Beginners help with Zerene. Examples posted.

[DrLazer]

Hey guys. I finally got a bellows unit, and attempted my first ever stack with a micro 4/3 camera. I am using a Lumix G10, a Nikon SB24 speedlight and a reverse mounted 50mm Minolta Rokkor 50mm. I was lucky enough to find this spider (no idea what type it is) in a decapitated crocus flower, it wasn't moving much. I brought it inside for a bit of stacking, it twitched occasionally but I think it's days are numbered.

Anyway, I run these stacks in CS5 and the results were awful. I downloaded a demo of Zerene and just used PMAX because I have no idea what I am doing. Both stacks are at 2:1. The more front facing images is stacked from 18 exposures, the other one is 17. Afew areas have a few artifacts . The less forward facing one has some strange ghosting effect going off. I'm sure it is all my fault, I could probably retouch and rescue all this in photoshop but I definately want to become more familiar with any techniques available to me in Zerene.

Thanks in advance everyone.


http://farm6.static.flickr.com/5212/552 ... f4b2_o.jpg



http://farm6.static.flickr.com/5296/552 ... fb5a_o.jpg

[rjlittlefield]

Hey guys. I finally got a bellows unit, and attempted my first ever stack with a micro 4/3 camera.

It's a great start. The problems you're running into are quite common for people new to stacking.

I am using a Lumix G10, a Nikon SB24 speedlight and a reverse mounted 50mm Minolta Rokkor 50mm.

My first perception is that the images are low contrast. The darks especially are not as dark as I would expect them to be. I suspect you're getting flare from stray light bouncing around inside the lens/bellows/camera. See the discussion HERE for tips about how to identify and cure that problem.

The less forward facing one has some strange ghosting effect going off.

The ghosting was caused by subject movement. Notice that the spider's leg slightly below left of center is obviously ghosted, while all the flower parts look fine except the bit at extreme lower left. PMax is particularly vulnerable to ghosting because it is relentless about finding and preserving detail at all scales and at all pixel positions. If a leg appears in different places in two frames, you'll see both of them.

In most cases, the cure for this problem is retouching, to replace the ghosted areas with content from just one image instead of two. It's best to do that retouching inside ZS because ZS knows how the images are aligned. The Zerene Stacker website has a couple of video tutorials on retouching, linked from the once-over-lightly description HERE.

However, those tutorials focus on how to combine PMax and DMap outputs. For the current situation that's not a good approach -- the combination of DMap, a new user, and a hairy bristly subject pretty much invites frustration.

Instead, I'll recommend a different approach: stick with just PMax and retouch from individual source frames.

In addition to the flare and ghosting, I see some focus banding. This term describes alternating fuzzy and sharp zones caused by having a focus increment that was too large compared to the DOF in each frame.

What method are you using to step focus? And what f-number are you shooting at?

--Rik

[lauriek]

A nice start indeed!

Maybe a wolf spider?

[DrLazer]

Hey Rik, Thanks for your detailed reply. A bit of a fail overall eh. Nevermind, there is only one direction to head in. I might be able to crop the best area from the stack and photoshop it to death to produce something i might keep.

Stray light. I'll go have a read of that thread. I thought that might be the case actually. All the adapter rings and stuff are chrome. No rubber or textile seals anywhere.

Thanks for the link to the tutorials. I'll certainly check them out and see if I can retouch the image. I guess if I intend to shoot living (even if barely) subjects, I should get used to having to retouch.

Is there a particular area where you can see focus banding for definate? I just want to be able to identify it for future reference. I am using a terrible method to step focus at the moment. I have the entire bellows unit, cam and lens on a cheap chinese focus rail that wobbles all over the place. Just taking shots with a 10 second timer to try and prevent that. It doesn't even move forwards in a straight line, the scale is on the left hand side and the smallest notch is 1mm, which i dont believe is fine enough for 2:1. I turn the knob a little until I think I see the next area in focus on the viewfinder and then set the timer off. I could really do with a better solution. I know I will need a linear stage and micrometer eventually, but if there is some other piece of equipment between what I have and a linear stage I would snap it up. Maybe there are more precise focus rails that are cheap?

My apeture was set to f/5.8. The Rokkor I am using has a range of f4 - f11. I just went for a wild guess as to where the sharpest f stop might be.

[DrLazer]

Maybe a wolf spider?

That crossed my mind, but it doesn't have two large eyes like most species of wolf spider do. I thought it might be a nursery spider, but then the two rows of eyes are quite far apart. Hmmmmmm

[ChrisR]

Rik has to go to bed sometime... the most obvious band of fuzz I've drawn round in green - you can see how it cuts across the spider.



At f/5.8 (!) and 2:1 your DOF is only about 0.2 mm, so your bellows will be possible, but tricky, to use.
You'll find that if you do the sums, and if you can just move the rear end of the bellows, you'll need much bigger movements, which are easier.
I did a small spreadsheet back here which was for 5x, but as it happens 0.2mm, The "rear standard" movement (first red column) was around 5mm per step.


Small lateral movements between images aren't as important as you might think - the stackers track the image pretty well, though perspective can change if you overdo it.
With flash you won't have to wait for things to settle much , at 2:1, so you don't need 10 seconds. Use a remote release though!
If you just give a bellows knob tiny nudges you might have too many steps, but it doesn't matter a lot.

[rjlittlefield]

Hey Rik, Thanks for your detailed reply. A bit of a fail overall eh.

You're welcome. I'd actually give it a "B" -- lots of room for improvement but no fundamental errors, basic technique sound.

I guess if I intend to shoot living (even if barely) subjects, I should get used to having to retouch.

That's definitely the way to bet. Things twitch.

ChrisR has offered good advice. I calculate a DOF that's even smaller than he does, closer to 0.1 mm at f/5.6. You might try stopping down to f/8. That will increase DOF by roughly 1.4X, and while there will be some softening from diffraction, it's liable to be compensated by less likelihood of softening from focus banding.

If your bellows allows moving the rear standard, then ChrisR's idea of adjusting bellows draw works great in this magnification range. One caution is that it tends to produce a fairly large change in scale from one end of the stack to the other. Be sure to frame at long extension, since that gives you the narrowest field. Then focus on the front of the subject and progressively reduce the extension to push the focus backward. Using this approach, you may also need to increase the maximum scale adjustment at Options > Preferences. For this sort of subject, you can just push that up to 100% and you should be fine.

--Rik

[ChrisR]

I calculate a DOF that's even smaller than he does, closer to 0.1 mm at f/5.6.

Alarm! - where's my abacus...

The only thing to change the calculation is the CofC, which as sensor/1000 gives:

fov 6mm, sensor 12mm, dof 0.1mm
fov 12mm, sensor 24mm, dof 0.2mm
fov 18mm, sensor 36mm, dof 0.3mm

Micro 4/3rds sensor image area: 17.3 mm × 13.0 mm (21.6 mm diagonal)

So, when you go for your bellows knob, it's the difference between a tad and a smidjin.

(Field of View
Circle of Confusion
Depth of Field)

[rjlittlefield]

The only thing to change the calculation is the CofC, which as sensor/1000 gives:

True. In honor of pixel-peeping, I've taken to using width/1500 instead of the more traditional diagonal/1500 or width/1000.

A pretty wide range of values can be justified; see for example Wikipedia's article.

--Rik

[DrLazer]

Good lord, I thought I had gained a resonable grasp of photography, then I became interested in Macro and it feels like day 1 again

@ChrisR Thankyou kindly for taking the time to point out the focus band. It is definately clear to me where the band is now. My bellows unit does have a moving rear standard! I will certainly try out moving the snesor on my next stack. Could you provide me the formula for working out DoF. I can google it easily enough, the only problem is .... I've found about 8 different formulas and have no clue which one to use. Apparently for Close Up work, yet another different forumla should be used too

Looking at your excel chart, it would seem that as magnification decreases, so does the amount that you have to move the camera forwards? Am I reading that correctly?

I plan on getting a remote release, just need to order one from the far east.

ooo, one more thing. As I am using a 50mm lens in reverse, do you have any idea how I can work out f?

@Rik, Thanks again, Ill give f/8 a whirl once I have figured oout the contrast problem. I read the thread and took the camera off the bellows to have a look through. Shining a maglite from all around the unit and the seals, I cannot find a single place I would suspect is leaking light. One thing that does concern me however, the stepdown ring used to hold the lens in place is glossy black. When I look through the bellows, the light coming through the lens is making it shine quite brightly. Do you think this could cause the contrasty problem?

Thanks for the advice with Zerene, I'll run a new stack really soon and will defo jimmy with that setting.

[ChrisR]

I thought I had gained a resonable grasp of photography, then I became interested in Macro and it feels like day 1 again

Know what you mean ! Some aspects need a closer look. That can become a narrow view - you'll have to decide for yourself where the boundaries lie between those!

DOF - start here, in the FAQ section:
http://www.photomacrography.net/forum/v ... php?t=7359

Diffraction - start here
http://www.cambridgeincolour.com/tutori ... graphy.htm
Your pixels are I think about 16µm² which doesn't appear in their table. You'll need to do some tests to see what happens. You may find say, that f/8 or 11 is about your limit as effective aperture, going narrower than which, you lose more to diffraction that you gain from DOF.

Effective aperture is marked aperture x (magnification + 1). So at 1:1 your f/5.6 is effectively f/11.

For my fatter camera I use "f twentysomething" as a limit, with the simplified calculation which seems to work:
DOF = 2 x CofC x (marked f no)x (m+1)/(m²)
For CofC, see previous posts.

Then remember your focus steps have to overlap. No harm in a few too many steps. About 3/4 of DOF seems to avoid me any trouble, though others may choose 95%.

The spreadsheet is worked from the simple "thin lens" model:
( (1/image distance) + (1/object distance) = (1/focal length) )
and magnification = image distance / object distance.

Lenses aren't thin of course, but for relative movements I think the model's good enough. (Rik??)

You can use the lens' marked "f" number whichever way round it is.

Right, off to do some plumbing. Calculator stays at home...

[rjlittlefield]

The spreadsheet is worked from the simple "thin lens" model:
( (1/image distance) + (1/object distance) = (1/focal length) )
and magnification = image distance / object distance.

Lenses aren't thin of course, but for relative movements I think the model's good enough. (Rik??)

Yes, for relative movements that model gives very good answers.

So, added_magnification = added_extension / focal_length .

The reverse of that is a good way to measure the focal length of a lens that is not marked:

focal_length = added_extension / added_magnification

These calculations are very unusual in that they stay as simple for real lenses as they are for the mythical thin ones.

You can use the lens' marked "f" number whichever way round it is.

Be cautious here. With some lenses, you will get quite different DOF at the same setting depending on whether the lens is reversed, even after adjusting the bellows to give the same magnification in both cases. See FAQ: What is "pupil ratio" and why would I care? for gory details.

But the "relative" version works in this case too. Doubling the f-number doubles the DOF and requires four times longer exposure, no matter what the lens or magnification.

--Rik

[rjlittlefield]

One thing that does concern me however, the stepdown ring used to hold the lens in place is glossy black. When I look through the bellows, the light coming through the lens is making it shine quite brightly. Do you think this could cause the contrasty problem?

Yes, it surely could. You may be able to cut some paper baffle to shadow the stepdown ring. Or perhaps shadow it indirectly, by using a conical shade on the front of the lens as shown HERE, first panel, lower section.

--Rik

[DrLazer]

I had chane to do another stack today. I used the "moving rear standard" method which seems to work a treat at 2:1. I don't think there are any obvious area where this image has focus bands.

I also used the conical black paper shade. I think the contrast has improved somewhat. I screwed the lighting up on this stack duhhhhhh. There's always something. Zerene stacked my 22 exposures on my crud laptop in just over 5 minutes. Pretty impressive.


http://farm6.static.flickr.com/5057/554 ... fed3_o.jpg

I think I need to get this remote release ordered, getting a bit tired of the timer beeps :p

[rjlittlefield]

Bingo -- I believe you've arrived! Quick study of the large version on Flickr shows no focus banding, clean and sharp throughout. Nice job.

--Rik