutely sure no anomalous dark pixels interfere with true background detection). The 'Temporary Autodev' setting lets us see what we're doing. Considering the apparent lack of flats, the result doesn't look to bad.
I choose to do a manual Develop (I don't usually), as I really want to keep that noise floor under control. I could've use an AutoDev with an ROI and turning Outside ROI way down. AutoDev can struggle a little with sever noise like this - it just wants to bring it out. [Digital Development] set to [87.08 %]
Not expecting too much in the face of so much noise, I try my luck at Decon, I choose a radius of 1.4px, Parameter [Regularization] set to [1.15 (smoother, less detail)].
I use the HDR 'Reveal' preset and set 'Detail Size Range' to 32px to make it a bit more aggressive.
I perform some wavelet sharpening, setting Amount 200%, setting scale 1 to 0% (as to not exacerbate any fine noise).
In the Color module I let it do its thing, and set Cap Green to Yellow. The muted colors are typical of an image that was shot with a CLS filter in place. Specific hues spedtral lines were irreversibly removed from the red and green channels and no amount of rebalancing won't bring them back.
In a last attempt to push back noise a little, I use the Life module's Isolate preset. I overdrive the Saturation by 200% and set Strength to 50%.
Finally I stop Tracking and perform final noise reduction;
Parameter [Color Detail Loss] set to [6 %]
Parameter [Brightness Detail Loss] set to [20 %]
Parameter [Redistribution Kernel] set to [5.6 pixels]
Parameter [Read Noise Compensation] set to [33.54 %]
Parameter [Smoothness] set to [51 %]
And this is what I ended up with;
Ivo Jager
StarTools creator and astronomy enthusiast
StarTools creator and astronomy enthusiast
Re: M101 noise problem
Thanks for the quick reply Ivo. What an inspiration to other software developers.
As I said in the intro I will keep the unmodded 600D for galaxies in the future and leave the modded 400D for nebula work.
I used the life module, for the shot I got over the weekend below, which caught a bit of IC 434 using the modded 400D.
I think this has come out quite well. Still waiting for the field flattener though as can be seen from the stars all round the edge.
Regards
Steve
As I said in the intro I will keep the unmodded 600D for galaxies in the future and leave the modded 400D for nebula work.
I used the life module, for the shot I got over the weekend below, which caught a bit of IC 434 using the modded 400D.
I think this has come out quite well. Still waiting for the field flattener though as can be seen from the stars all round the edge.
Regards
Steve
Re: M101 noise problem
If it helps, I think the dark specks in the image look very much the result of mismatched dark frames. Ivo has expertly dealt with those. This type of artefact can be dealt with by a combination of better temperature matching of darks and dithering. DSLR image calibration is more subject to temperature differences between frames than a cooled ccd, for instance.