"Sparkles" in Zerene Pmax
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
"Sparkles" in Zerene Pmax
Lately, I've noticed glittering washed out areas in my Pmax stacks.
These don't appear in the unstacked images.
I told it not to preserve dynamic range and it got a little better, but it's still unacceptable. I changed the threshold for Dmap from 20 to 40 and the issue is much less that way. However I prefer Pmax because I get fewer artifacts.
Am I missing a setting somewhere?
These don't appear in the unstacked images.
I told it not to preserve dynamic range and it got a little better, but it's still unacceptable. I changed the threshold for Dmap from 20 to 40 and the issue is much less that way. However I prefer Pmax because I get fewer artifacts.
Am I missing a setting somewhere?
- rjlittlefield
- Site Admin
- Posts: 23625
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
There is no setting for PMax that will make the sparkles go away.
Typically sparkles occur when you have relatively "hard" illumination that produces very bright specular highlights. Those highlights actually spread out and move around a little from one frame to the next. In each different position they look like focused detail. PMax is relentless about preserving focused detail, so it ends up producing a result that is an amalgam of all those different highlights from multiple frames.
The most powerful attack on sparkles is to use more diffused "soft" illumination. See Reflections of hard and soft light in a spider's eye for illustration and discussion.
As you've noticed, DMap has much less tendency to produce sparkles. That's because DMap is not as relentless about preserving focused detail, and unfortunately that goes hand-in-hand with producing more artifacts when you have complicated geometry with lots of foreground/background overlaps.
Often the best final result is produced by using the retouching tool to merge the best bits of DMap and PMax outputs. That way you get to choose which is the better rendition for each part of your subject.
By the way, the percentage number shown for the DMap threshold slider is just the fraction of frame area that is covered by the corresponding mask. The best setting can vary from 0 to almost 100%, depending largely on composition. If there's a clear distinction between focused subject and unfocused surroundings, the best setting is usually whatever makes the mask cover all the unfocused areas. But there are special cases where other settings can produce more useful results, especially if you're retouching. For more information, see the tutorials How To Use DMap and
Advanced Retouching .
--Rik
Typically sparkles occur when you have relatively "hard" illumination that produces very bright specular highlights. Those highlights actually spread out and move around a little from one frame to the next. In each different position they look like focused detail. PMax is relentless about preserving focused detail, so it ends up producing a result that is an amalgam of all those different highlights from multiple frames.
The most powerful attack on sparkles is to use more diffused "soft" illumination. See Reflections of hard and soft light in a spider's eye for illustration and discussion.
As you've noticed, DMap has much less tendency to produce sparkles. That's because DMap is not as relentless about preserving focused detail, and unfortunately that goes hand-in-hand with producing more artifacts when you have complicated geometry with lots of foreground/background overlaps.
Often the best final result is produced by using the retouching tool to merge the best bits of DMap and PMax outputs. That way you get to choose which is the better rendition for each part of your subject.
By the way, the percentage number shown for the DMap threshold slider is just the fraction of frame area that is covered by the corresponding mask. The best setting can vary from 0 to almost 100%, depending largely on composition. If there's a clear distinction between focused subject and unfocused surroundings, the best setting is usually whatever makes the mask cover all the unfocused areas. But there are special cases where other settings can produce more useful results, especially if you're retouching. For more information, see the tutorials How To Use DMap and
Advanced Retouching .
--Rik
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
Thanks.rjlittlefield wrote:There is no setting for PMax that will make the sparkles go away.
Typically sparkles occur when you have relatively "hard" illumination that produces very bright specular highlights. Those highlights actually spread out and move around a little from one frame to the next. In each different position they look like focused detail. PMax is relentless about preserving focused detail, so it ends up producing a result that is an amalgam of all those different highlights from multiple frames.
The most powerful attack on sparkles is to use more diffused "soft" illumination. See Reflections of hard and soft light in a spider's eye for illustration and discussion.
As you've noticed, DMap has much less tendency to produce sparkles. That's because DMap is not as relentless about preserving focused detail, and unfortunately that goes hand-in-hand with producing more artifacts when you have complicated geometry with lots of foreground/background overlaps.
Often the best final result is produced by using the retouching tool to merge the best bits of DMap and PMax outputs. That way you get to choose which is the better rendition for each part of your subject.
By the way, the percentage number shown for the DMap threshold slider is just the fraction of frame area that is covered by the corresponding mask. The best setting can vary from 0 to almost 100%, depending largely on composition. If there's a clear distinction between focused subject and unfocused surroundings, the best setting is usually whatever makes the mask cover all the unfocused areas. But there are special cases where other settings can produce more useful results, especially if you're retouching. For more information, see the tutorials How To Use DMap and
Advanced Retouching .
--Rik
Here's the final result I got with Dmap before I had to go to work:
[/img]
This is Pmax retouched with DMap??
Diffuse diffuse diffuse.
(+ make sure lens is clean, & without any filter.)
Look very carefully to see if a point's blur-sparkle goes from (say) North-South to East-West when you focus through it. That longitudinal variation in astigmatism is annoyingly common. Try to diffuse until there isn't any of that sort of blur, or it all looks like detail, to the stacker, and you end up with a sort of cross. Hard to tell at this scale.
Diffuse diffuse diffuse.
(+ make sure lens is clean, & without any filter.)
Look very carefully to see if a point's blur-sparkle goes from (say) North-South to East-West when you focus through it. That longitudinal variation in astigmatism is annoyingly common. Try to diffuse until there isn't any of that sort of blur, or it all looks like detail, to the stacker, and you end up with a sort of cross. Hard to tell at this scale.
Chris R
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
No, it's straight Dmap with a few tweaks in settings.ChrisR wrote:This is Pmax retouched with DMap??
This image was taken before I started this thread.
Since I'm still working on my new interim macro rig, I was using the Jansjos instead of flash.
I had the Jansjos relatively close to the subject (a 3/8" to 1/4" ballhead adapter). They have pingpong ball diffusers on them. One was positioned to light from above and the other from front left.
About how far away should they be for something that size?
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
I'll have to make something, time permitting.Lou Jost wrote:Diffusers on the lamps don't do much good; they will take up only a small proportion of the "sky" as seen by the subject. You need diffusers close to the subject, so the whole "sky" looks evenly lit from the subject's point of view. .
My regular light tent for still lifes is too big.
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
An interesting visual demonstration of what others have posted here. I think I have in fact seen the positional shift illusion recently while shooting similar reflective subjects, with insufficient diffusion.
http://www.stackrail.info/index.php/en/ ... ht-or-else
http://www.stackrail.info/index.php/en/ ... ht-or-else
eBay has some small inexpensive tents, I have that's only a 1' cube. It can easily fit inside another larger one. I've used this double diffused tent often, and also found that a small foam cup over the lens end inside a small tent works well of some subjects.Deanimator wrote:I'll have to make something, time permitting.Lou Jost wrote:Diffusers on the lamps don't do much good; they will take up only a small proportion of the "sky" as seen by the subject. You need diffusers close to the subject, so the whole "sky" looks evenly lit from the subject's point of view. .
My regular light tent for still lifes is too big.
With metallic or shinny subjects as Chris said "Diffuse, diffuse...diffuse".
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
I've gotten very good at making these things, so I'll probably build something from PVC and packing foam tomorrow.mawyatt wrote:eBay has some small inexpensive tents, I have that's only a 1' cube. It can easily fit inside another larger one. I've used this double diffused tent often, and also found that a small foam cup over the lens end inside a small tent works well of some subjects.
That does appear to be the case. I didn't notice it until I started shooting small chrome plated items.mawyatt wrote:With metallic or shinny subjects as Chris said "Diffuse, diffuse...diffuse".
Think I paid about $10 for the small 1' cube tent.Deanimator wrote:I've gotten very good at making these things, so I'll probably build something from PVC and packing foam tomorrow.mawyatt wrote:eBay has some small inexpensive tents, I have that's only a 1' cube. It can easily fit inside another larger one. I've used this double diffused tent often, and also found that a small foam cup over the lens end inside a small tent works well of some subjects.
That does appear to be the case. I didn't notice it until I started shooting small chrome plated items.mawyatt wrote:With metallic or shinny subjects as Chris said "Diffuse, diffuse...diffuse".
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
I've got everything I need (for several very different configurations) just lying around my apartment.mawyatt wrote:Think I paid about $10 for the small 1' cube tent.
I should have something by dinner time tomorrow. I'd have something by lunch time, but I just got home from work and dinner and it's 4:15am.
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
Simple paper draped near the subject, is usually fine. Ideas here: http://www.photomacrography.net/forum/v ... hp?t=25935
Chris R
-
- Posts: 870
- Joined: Tue Oct 23, 2012 7:01 pm
- Location: North Olmsted, Ohio, U.S.A.
So far the card stock seems to work relatively well. It has the advantage of being stiff enough to have its own structural strength.ChrisR wrote:Simple paper draped near the subject, is usually fine. Ideas here: http://www.photomacrography.net/forum/v ... hp?t=25935
I'm doing a 103 image stack of a silver cufflink right now. I pulled the lamps back a bit. We'll see how that comes out.
Update: That one didn't come out well due to weird reflections (the camera) in the subject.
I just made a much smaller diffuser which sits right over the subject.