CCDSharp Deconvolution

[RobF (Rob)]

I can't seem to get this to work for my DSLR stacked images. I've tried converting TIFs to FITs in various ways.

Does anyone have any success with CCDSharp with DSLRs? I expect the CCD users using SBIG format will be ok?

[bojan]

No joy here either (it says "unknown error.. etc").
I think it accepts only certain size of the image? Or, it needs specific header to be recognised by CCDSharp. Only guessing...

[jjjnettie (Jeanette)]

It may only work with 16bit or 8bit images?

[cfranks (Charles)]

From the Help menu:

CCDSharp supports the SBIG compressed and uncompressed image and most 16 Bits FITS formats only. You can not read TIFF or JPEG format images with CCDSharp.

[RobF (Rob)]

I've tried 8 bit, 16 bit, and everything else I can think of.

I had a cunning plan that I could use CCDSharp to convolute the heck out of an unsaturated long exposure stack, then use this as a luminescence channel against the original RBG stack. Am I dreaming!?
(I can't get convolution working anyway.....)

[Bassnut (Fred)]

Rob

Ive used Deconvolute in CCD soft with great success, I dont know if its the same as CCD sharp, but bar none, its the single biggest improvement you can do to an image, given sufficient data for noise not to become obvious.

It deffinitely suits longer FLs though, for 4 pixel stars it wont make much difference.

I dont think the cam is the point, if its in FITs, it should work, I never use SBIG format with by SBIG, FITs only. Are you haveing trouble with the conversion to fits, or using Deconvolute?.

Yes, deconvolute is used on the Lum channel to bring out detail, not on RGB, your on the right track there. In fact, decon is often used to extreme with a layer mask and only then selectively revieled on highlights by painting on the mask.

[bojan]

Originally Posted by cfranks

From the Help menu:

CCDSharp supports the SBIG compressed and uncompressed image and most 16 Bits FITS formats only. You can not read TIFF or JPEG format images with CCDSharp.

We know that.
I tried all FITS formats (generated from DSS) and it does not work.
As I said before, it comes back with error message.

[[1ponders] (Paul)]

If you can save the fits image in DDS before DDS reapplies the Bayer ( it converts the images to monochrome fits before processing them and then reapplies the appropriate bayer filter after stacking......from memory that how I think it works) then you may have more luck. You would then have to work out how to then reapply the Bayer filter. If it can be done it may work.

[peter_4059 (Peter)]

Rob,

I have the AstraImage photoshop plug-in deconvolution filter and have a problem with DSLR images. It only works if you reduce the image size (I think below 2000 pixels in each dimension). It might be worth re-sizing the image to something smaller and giving it a try?

Peter

[bojan]

I think it may have something to with the algorithm itself.. and how it is implemented.
Iris can do it, but only on cropped picture (and it has to be square, x=y).

[RobF (Rob)]

Hmm - couple of tempting ideas to try there. Don't really want to crop or resize, but will be interesting to see if it works and if deconv helps.

Fred, it sounds like deconv isn't just for CCDs then? There's no law against doing a severe deconv and then stacking in as a lum channel? (I always thought L was something you only needed when doing filter wheel separate R, G, B exposures, but looking at some articles on it the other day got me thinking there's no reason it couldn't be done in Photoshop. I've had a bit of a play making a layer copy, desaturating, applying heavy unsharp mask, then layering under original RGB as a lumin channel. The sky hasn't fallen in as yet.....

BTW, is there anywhere on the net people recommend to read up on these more advanced imaging tricks?

R