Hi,
yes I think the preview issue is not in Guy's notes, but the detailed description is Guy's work!
Regrarding the stars- that seems to be an aesthetic/artistic issue. In a way as Astrophotographers we probably had been spoiled by all the shortcomings of the gear used in the past. We are so used to the "old school" look we assume its actually natural.
Decon tries to remove any dispersion caused by athmosphere and/or optical train. The PSF is called that way, because the star in fact is a point spreaded over multiple pixels. There is no "natural"fade to a point. It's just that we got used to see stars like that.
If You were at ISS or expose Your image using HSS or the like, You would see the stars just as airy disks only limited by the optics. Now SVD is trying to "undo" the way a point has been spread over many pixels,(point spread function = PSF), it offers to undo as many PSFs as needed, varying throughout the image.
Whether or not You like the "unnatural" point look is a matter of taste, just as with color: white color in blown out areas (e.g. M42 center) seems "natural to us, but is simply plain wrong, we just got used to that, by looking at thousands of white-core M42 images, so we believe it is white by nature. Similar is valid e.g. for purple vs red H-II regions.
ST's approach is to enable scientificly "correct" representation as far as possible, but ST does not nail You on that. There are always options to create images in "old-school" look e.g. FilmDev as stretching tool, or "Artistic" mode in Color.
Clear Skies,
Jochen
Version 1.8 HDR and SV Decon
Re: Version 1.8 HDR and SV Decon
Jochen,
Thanks for your reply
In regard to selecting a Star sample in the preview area, my preview area is obviously located within the main structure of the object ( nebulosity, gas, dust etc… ) so I can access the effect of SV Decon before and after, just adjacent to the preview area, however isn’t one of the prerequisites of selection of Star samples to avoid selecting Star samples within nebulosity areas. This is a slight contradiction in terms ??
Maybe one sample in the preview area and nine outside the preview area doesn’t affect the algorithm too much , I don’t know ??
Await your advice
Cheers
Clear Skies
Martin
Thanks for your reply
In regard to selecting a Star sample in the preview area, my preview area is obviously located within the main structure of the object ( nebulosity, gas, dust etc… ) so I can access the effect of SV Decon before and after, just adjacent to the preview area, however isn’t one of the prerequisites of selection of Star samples to avoid selecting Star samples within nebulosity areas. This is a slight contradiction in terms ??
Maybe one sample in the preview area and nine outside the preview area doesn’t affect the algorithm too much , I don’t know ??
Await your advice
Cheers
Clear Skies
Martin
Re: Version 1.8 HDR and SV Decon
Hi Martin,
well it depends on the subject and Your computer. I stopped using preview altogether. Most PSF distortion changes are towards the corner anyway and the subject is usually centered. Instead, to save rendering time I start off using less iterations (5-7) and no centroid tracking. If Your computer is not to outdated and has some middle of the road GPU at least,
I start selecting green samples, at least in each corner, in the middle of the image borders and towards the center, surrounding the object. Each selection launches rendering from scratch, so You can just go ahead, select all samples. Once You are confident You covered all areas using like 10-20 stars, You can wait for the rendering to complete. Switch to result view and if You like what You see, just switch on centroid tracking and increase iteration until no further improvement or to taste (I use 10-13, You'll probably be fine with less). Should You like the ST7 Decon result better, try the Synthetic PSF section. Don't select sampling stars to enable the atmospheric models.
If You choose to go the preview route, I'd recommend to select at least 2-3 stars within it. You can and should add more outside the preview, but those wouldn't affect the result within it, instead ensure even in the corners a slightly different PSF would also be reversed with a result comparable to the one in the preview area.
Cheers,
Jochen
well it depends on the subject and Your computer. I stopped using preview altogether. Most PSF distortion changes are towards the corner anyway and the subject is usually centered. Instead, to save rendering time I start off using less iterations (5-7) and no centroid tracking. If Your computer is not to outdated and has some middle of the road GPU at least,
I start selecting green samples, at least in each corner, in the middle of the image borders and towards the center, surrounding the object. Each selection launches rendering from scratch, so You can just go ahead, select all samples. Once You are confident You covered all areas using like 10-20 stars, You can wait for the rendering to complete. Switch to result view and if You like what You see, just switch on centroid tracking and increase iteration until no further improvement or to taste (I use 10-13, You'll probably be fine with less). Should You like the ST7 Decon result better, try the Synthetic PSF section. Don't select sampling stars to enable the atmospheric models.
If You choose to go the preview route, I'd recommend to select at least 2-3 stars within it. You can and should add more outside the preview, but those wouldn't affect the result within it, instead ensure even in the corners a slightly different PSF would also be reversed with a result comparable to the one in the preview area.
Cheers,
Jochen
Re: Version 1.8 HDR and SV Decon
Thanks Again !!
Post closed from my end
Martin
Post closed from my end
Martin