Some trouble with stacking artifacts...

etalon · Post by **etalon** » Tue Jan 05, 2016 8:17 pm

Hello to all,

hope I´m in the right board...

Since yesterday I own a beautiful Mitutoyo Plan Apo 10x/0.28 which I use together with a Raynox DCR150 as tube lens and my Canon EOS5D mkII on my micro rig. Now I could present the very first results with that combo:

Pyromorphite, 200 layer stack, 0.005mm steps, Helicon Focus (B/C):

Arisite, 275 layer stack, 0.005 mm steps, Zerene Stacker PMax/DMap:

The FOV is always ca.3.5mm.

Till now, I process every pic in both programs (Zerene and Helicon) with all algorithm provided and process the final image with only the best parts of each image created by the stacking software in PS CC2015. My experience till now is, that mostly Zerene would give the best results, especially when needle like crystals are in the picture. But on crystals with "big" surfaces (like the Pyromorphite), normally Helicon would bring the better results. But, and thats my biggest problem at the moment, on that picture from Uranophane:

(Uranophane, 300 layer stack, 0.005mm steps, Zerene PMax and DMap, FOV ca. 3.5mm)

I got only s**t with all software I tried. The best results would provide Zerene, but it looks only good in the small shape and after a hard post processing in PS to eliminate all the stacking artifacts. In full resolution it looks terrible. Also some picture areas are unsharp through the stacking process, especially when the DOF increasing fast. Can anyone give me some advice to solve that problems, or is the objects structure too complex for the stacking programs?

Thanks in advance,
Markus

etalon · Post by **etalon** » Wed Jan 06, 2016 3:32 pm

Hello folks,

and an other one:

Erythrine, 70 layer stack, 0.005mm steps, Helicon Focus (A/B/C), post processing PS CC2015, FOV ca. 3.5mm:

Cheers Markus

ChrisR · Post by **ChrisR** » Wed Jan 06, 2016 4:05 pm

Hi Markus
Welcome to the forum!

These look pretty good at this size.
(By the way, the forum limit on image dimesions is 1024 x 1024, because the images run off people's monitors. If you could resize we'd be grateful).

Can you be a little more specific about the stacking problems you have?

What kind of artifacts - halos or transparent foreground? Or dust on your sensor?
DMap should give you a good image where things do not overlap, but you do have to get the settings about right (the three sliders). There's a good tutorial on the Zerene Stacker website.
For overlapping needle-like structures, "physics" gets in the way. There will be some "transparent forground" where the lens sees around a foreground feature, particularly to higher contast things behind, but on large surfaces you wouldn't see those.
For the overlaps, Pmax is about as good as the stackers get. Often where an image has a lot of difficulties, sub-stacks (referred to as "slabs" or "slabbing") can help a lot. If you do a search on those terms you'll find some programs (linked to from the forum) for batch scripts which will generate slabs for you.

etalon · Post by **etalon** » Thu Jan 07, 2016 9:41 am

Hello Chris,

oops, I´m sorry. I´ve resized the pics...

Well, let me show you my problems:

ZS DMap

ZS DMap

ZS PMax

All pics are 100% crops of the stacked image without any postprocessing.

I tried very much different settings and this one, I used, seems to be the best for me. I tried slabs, too. But there was no improvement in the results. Dust on the sensor is no problem, and the unsharp edges are maybe a problem of my setup (see below). But the other artifacts, shown above, destroying the whole picture, especially if you like to make poster prints...

I also did some tests today and find a problem in my setup which I can´t explain to myself. Test target was a CMOS-Sensor silicon chip. In case it was a little bit tilted under the objective, I do a small Focusstack with 7 layers to cover it completely sharp. Well, the test shows me a little distorsion and a little field curvature toward the edges. But it also show me following:

Focused on the lower left edge (highest point) and started the acquisition with mirror lockup and 10s delay to the shots, the sharpest result on the lower left edge I got was that:

In the same stack the best result for the upper right edge (lowest point) was that:

The upper right edge looks pretty sharp while the lower left edge is unsharp. Now I did the same without mirror lockup and capture while I use live view. Than the results are these:

Lower left corner:

Upper right corner:

All pics are 100% crops.
Nothing else was touched or changed. What happens here? I checked it several times and could reproduce the results. I also checked all mechanical stuff, tighten all threads, etc. I could not improve the results. Optic seems to be ok, because I got a sharp image all over the target, but not at the same procedure. Can anyone help me?

Thanks,
Markus

ChrisR · Post by **ChrisR** » Thu Jan 07, 2016 10:48 am

Are your "alignments" settings the default ones?
(=20% if I remember correctly)

It's the "blurred" effects which I don't understand. The others can be dealt with, more or less.

Are all the pictures in correct sequence? Ummm...

etalon · Post by **etalon** » Thu Jan 07, 2016 11:17 am

Yes, all alignment settings are default (20%) and also all pics have been in correct sequence.

JH · Post by JH » Thu Jan 07, 2016 11:37 am

Hi

I like your pictures and your impressive set-up that you show in your presentation.

The 100% crops looks like there is some vibrations.
At short exposure-times the entire sensor will not be open at the same time, so it
is possible to have vibration in top and not in the bottom of the picture.
If this is the case I suggest longer exposure times or flash.

There is of course an artistic value with shadows in a picture, shadowless pictures
can be "flat" and not so interesting. But I suggest that you test a stack with
a more diffused light to see if it makes the "hairs" easier to stack.

Regards Jörgen

ChrisR · Post by **ChrisR** » Thu Jan 07, 2016 11:45 am

Ah, I understood that individual frames were sharp.
No?

I don't remember seeing the setup. On Markus' website I was looking at pitcures of Jupiter

FOund it :
http://www.photomacrography.net/forum/v ... p?p=182650

Yes you must use "EFSC", or flash.
That setup looks quite likely to suffer from vibrations

etalon · Post by **etalon** » Thu Jan 07, 2016 1:10 pm

Hi Jörgen,

thanks for your minds. Well, it is possible to have some vibrations, but, as I tried it, shorter exposures would deliver better results as longer ones. When using extreme short exposure times, the cameras curtain is not completely open at the same time, thats right. But that is only the case with exposure times less than 1/1000sec and that is much to short for my exposures.

I also don´t know vibrations wich are damped so fast (only a part of a 1/1000sec), that only a part of the picture suffers from it. Flash would be one possibility, but not my favorite one, because lighting a specimen is much easier with continuous light sources. So I hope there is an other possibility to fix that problem.

I´ll test an other stack with more diffuse light, like you recommend it.

Chris, you understand right. The pictures was sharp, but I guess not over the entire FOV. There are some unsharp regions in one corner (however), like I showed it in my other test.

That setup looks quite likely to suffer from vibrations

What let you assume that? Can you please be a little more specific, what you think, that let vibrations not be damped enough or avoid it? It´s important for me to figure out the problem.

I personally think it´s potentially not (only?) a problem of vibrations. I do all the acquisitions in the same way, only the exposure time varies. As you can see in the other postet test-photos, that there are pretty sharp pics possible and available. If there where some intrinsic vibrations, they should affect all acquisitions in the same manner, or I´m wrong?

Some more ideas what the problem could be and how I can fix it?

Thanks for your help, guys.

Cheers Markus

BTW, what is EFSC?

JH · Post by JH » Thu Jan 07, 2016 1:42 pm

Hi Markus
I also prefer continuous light when possible.
If you do not want to use flash permanent just try it once to determine if there is a vibration issue or not.
Regards Jörgen

ChrisR · Post by **ChrisR** » Thu Jan 07, 2016 4:58 pm

Jörgen's advice to try flash, is sound - just remember to diffuse it quite a lot! Use the fastest shutter speed you can (1/200th?)

I thought the "rig" looked susceptible to vibration problems because it has quite a high angular momentum. In other words, quite a long piece of support with quite a heavy mass connected.
"Connected mass" is often overlooked when considering vibrations. It extends damping periods. After the rail has moved, it will take some time to stop moving.
Also, any vibration transmitted through the base, will set up a movement. I notice that the blur ( if it is blur) is directional in some cases, that suggests a mechanical constraint "swinging". SOme vibration-absorbing feet should help with that.

You will also get vibrations due to both the mirror, and the shutter opening.

You can avoid mirror vibration by using lockup, and waiting, but the shutter is also significant.
As you have a Canon 5d2, you have a couple of modes which will avoid shutter vibration, using "Electronic First (or Front) Shutter Curtain."

The only snag is that you can't use flash as well.

A forum search on EFSC ( OR EFCS...!) will give you a number of hits.
If you start here, you will get the picture:

http://photomacrography.net/forum/viewtopic.php?p=72108

Charles Krebs · Post by **Charles Krebs** » Thu Jan 07, 2016 5:14 pm

BTW, what is EFSC

It is a mode where the camera will utilize an electronic first shutter curtain. This reduces camera induced vibration greatly. You must be shooting with live-view turned on, and have the camera "Silent Shoot" mode set to "Mode1" or "Mode 2" as indicated (see below) in the EOS 5DII manual. This way you will have no mirror or shutter vibration when an exposure starts.:

etalon · Post by **etalon** » Thu Jan 07, 2016 5:53 pm

Thank you guys!

That helps me a lot. Tomorrow I´ll try the EFSC mode. It seems to be exactly that what I´m looking for. I had never heard of that before...

"Connected mass" is often overlooked when considering vibrations. It extends damping periods.

Chris, you are right, but it also takes a longer time to accelerate that mass...
Would you think, it is better to move the object instead the camera for the stacking steps?

Did anybody know, wether the EFSC mode works with an external automatic releaser?

Have anybody an idea what happens in my test above? Why is in different shutter modes one corner sharp and the opposite corner not and vice versa with an other mode? Vibrations through the stepping mechanics should affect the whole and every picture, not only one... I´m a little bit confused about that results.

Cheers Markus

etalon · Post by **etalon** » Thu Jan 07, 2016 6:38 pm

...I can´t sleep with that in mind, so I started one more test right now. When I going to make the settings for EFSC, I saw, I work still with that settings (without my knowledge)!

That means, the pictures in my test above, which I shot during live view, was captured in EFSC mode1. But that would not declare, why I have that unsharp corner...

And why is captured with mirror lock the opposite corner unsharp without changing anything else???

ChrisR · Post by **ChrisR** » Thu Jan 07, 2016 6:53 pm

Perhaps it's sharp in multiple frames, but it's moved for some reason so you get an image from two frames.
This is Dmap, yes?
You should be able to see the frames, and any movement, in Zerene Stacker if you scroll the mouse wheel to "fly" up and down the source images. You can do that with the zoom set high.