Stacking exercise

[rudi]

Up till now I stacked my photo's in PS CS. Reading that there ar better programs to the job, I tried Zerene Stacker and Helicon focus. And I was a bit surprised about the outcome. Because I have no experience with Zerene nor Helicon I guess I am doing something wrong.....Can someone shed some light on this ????

As exercise, I stacked 3 pictures of a hoverfly, jpg out off camera, no other PP has been done. Pictures were taken outside, handheld.

1: Stacked in CS6


2: Stacked in Zerene Pmax


3: Stacked in Helicon Meth A, R8, S6

[Kodiak]

Can someone shed some light on this ????

The light I will shed is on your test protocol. HORRIBLE!
You are working against yourself, your software… and you are expecting
miracles.

Obviously here, some movements were induced while capturing since the
three take do not stack correctly. By focus stacking, immobility is a must
since no software will stretch or skew a slice to make it fit among the others.

Rudi, do such tests with absolutely immobile subjects and immobile camera.

I would be very interested to see your new tests with the same softwares. I
went through this some years ago and selected one of your three candidates.

[rudi]

The light I will shed is on your test protocol. HORRIBLE!
You are working against yourself, your software… and you are expecting
miracles.

Obviously here, some movements were induced while capturing since the
three take do not stack correctly. By focus stacking, immobility is a must
since no software will stretch or skew a slice to make it fit among the others.

Rudi, do such tests with absolutely immobile subjects and immobile camera.

I would be very interested to see your new tests with the same softwares. I
went through this some years ago and selected one of your three candidates.

Hi Kodiak,

you are not helping here. I want to stack LIVE critters, and live critters tend to MOVE.
I don't expect miracles. I was just wondering why the so called better programs aren't doing so good in MY case.

[Kodiak]

Hi Kodiak, you are not helping here. I want to stack LIVE critters, and live critters tend to MOVE.

My reply was based on the assumption that you were testing the softwares.
If true, my recommendations hold.

If NOT true, I can only say that all future stacking attempts will be based
on 95% luck! …since no known software will render what you saw but what
they see when stacking your shots.

Good luck!

[DanK]

Kodiak, this does evaluate part of what the software does. Most stacking software does two things, in sequence: it tries to align the images, and then in creates a composite. This set of images is so badly aligned that it can only test the first, but it does test that well, and I am frankly surprised and a bit puzzled.

First, why I am surprised: I didn't expect that PS would align the fly's head better than the other two.

Second, why I am puzzled: compare the areas of the image NOT close to the fly, like the leaf at the top left. Although PS did a better job of aligning the fly's head, it did an awful job with the leaf, while Zerene did quite well. Helicon handled it worst. This must have something to do with the alignment algorithms the three use.

I'm also puzzled why Zerene got the left hand side of the image fairly well (check the bottom edge of the lower leaf too), while it didn't get the eyes. I wonder if the movement between images was rotation around an axis near the left side of the image.

I do agree with Kodiak that it would be helpful to test the stacking portion of the software rather than the alignment portion, and for that, you will need images that are withing the alignment tolerance of all three.

[chauncey]

I have been using Zerene for about 3 weeks on my floral stuff and found that it does do a great job on,
as Kodiak mentioned, immobile subjects, except...it does introduce unacceptable noise.
Prior to that, PS CC was my weapon of choice, albeit, a lot of artifacts. Then it quit stacking at all.

But PS CC upgraded last week...the stacking works much better than before with virtually no artifacts.
As a side note, I took a stack from PP CC and Zerene and stacked them together...looked good.
They both need hard crisp edges to lock onto and flowers are tough in that regard.

With three images Rudi, repairing the PS CC image would consume about 30 seconds, but the exposure
problems...what happened with that and why jpeg images?

[DanK]

I have been using Zerene for about 3 weeks on my floral stuff and found that it does do a great job on,
as Kodiak mentioned, immobile subjects, except...it does introduce unacceptable noise.
Prior to that, PS CC was my weapon of choice, albeit, a lot of artifacts. Then it quit stacking at all.

But PS CC upgraded last week...the stacking works much better than before with virtually no artifacts.
As a side note, I took a stack from PP CC and Zerene and stacked them together...looked good.
They both need hard crisp edges to lock onto and flowers are tough in that regard.

With three images Rudi, repairing the PS CC image would consume about 30 seconds, but the exposure
problems...what happened with that and why jpeg images?

Chauncey,

Very interesting. Thanks for posting this detail. I won't have time for a few weeks, but I will plan to do some comparisons of Zerene with the new PS.

Have you done comparisons of PMax and Dmap? they are often quite different for flowers. In general, I prefer DMAP for that purpose, but PMax does create fewer halos.

Dan

[chauncey]

Dan, when using Zerene, I routinely use both methods and compare but, one of the requires an
adjustment along the way (don't remember which method does) and it can make a huge difference.

Regardless, now that PS CC is working well I won't be buying Zerene.

[rudi]

Thanks Dan for your input.

[rudi]

[QUOTE=chauncey;454797....With three images Rudi, repairing the PS CC image would consume about 30 seconds, but the exposure
problems...what happened with that and why jpeg images? [/QUOTE]

I tried a few times importing the CS6 stack into Zerene and retouch from there.

Exposure problem: need to work on my flashdiffusion....but actually the Zerene stack comes closest to the raw file, which isn't that bad...

Why jpg: when I see a possible stack I import the jpg's as layers in CS6, auto align auto blend, which goes relatif quickly. Iff I see the stack works I start from scratch with the raw file.

[Glenn NK]

My first reaction was, "only three images?" Then I wondered about the FL and f/stop used? (no EXIF obviously on a stack). I've done stacks using three images, but the resultant DOF was still quite thin.

Considering the fly moved - this is not a bad job at all. Three dimensional geometry isn't everyone's forte, but it's useful to think about it when doing stacks. If the fly moved in a straight line and the camera was a considerable distance from the fly (very long FL), then the misalignment of the images could be adjusted for. But the insects I know don't walk in straight lines, and this completely changes the problem as different parts of the body come into view and/or disappear as the fly rotates (changes direction), and these different body parts will be in different focus (especially as the fly moves).

I've stacked a few images, and I've attempted to stack flowers with a very small breeze and it's not possible to achieve perfection. A moving fly is another level of difficulty.

There is a fairly well known macro/stacker that goes by the name of "Lord V", and he has posted a fair bit of helpful information on the topic. I just googled him:

http://photography-on-the.net/forum/...d.php?t=807056

This is one link - there are many others.

BTW, this guy hand holds his camera - something I consider very difficult.

Glenn

[DanK]

I couldn't resist.

I took a 7-image stack that is very easy to stack: the images are well aligned, and it is not the type of image that would generally generate stacking images. I adjusted only white balance. I loaded the cr2 files into CC via bridge and stacked. I saved the stack as a level-9 jpeg. I then exported 16 bit TIFFS to zerene and stacked both ways, setting the DMap contrast slider to zero. I saved as TIFFs back into LR and exported as JPEGs with approximately the same quality level.

The times for aligning and stacking, not counting the time to load the images:

CC: 2:10
Zerene PMax: 0:45
Zerene DMap: 0:46 (usually the difference is greater).

Here are the images. I did no further editing.

CC:



PMax:



DMap:



My take is that at this resolution, with images this straightforward, the biggest difference is the time. Of course, these are very low-resolution images, and the differences might be clearer if printed or blown up. However, I think its probably the case that to really see the differences, you need to try more problematic images--for example, brightly colored images, images with subtler textures, or images prone to halos. I haven't compared CC, but with Zerene, the first two favor DMAP, and the third favors PMax. The differences are often easily visible, even at this resolution. I don't have the time now, but the next time I am stacking a more difficult images, I will try to remember to do a CC stack along with Zerene to post.

[Geoff F]

I've tried various stacking software, Rudi, and without getting a satisfactory result - except from CS 5 which works most of the time.

Amongst the software which failed for me; one just ground to a halt with a three image stack while another worked but produced a seriously over sharpened unnatural looking result.

Yes, two shots is the average for me as well when using live subjects in outdoor conditions. Using more images just tends to slow my computer, or stop it altogether, unless I'm working with smaller sized Jpeg images. But Adobe CS will work with full sized Raw conversions (converted to Psd) on two or three layers.

When working with real life conditions and there is quite a bit of movement between shots I have overcome this by manually moving the layers closer to alignment. I drag some guide lines onto suitable points and use that as a rough alignment position.

[chauncey]

Humor me Dan, take the zerene images into PS>resize them to the appropriate size of 300 ppi>layer the
two zerene images on top of the PS stack>align and stack them all together and let us see the results,
along with a screen shot of the black boxes.

[DanK]

Humor me Dan, take the zerene images into PS>resize them to the appropriate size of 300 ppi>layer the
two zerene images on top of the PS stack>align and stack them all together and let us see the results,
along with a screen shot of the black boxes.

Why? once they are stacked with either program, there is nothing to be gained by stacking them again. Seems to me it is just a chance to introduce errors into the file.

I do sometimes stack in two different ways and use one composite to touch up the other, but I have never bothered to composite two composites. If the images are in focus, I am happy.

[AsproUK]

Focus stacking is my special interest and have used various programs. In this discussion, no mention has been made of the fact that, unless you use a racking plate on a tripod ( or move the camera towards the subject ) re-focussing at successive image planes will produce an image of a slightly different magnification. This will display as if the subject moved, even if it did not.

Alignment may then "interpreted" differently by different algorithms. If the camera/ lens has a constant focus point and moved incrementally, the software can pixel match accurately. If the subject moves, this will deliver an imperfect match, no matter which software is used.

But it is fun anyway!
PS Software uses contrast to pixel match and if contrast is low, results can be variable.

[deetheturk]

Welcome to the forum Michelle, looking forward to seeing your images

David

[chauncey]

Hey Michelle, glad to have you aboard...
FWIW, lately I've been combining/stacking my Zerene stacks with my PS CC stacks and sometimes it helps and other times not. Repeating the stack, regardless of software used, gives inconsistent results. Don't understand why that is the case.

PS Michelle...are you suggesting that you get better results using a rail system?

[DanK]

Focus stacking is my special interest and have used various programs. In this discussion, no mention has been made of the fact that, unless you use a racking plate on a tripod ( or move the camera towards the subject ) re-focussing at successive image planes will produce an image of a slightly different magnification. This will display as if the subject moved, even if it did not.

Alignment may then "interpreted" differently by different algorithms. If the camera/ lens has a constant focus point and moved incrementally, the software can pixel match accurately. If the subject moves, this will deliver an imperfect match, no matter which software is used.

But it is fun anyway!
PS Software uses contrast to pixel match and if contrast is low, results can be variable.

Michelle,

Glad to have you join the forum.

I think you have two different issues: alignment and magnification. here's my two cents.

We had another thread on magnification recently. Regardless of whether you use a rack or refocus, magnification will change between slices. The software packages I have used all deal with this well, as long as the changes in perspective are not huge. One of the old Combine packages tended to mess this up around the borders of images, but I haven't seen any problems with this in ages using either Zerene ( my usual) or PS, which I have used for stacking a few times.

Alignment is often used to mean something very different, which is successive images that are not centered the same. In my experience, there are serious limits on the ability of software to handle these differences. This is why my efforts to stack bug shots have been such a flop. I don't have problems with work on a tripod.

Dan

[Glenn NK]

Michelle,

Glad to have you join the forum.

I think you have two different issues: alignment and magnification. here's my two cents.

We had another thread on magnification recently. Regardless of whether you use a rack or refocus, magnification will change between slices. The software packages I have used all deal with this well, as long as the changes in perspective are not huge. One of the old Combine packages tended to mess this up around the borders of images, but I haven't seen any problems with this in ages using either Zerene ( my usual) or PS, which I have used for stacking a few times.

Alignment is often used to mean something very different, which is successive images that are not centered the same. In my experience, there are serious limits on the ability of software to handle these differences. This is why my efforts to stack bug shots have been such a flop. I don't have problems with work on a tripod.

Dan

Dan:

I agree with you.

If the distance between subject and camera changes, then parallax error is introduced. This can only be eliminated by shooting from infinity - but at infinity the camera couldn't "see" the image.

If the lens is re-focused for each shot, and the nodal point does not move (an impossibility perhaps?), then everything is fine (except I think the nodal point moves - but could be wrong).

In the attached diagram, the amount of the object seen varies with the camera to subject distance. As the camera gets closer, less and less of the object is seen. To some extent, each image is of a different object than the one next to it and so on.

To add to this problem, suppose that the object has little marks at even spacing along its surface (veins in a leaf or petal). In successive images, the amount of the subject will not only change, but the ratio of the distance of the spacings will change. This is a very difficult task for the stacking program. The fact that it works at all, is remarkable.

However as the camera is moved farther away from the subject, (longer focal length) then differences between frames becomes less - hence the longer the focal length, the less problems will occur.



The best option (IMO) is to use as long a focal length as is practical - this puts the camera/lens farther from the object which to a small degree is akin to being closer to infinity.

In my personal experience, when I used a 100 macro on a crop sensor, my stacks were better than when I used the same lens on a FF body.

With the 1.6 crop body, for the same subject and same framing the camera would be 1.6 as far from the subject as is the FF body.

Anecdotally (not a rigourous proof by any means), the best stacks I've seen by others were done with longer FL lenses, particularly if a FF body was used (in a Canon example, 180 mm + FF body).

This is one of the reasons I acquired a 70-200 zoom lens (it's also an effective landscape lens, but that's another topic).

Glenn