Learning Center What is a mineral?The most common minerals on earth Information for Educators Mindat Articles The Elements The Rock H. Currier Digital Library Geologic Time

Minerals by Properties Minerals by Chemistry Advanced Locality Search Random Mineral Random Locality Search by minID Localities Near Me Search Articles Search Glossary More Search Options

The Mindat Manual Add a New Photo Rate Photos Locality Edit Report Coordinate Completion Report Add Glossary Item

Mining Companies Statistics Users Mineral Museums Clubs & Organizations Mineral Shows & Events The Mindat Directory Device Settings The Mineral Quiz

Photo Search Photo Galleries Search by Color New Photos Today New Photos Yesterday Members' Photo Galleries Past Photo of the Day Gallery Photography

Discussions

💬 Home 🔎 Search 📅 Latest

╳Discussions

💬 Home 🔎 Search 📅 Latest

Groups

Recent Images in Discussions

Mineral PhotographyDepth perception in macro photos

16th Aug 2015 19:24 UTCRonald J. Pellar Expert

In another thread, that was closed to further discussion prematurely, Bob Harmon alluded to a perceptual difference between some hi-def macro photos. One respondent thought it might be do to stacking. There is a real difference in photos when "rail stacking" is used as opposed to "focus stacking". Rail stacking is when the camera or specimen is moved to move the focal plane through the specimen to create the stack. Focus stacking is performed by keeping the specimen and camera stationary and changing the focus of the lens to create the photo stack. Rail stacking completely eliminates parallax in the resulting images. Whereas "focus stacking" preserves parallax throughout the stack. This difference can be perceived as an apparent presence or lack thereof of depth. In photomicrography, stacked images are all necessarily "rail stacked", so their is no basis for comparison.

It would be interesting to find out how the photos compare in the method that they were "stacked".

16th Aug 2015 22:10 UTCDavid Baldwin

I was about to reply to that discussion when it was abruptly closed. I see what Bob meant in the photos concerned, and I think the problem lies with the fact that the entire specimen is in focus, whether it be focus or rail stacked. This is not a natural view and not what our brain expects to see, especially when close up, and just looks wrong when viewed on screen. With the larger, complete specimens, the effect is worsened by the neutral background and in some cases, over zealous lighting which eliminates most shadows.

Just to be clear, this is not a complaint about photo quality, it's just an observation of the effect of certain modern photographic techniques. The photos look great, and are one of the main reasons I visit this site regularly.

16th Aug 2015 23:29 UTCBob Harman

I am a little dismayed that my thread was closed (prematurely as RON P noted) and thanks for your comments more or less supporting my thoughts. That is all I wanted to note…..nothing more! CHEERS……BOB

14th Sep 2015 13:05 UTCOwen Melfyn Lewis

Bob,

My sense of it is - as Dave explains - that the brain is presented in 2-D with an all-sharp image of a 3-D object in a way that eyes used directly never can. Hence the slight sense 'weirdness' sometimes felt when looking as a well-stacked photo.

I also see where Ron is taking his explanation but I remain to be convinced. In simple terms of keeping the lens-to-object-critical-focus distance the same or not, I'm sure he's right - but without understanding (and I don't) the detail of the software image stacking process there is the lack of a complete explanation.

AIR all the main 'stacker' software packages actually contain more than one stacking algorithm for keenies to choose between in combining their stacked images, so as to be able to select the one that provides the most pleasing effect from the multiple stored images of a single object delivered from the digital camera. AIR, the parallax effect is one that is treated differently according to the algorithm selected?

It's worth adding that there *is* a macro-rail solution for photomicrography. Made by a Spanish company, it replaces the vertical support and focusing arrangement of the microscope, replacing it with custom-engineered metalwork, gearing, step-motor and software control to govern the lowering of the optical pod secure to the top of is in accurately stepping the descent of the pod by a few microns between each exposure. About a year or so ago this was discussed at some length in this topic area but a search today can't bring the exchanges to light, so I guess they have been deleted. :-X However, some pics can be seen here http://macrorail.com/VerProductoEng.php?P=2

17th Sep 2015 01:59 UTCEugene & Sharon Cisneros Expert

Owen,

There is a difference in the perspective when doing rail stacking. As the camera moves closer to the subject the view of the subject and its surroundings change. Each frame will have a slightly different perspective and an apparent distortion of the subject with respect to the reference frame.

Focus stacking, depending upon the lens, can eliminate some of the afore mentioned issues. If the lens has internal focusing (doesn't change length) then much of the shortcomings of rail stacking are eliminated. An issue with this method is when the lens does not have internal focusing and it does change in mechanical length. In this case, some of the problems associated with rail stacking occur.

The ratio of the focal distance to the required rail movement also is a factor to consider. So, for subjects that don't have a lot of depth, the issues are less.

As Ron mentioned, it is almost necessary to employ rail focusing for micro subjects.....but, not always. Some have used internal electronic focus lenses with infinite focus microscope objectives affixed to the front. The focusing stacking is done by incrementing the internal focus motor. If the subject is not too deep, this can work well. But, for most micro stacking, the rail method works the best.

You mentioned motorized stacking rails and here is one commercial unit that is available and from reports is very good.

Cheers,

Gene

17th Sep 2015 09:14 UTCVolker Betz 🌟 Expert

Hi,

the attached stack of only two pictures may expain the situation.

The top picture was made with focus stacking, the result has a persective (entocentric)

The lower picture was made with rail stacking, the result has no persevtive /telecentric).

Focus stacking is easy made macro lenses, i do that frequent.

To do focus stacking in the macro range the equipment has to be built.

Volker

20th Sep 2015 16:24 UTCJay I. G. Roland

This is a very interesting topic and one I shall be following as I too like to make macro images. One thing Volker, why is there a halo around the left nut in the top image and around the right nut in the bottom image?

I use the focus rail method along with Zerene Stacker software and oftentimes get a similar halo effect but have never managed to pin down quite why it happens.

Regards,

Jay.

20th Sep 2015 19:24 UTCHarald Schillhammer Expert

That is the reason why many of the images shot with microscope lenses appear rather flat. Sometimes they don't, but that obviously has to do rather with lighting that fakes some kind of perspective.

As Volker suggested, a microscope lens (e.g. a 10/0.28) with internal focussing would be "the" hit, but I am afraid a construction like this at this scale is very difficult. Just imagine - keeping in mind the razor thin DOF of such lenses - how fine the focussing increments would have to be. Such a lens would be very costly and we probably would have to sacrifice some of the imaging quality that, e.g., a Mitutoyo lens delivers.

21st Sep 2015 09:44 UTCHarald Schillhammer Expert

Came to think of this again.

I think the easiest solution would be to modify a bellows in a way that the front standard stays fixed while the rear standard with the camera is attached to the focus rail, so every step of the focus rail slightly changes the magnification ratio and focus plane and thus preserves the perspective.

Charles Krebs did a similar thing to achieve a "macro-infinity" setup. The results were quite breath-taking, see here.

Why shouldn't this work with extreme magnifications as well?

21st Sep 2015 11:15 UTCVolker Betz 🌟 Expert

Hi

I am on the way to do experiments. To di this is a bit tricky but Rainer Ernst is bulding a variable lenght extention tube with a fixed front and moveable camera position. As soon I have the bits I will do test pictures with manual stacking. To add a motor is not so difficult.

Volker

21st Sep 2015 11:41 UTCJolyon Ralph Founder

This is where a lighter mirrorless camera really becomes a significant benefit.

21st Sep 2015 12:15 UTCHarald Schillhammer Expert

Jolyon & Katya Ralph Wrote:

-------------------------------------------------------

> This is where a lighter mirrorless camera really

> becomes a significant benefit.

Not necessarily. The stackshot rail easily carries my D4 + Novoflex bellows + lens (1kg +).

The tricky thing actually is that you have to disassemble the bellows because the rear standard has to be mounted to the stackshot rail. But how to fix the front standard? It cannot dangle in the air and you have to align it perfectly to achieve the precise optical path. Should be no problem though for a DIY experienced person :).

21st Sep 2015 14:22 UTCVolker Betz 🌟 Expert

Hi,

Rainer Ernst ( Stonemaster.eu ) has already designed a rack for variable tube lenght, replacing a bellows.

There is a prototype version with a fixed front stand and a variable rear stand for the camera which can be adjusted by a spindle drive.

This alllows to work with a fixed object distance and a variable image distance. I wait now for the variable tube also have to to solve come mounting issues.

Volker

22nd Sep 2015 00:15 UTCGeoff Van Horn Expert

Jolyon & Katya Ralph Wrote:

-------------------------------------------------------

> This is where a lighter mirrorless camera really

> becomes a significant benefit.

Another option is to have the camera mounted to a fixed object and use the stackshot to move the subject. I haven't had a chance to try this method yet, but saw many people are using it, especially with a bellows setup. Once I get the camera on my trinocular microscope running again I plan to use the stackshot in this manner.

22nd Sep 2015 18:10 UTCRonald J. Pellar Expert

Geoff,

Fixing the camera and moving the specimen is exactly like fixing the specimen and moving the camera. They are both telecentric setups and you lose depth perception. Changing the lens-to- image plane distance is the only way to retain parallax effects (where image plane is defined by the plane of finest, or best, focus).

22nd Sep 2015 19:11 UTCGeoff Van Horn Expert

i was speaking about the weight of the camera being a burden to the stackshot, not the original issue of loss of depth.

22nd Sep 2015 20:43 UTCHarald Schillhammer Expert

Geoff Van Horn Wrote:

-------------------------------------------------------

> i was speaking about the weight of the camera

> being a burden to the stackshot, not the original

> issue of loss of depth.

The stackshot can carry quite a burden. My setup with more than 1 kg is not even a challenge for it ;).

23rd Sep 2015 05:15 UTCGeoff Van Horn Expert

Harald Schillhammer Wrote:

-------------------------------------------------------

> Geoff Van Horn Wrote:

> --------------------------------------------------

> -----

> > i was speaking about the weight of the camera

> > being a burden to the stackshot, not the

> original

> > issue of loss of depth.

>

>

> The stackshot can carry quite a burden. My setup

> with more than 1 kg is not even a challenge for it

> ;).

No doubt, but I don't think it would stand up to my O.C White microscope.

[img]http://img.photobucket.com/albums/v737/airborne200208/DB0B6788-353D-413E-8BE1-6D136C0A85CA.jpg[/img]

[img]http://img.photobucket.com/albums/v737/airborne200208/F70343AE-B890-4E7A-96CC-3FD461D15754.jpg[/img]

27th Sep 2015 15:36 UTCVolker Betz 🌟 Expert

Does the stacking method (rail stacking or focus stacking) have an influence on the depth perception of the final picture ?

Depth perception is a result of human vision which can “see” a depth on a two dimensional picture if parallel lines are reproduced perspectively.

Perspective distortion is a known phenomen in photography.

( https://en.wikipedia.org/wiki/Perspective_distortion_%28photography%29 ) and caused by lenses with a low angle of view. In normal photography this are tele lenses.

In macro photography all lenses have a low angle of view and so also a distorted depth perception independent from their focus length.

Focus stacking is (typical) done with the focus of the macro lens with constant object distance.

Rail stacking is done by variable object distance and fixed picture distance.

I compared focus stacking and rail stacking on a home made stand using a rack made by Rainer Ernst (stonemaster.eu) and a 63 mm Zeis Luminar at a reproduction scale of ~2.75. As target I used paper with a mm grid mounted with 45 ° tilt against the optical axis. The depth of view is 5.7 mm and field of view 8 mm. Stacking was made with Helicon focus method B.

Results and interpretation:

Rail stacking needs 5.7 mm variation in object distance. Reproduction scale is constant.

Focus stacking needs a 30 mm variation in picture distance and reproduction scale changes according.

Almost identical stacks are available with focus stacking and rail stacking.

It is no significant difference in perspective to a single shot made with closed aperture.

The lack of depth perception in 2D macro pictures is caused by the low picture angel of the used lenses and not a result of stacking. (tele lens effect)

Depth perception is well possible with stereo pairs (3D).

As I use focus stacking frequently with a macro lens my conclusion is:

Focus stacking works well up to reproduction scale of 1:1. Here small changes in focus( picture distance) cause equivalent changes of focus towards object distance

In higher reproduction scales the changes in picture distance must be much greater to get a focus change. This must be closer tested.

For reproduction scales above 1:1 (up to 20:1 are practical possible) rail stacking is at present easier to do.

Both methods are equal concerning limited depth perception.

27th Sep 2015 17:36 UTCOwen Melfyn Lewis

Thanks Volker. A useful exercise. But we are still not considering one factor that I think may be vital or unraveling this. Namely, creating correct perspective in a 2-D image obeys strict mathematical rules as has been understood since the 15th(?) century. It seems to me (though I do not know it to be so) that the mathematics for perspective correction can (and should) be included in the any of the several stacking algorithms of the different proprietary brands of stacking software.

It follows that it would be interesting, having made (say) 20 image of an appreciably 3-D object (I'd suggest a cube for easy detection of distortions) by (a) the rail focusing method and (b) by lens remote-controlled re-focusing. Then to compile these as stacked images and to compare the results across the different algorithms Helicon provides or - even better - to do the same for all the algorithms of Helicon Focus, Zerene Stacker and Combine-Z. I'll be surprised if there is not a correct perspective solution for both focusing methods obtainable with at least one of the first two names companies algorithms. But I don't thing one can get correct perspective in the final image for both of the focusing methods. So a different algorithm for each of the methods should give the best result?

Just working it as a mental exercise, it seems to me that the maths for correct perspective in stacking a given number of images produces be (a) the moving camera method and (b) change of lens focal length can't be the same. It would then follow the correctness of perspective must rely:

1. On the skills of the software engineer.

2. Correct algorithm selection by the *user* of the best software algorithm for the make of software at his disposal and for the focusing method the user has chosen to employ.

3rd Oct 2015 22:14 UTCRonald J. Pellar Expert

Macrophotography Stacking Study

The recent discussion concerning the two methods of photo stacking, to provide increased depth-of-filed in macrophotography, has stimulated me to perform a comparison between the two in as controlled a way as possible. The specimen in this study is a machinist’s “Johansson” gauge block – used for verification of accuracy of measuring instruments, e.g., micrometers – that is 1.3434 inches square. In addition, two nuts of the same size were placed, one in the foreground and one in the background, as out-of-focus objects for perspective. The camera was mounted on a rail on the tripod. The camera was focused on the left front corner of the “Jo” block. A series of 45 photos were taken with the rail advanced manually between shots (finer increments than were actually needed). The shots were controlled from a computer using the USB link from the camera after a time interval to let the rig settle down. The focus point was checked from the computer and “live view” mode set to 200%. The photo series was ended when the right rear corner of the “Jo” block was in focus. At this point another series of 20 photos were taken using the smallest focusing increment in the camera control software without touching the physical setup. This was a “Focus Stack” series and was completed when the left front corner of the “Jo” block was verified as in focus with “Live View” set at 200%.

The alignment and stacking of both the “Rail” and “Focus” series of photographs were performed using the Adobe Photoshop CS6 software.

Rail Stacked Photo:

Focus Stacked Photo:

In the “Focus Stacked” photo, normal perspective is maintained and the “Jo” block appears to be square and appropriate to the out-of-focus nuts, i.e., all resting in the same plane. However, in the “Rail Stacked” photo, the “Jo” block appears trapezoidal and not in the same plane as the nuts, at least to my eyes. The removal of normal perspective in “rail stacking” can cause a “distorted” appearance to items being photographed and stacked. In my opinion, “focus stacking” is the preferred way to go in macrophotography, where possible. It is also easier to use with the camera control software. In particular, the ability to use magnified “live view” is really handy for checking critical focus without disturbing the camera and/or specimen setup.

Definitions:

Rail Stacking: A series of photographs are taken while moving either the specimen or the camera with respect to each other to achieve in-focus slices, through the specimen of interest, to be assembled into a stacked photo having a greater depth-of-field than is possible with a single photo. The camera focus is not altered during the taking of the series of photos.
Focus Stacking: A series of photographs are taken without altering the specimen or the camera distance. The series of photos are acquired by changing the camera focus only to achieve a stacked photo having a greater depth-of-field than is possible with a single photo.

Equipment

Camera: Nikon D300

Settings:

Color Space = AdobeRGB

Bit Depth = 14 bits

File Format = Camera RAW

White Balance = Preset, using a gray card

Exposure Mode = Apertures Priority
Lens: Micro-Nikkor 105mm VR
Lights: Four goose neck 20w Tungsten-Halogen lamps
Rail: Valbon Super Mag Slider
Tripod: Gitza

Software

Camera Control: Nikon Camera Control Pro 2
Stacking: Photoshop CS6

3rd Oct 2015 22:47 UTCDoug Daniels

To tell the truth, I can't tell any difference between the two photos. Then again, I'm reduce to viewing them using a notebook computer (my desktop won't power up). So, a complicated subject, to be sure, and probably no one real answer other than "to each his own".

4th Oct 2015 13:11 UTCOwen Melfyn Lewis

Ron,

An interesting test.

I think that the trapezoidal distortion you refer to may not exist. Try the following as a crude check.

- For one of your images, place a paper straight edge along the forward edge of the square block. Mark off the length of this forward face.

- Move the strip to align with the rear top square face, using your right-hand marker to align with the right-hand end of the face.

- As one should expect, the perspective in the image assures that the rear square face is shorter on your screen that in the front face. Make the third mark in the paper to denote the seemingly reduced length of the rear face.

-Now repeat the exercise with the other image and using your already marked strip of paper. You will find (hopefully) that that the marked off distances from the first image exactly fit the dimensions of the secong image also.

The rail stacking and focus stacking methods return identically dimensioned results from both methods when the stacking is done with the PS CS6 image stacking algorithm. Is the stacking programme 'aware' of the multiple image making technique that you are choosing to employ? The counter-intuitive result obtained in the above test suggests strongly to me that the programme must be 'aware' of the change in camera set up and, invisibly to the user, swaps its perspective-dealing algorithm accordingly.

Whilst the conclusion in the above paragraph is speculative, from the simple test it is determined that the perspective in your two stacked images is the same. This does not prove that the perspective is correct (you can repeat the test making accurate measurement and doing the maths if you (really) want). However, the perspective, for better or worse, is the same.

4th Oct 2015 15:08 UTCJay I. G. Roland

Ronald, I have to agree with both Doug and Owen as I can percieve no difference in perspective (only in shading) twixt your two images.

Perhaps you might explain what you can see and why one image is preferable over the other?

Regards,

Jay.

4th Oct 2015 18:14 UTCRonald J. Pellar Expert

I got to thinking about the trial and realized that I had not used the same camera distance to the front edge of the block, therefor, the magnification in the focus stack photo is slightly greater than in the rail stack photo. The back edge of the block is the same dimension in both photos. However, the front edge is larger in the focus stack than in the rail stack. I will have change my procedure to make sure that the camera is the same distance to the front edge of the block at the start of both runs. I will do this later today and repost.

Owen: I have printed both images and measured them ant they are different! What has complicated the perspective changes is that the start of the rail stack is the front edge of the block and the start of the focus stack is the back edge of the block. Therefor, the distance between camera and the sharp focus plane are different in the two images.

4th Oct 2015 22:25 UTCRonald J. Pellar Expert

Owen:

You were right! Photoshop messes with rail stack and introduces perspective into the blended image!!!! The focus stacked photo is correct since it does have perspective. I verified these conclusions by comparing first and last photos to the blended image in each case, rail and focus stacking. I will have to acquire an appropriate stacking software to redo the comparison!

Photoshop CS6 has a lot more "context aware" functions. The blend tool will create panorama blends, as well as stacking blends, which can require skew and parallax corrections. This context awareness is actually adding parallax back into the rail stacked images, actually improving them from my perspective (to make a pun).

You learn something new every day!!!! I would love to receive any recommendations for which stacking software to consider and, if possible, why.

Ron

5th Oct 2015 11:57 UTCVolker Betz 🌟 Expert

Hi Ron,

Recently I did some experiment with focus stacking in the macro range (4:1).

It works there also, but you neeed very large variation in image distance to obtain small focus distances at object side.

This large differences in image distance cause also substantial differences in reproduction scale from picture to picture..

The stracking algorithm will rescale the difference for seamless integration of the sharp zones. Outlaying parts of the picture are discarded. As a side effect the picture losses perspective.

But greatest los off perspective is caused by the tele-lens effect which distores perspective.

All macro lenses work in the low angle range and surpress perspective.This is the price we pay for working distance !

See:

https://en.wikipedia.org/wiki/Perspective_distortion_%28photography%29

The only and best way to get perspective is 3 D pictures.

Concerning Stacking software: I am using Helicon focus since 2004. Zerene stacker may be better in some cases but is slow with large stacks.

My reason is: I Use Helicon remote for taking the stacks with a canon camera, stack with Helicon Focus and finish the picture with Helicon Filter.

Volker

5th Oct 2015 18:27 UTCRonald J. Pellar Expert

Thanks for the recommendation Volker. I will look into the various stacking programs available, but it helps to have some advance guidance.

In the focus stacked photo above the perspective is correct. The back edge of the block is 12% shorter than the front edge. The rail stacked photo has been "corrected" by Photoshop CS6 to reintroduce some perspective by some means unknown to me.

11th Oct 2015 19:42 UTCRonald J. Pellar Expert

I have downloaded trial versions of Zerene and Helicon and restacked the Rail shots with each of them. The results are identical to the above Photoshop CS6 rendition. They all restore parallax to the stacked images! In addition they slightly enlarge the image relative to the focus stacked image.

After further investigation, the stacking algorithms must provide alignment, scaling, and, in some cases, rotation in order to do a decent stack. The scaling of pixels outside of the fine focus area provides a natural perspective in the rail stacked images. This is very evident in a stacked image with "Scaling" turned off! In this case the block is equi-dimensional and all other out-of-focus items to the sides are rendered in multiple displaced images. With "Scaling" turned these images are properly superimposed and in proper perspective. There is an excellent explanation for the pros and cons for rail versus focus stacking by Zerene at

http://zerenesystems.com/cms/stacker/docs/troubleshooting/ringversusrail

In general, focus stacking is less problematic and produces a better rendering than rail stacking for mineral macrophotography.

19th Oct 2015 19:45 UTCRonald J. Pellar Expert

The very nice discussion by Zerene on this subject was placed in my previous message as an edit, not realizing that the edit would not reset the message to a "new" designation. It is sufficiently revealing comparison of rail and focus stacking to be worth repeating in this "new" message.

http://zerenesystems.com/cms/stacker/docs/troubleshooting/ringversusrail

It is well worthwhile reading and seeing the demo!

Ron

Mindat.org is an outreach project of the Hudson Institute of Mineralogy, a 501(c)(3) not-for-profit organization.
Copyright © mindat.org and the Hudson Institute of Mineralogy 1993-2024, except where stated. Most political location boundaries are © OpenStreetMap contributors. Mindat.org relies on the contributions of thousands of members and supporters. Founded in 2000 by Jolyon Ralph.
Privacy Policy - Terms & Conditions - Contact Us / DMCA issues - Report a bug/vulnerability Current server date and time: April 23, 2024 07:25:04

Go to top of page