First macro with extension Tubes

Images taken in a controlled environment or with a posed subject. All subject types.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

If I may - a personal feeling: I'd like to see the face of the caterpillar. Lighting is hard in macro, and you sometimes don't get detail where you want.
I'd use Curves or Levels, maybe selectively, in Photoshop or similar, to bring it up.

Objectives are lenses, but the ones on the subject end of a microscope. SOmetimes they're also called lenses.
WIth some of them, notably some Nikon ones, all the parameters are right to be able to use them on a camera bellows instead of a microscope. You can read about all that in the equipment forum.
The camera will be on Manual, so any make will do.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

thartl wrote:I definitely was talking about the top right leaf. not sure i their was any movement, but sometimes after I adjust my focus ring the camera is still moving slightly - and I need to wait longer to hit the shutter. I was also looking at the white smudges on the top of the caterpillars "head" in the first photo.
I missed the white smudges on the top of the caterpillar's head. (You're correct, it's not really his head. First thoracic segment.) Initially I thought that was just reflection from a shiny area. I still think that's part of the problem, but maybe something else is going on too. It's possible that the caterpillar moved its head while it was OOF, so that some of the bright blurs didn't stack out properly. That's another thing to check by scrolling through frames as a filmstrip.

By the way, when I try to shoot live bugs I very often see little movements in the stack that I surely did not notice while shooting. Dead bugs are a lot safer but even then not completely foolproof. One of my funniest stories (OK, maybe you had to be there...) is actually about my avatar. See 'The Answer is Blowing In The Wind'
lot of great insect/spider photos ... my conclusion is that the bug isn't moving.
Correct. The deep stacks are all immobile specimens, most dead. Occasionally something like a spider will sit still enough for a deep mechanical stack. (See 'Orb-weaving spider, up close and personal'.) But most stacks of live specimens are shallow, less than 10 frames, often more like 5.
New question - and might be out of my ability at this point. I hear alot of people describe their lenses as objectives on this site. Now I understand objective to be synonomous with Lens. However it is usually used with a 10x or 5x or something similar with it. So in this instance, are there special lenses that are classified as "objectives" and are they called objectives because they move beyond the realm of MacroPhotography into the the world of Microphotography? I see it alot with Nikon - is there a common set up for use with canon?
As ChrisR said, these are microscope objectives, usually on bellows, in front of any kind of camera. Nikon made a series of objectives designated "CF" (color free) because they have all of their aberration corrections built into the objective instead of relying on compensation in the eyepiece. Those objectives make superb high magnification macro lenses, but they also have very shallow depth of field -- something around 11 microns (0.011 mm) for a 10X NA 0.30 objective. So to use them you need some high precision focusing setup, typically based around either a screw table or a microscope focus block.

--Rik

thartl
Posts: 169
Joined: Wed Oct 28, 2009 8:07 pm
Location: Wyoming

Post by thartl »

Ok - I feel I am getting better. Thanks for sticking with me while I am working on improving. Some of my stacks have not come out very sharp, but I am posting some of the better ones. I am really proud of this first one, I think I stacked around 26 images.
Image
This flower (Gerbera?) was a larger stack, of 53. Much easier to do flowers than insects. Hahaha - not lit real well though.
Image
Finally this one, I am not sure on the flower, but it had an even larger stack, of 118, I think by far my largest stack, and yet the least amount of detail, but I think that was my subject.
Image

Does ZS accept camera RAW? I really like to shoot Raw.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

These look excellent. Attractive, no obvious artifacts. Well done.

No, ZS does not directly accept raw input. Shooting raw is fine. But then the recommended workflow is to batch convert raw to TIFF, using your favorite raw converter and whatever set of parameters make that converter produce the best result. Feed the TIFF files to ZS. When you're happy with the result, delete the TIFFs to save space. You can do this with either 8- or 16-bit TIFFs. Doing an explicit conversion sounds expensive, but in fact it's exactly the same processing that say Helicon Focus does behind the scenes. Explicit conversion takes more keystrokes, but lets you use a specialty raw converter with all the controls provided by that vendor.

--Rik

thartl
Posts: 169
Joined: Wed Oct 28, 2009 8:07 pm
Location: Wyoming

Post by thartl »

Rik I really appreciate the comments - thank you so much. The first few photos I felt it was important to work on focus and stacking, as I went on I felt I could look at composition and lighting more as I was getting that down too. I AM HAVING TOO MUCH FUN AREN'T I?

I have bellows coming on the way - not sure how well I will like them, as my extension tubes already put my 50mm lens almost right on top of the subject, perhaps my 135mm macro will need to be used on the bellows?

I also just purchased (didn't realize they were so inexpensive,) a reversing ring to mount my 50mm to my body, and then a filter ring to attach my 135mm face to the face of my 50mm. I will experiment with those a bit too when they arrive.

I currently have been using my portrait strobes to light, but I imaging by spring time I ought to invest in a ring flash. Are the BOWEN brand flashes worth it over the Canon?

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

thartl wrote:I currently have been using my portrait strobes to light, but I imagine by spring time I ought to invest in a ring flash.
I'd recommend spending your money on something else. Ring flash lighting tends to be overly flat and hard to control. Invest instead in a small collection of highly specialized and expensive diffusion devices like old styrofoam cups, white yogurt containers, and Kleenex tissue. Study the postings by NikonUser and Charles Krebs. They have taken pains to show us a wide variety of lighting schemes that are all very effective with small subjects, and scarcely a ring flash to be found. Consider getting a dual head fiber optic illuminator. They are very flexible and since they're continuous illumination, they're much easier to adjust for specific effects.
Are the BOWEN brand flashes worth it over the Canon?
I don't have a clue on that one. Probably somebody else does.

--Rik

PS. Looking again at your last set of pictures, I do see one obvious artifact. Third picture, center left, there's a "transparent foreground" petal. That's caused by the lens looking around the foreground petal when it is focused on stuff farther back. You end up getting detail at two different depths in the stack at the same pixel position. The stronger detail wins, and in this case that's on the farther back petal. Most probably it is not possible to eliminate that problem by any tuning of ZS parameters, but it is easily fixed by retouching within ZS.

thartl
Posts: 169
Joined: Wed Oct 28, 2009 8:07 pm
Location: Wyoming

Post by thartl »

Hahahaha - I get what you are saying. I will do some research around and see what I come up with for lighting. The dual head fiber optic illuminator sounds cool, does it attach to camera? Never mind, I will do the research.

Rik - can you elaborate on:
Most probably it is not possible to eliminate that problem by any tuning of ZS parameters, but it is easily fixed by retouching within ZS.
Also - I see a problem, and i am wondering if its dust on my sensor. (which would be wierd, my 5dmkii is not that old - and isn't suppose to be that susceptable to dust.) Look at the caterpillar (on the green flower.) To the left of the head, way back on the tail end of the caterpillar by where its legs would be, there is a thin black line almost looks to be made up of tiny dots. What would you think would cause that? (Easily repairable in photoshop - however - thats not the point.)

Here is a dead beetle I found. Same green flower (I like the color.) I am working on a deeper stack as I type. I also did some more of the caterpillar - but I am learning critters make very small movements that I can't see through the eye piece! This beetle was 1/200, f8, ISO 200 stack of 25.
Image
Also - I would love to be able to do my own species identification - and my searches online are fruitless. Any pointers to lead me in the right direction? A book? Experience? Classes? [/quote]

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

thartl wrote:Rik - can you elaborate on:
Most probably it is not possible to eliminate that problem by any tuning of ZS parameters, but it is easily fixed by retouching within ZS.
ZS has two stacking methods. For PMax, there are basically no options. DMap has three tuning parameters: two radii and a contrast threshold. When confronted with the "transparent foreground" artifact, it can be tempting to try tuning the radii to get rid of it. But in my experience that seldom works. It's quicker and more effective to just put the human in the loop via the retouching tool. At this time (meaning this decade and possibly the next), humans are capable of accurately figuring out the physical structure from the stack of images, while software is not.
Also - I see a problem, and i am wondering if its dust on my sensor. (which would be wierd, my 5dmkii is not that old - and isn't suppose to be that susceptable to dust.) Look at the caterpillar (on the green flower.) To the left of the head, way back on the tail end of the caterpillar by where its legs would be, there is a thin black line almost looks to be made up of tiny dots. What would you think would cause that?
Yes, that is a dust streak. It looks like there is another one, much fainter, extending to the left and up from near the caterpillar's tail end. If everything is locked down, and the focus movement is exactly along the optical axis, and the image itself is roughly symmetric, then the streaks are radial because they come solely from enlarging or shrinking frames to correct for changes in magnification across focus steps. If the focus movement is not exactly along the optical axis, then images will also shift horizontally or vertically, and then the streaks may not be radial. The alignment algorithm can also be misled somewhat by asymmetric patterns of light and dark causing apparent movement as they change focus.
Also - I would love to be able to do my own species identification - and my searches online are fruitless. Any pointers to lead me in the right direction? A book? Experience? Classes?
For some pointers about ID resources, see "FAQ: Where do I find ID guides and references for naming?" But please also read the philosophy written HERE and in the other posts in that topic. The bottom line is that identifying insects is challenging in the extreme. Often even a specialist cannot get closer than genus, sometimes no closer than family, without having the specimen in hand.

Perhaps the best resource at your disposal is the membership here at photomacrography.net. If you post a whole-body view, shot with a large f-number so that you get lots of DOF even if the image is not very sharp, then usually one of our entomologists can get pretty close.

--Rik

thartl
Posts: 169
Joined: Wed Oct 28, 2009 8:07 pm
Location: Wyoming

Post by thartl »

I took this intending to start a new thread - show off my new found skills - and I am a little frustrated. It isn't as sharp as I have been hoping for. I am wondering if my 50mm Lens is letting me down. (Canon 50mm f2.5 macro) It was definitely a cheaper piece of glass, but I don't think the 50mm f1.8 is a macro - although it probably would work with tubes as a macro? I am using a pretty sturdy tripod, but it is resting on carpet, could the shutter be causing vibrations? (At 1/200s?) Is it exposure? Is it becuase I am pushing the lens? I am shooting on a very sturdy tripod with pistol grip ball head, using a remote shutter release, and I was using strobes to light this.

Or is this a product of stacking? Or am I being too picky? It just isn't as sharp as I want. The subject was not alive.
Canon 5dmkII 50mm f2.8 + Kenko Tubes, 1/200s, f6.3, iso 200, stack of 73 images.
Image

100% crop of above stack

Image

This is a 100% crop of 1 image in the stack for comparison - if it helps?

Image

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

This looks pretty good, but let's see if we can make it better.

I'm presuming from your comments that this has not been sharpened in post-processing. Be aware that most of the high mag stacks you see here on photomacrography.net have been treated with "unsharp mask" (USM) or some similar sharpening method. My usual guidance is to sharpen as much as you can without adding artifacts or making the image misrepresent the subject.

Other thoughts...

The primary cause of sharpness loss in image processing is interpolation, also known as resampling. This can occur before, during, or after stacking.

Let's concentrate for a moment on stacking. To make stack frames line up properly, they have to be slightly increased or decreased in size. For illustration, suppose that in some frame you had captured a set of perfectly focused stripes, black/white/black/white, every other pixel, spanning 100 pixels, and now that frame needs to be aligned with another one in which the stripes were out-of-focus and spanned 105 pixels (5% scale change). To properly align the images, the software has to take those 100 pixels of perfectly focused stripes and resample them to span 105 pixels. Regardless of how the resampling is done, you can see that at some places in that 105 pixels span, the original pixels will line up almost perfectly with the new pixel positions, while at other places they'll be off by 1/2 pixel. Of course there's no way to represent black/white/black/white when you're off by 1/2 pixel. As a result, in this contrived example, what happens is that the stripes end up getting preserved well at some places and virtually eliminated (converted to gray/gray/gray/gray) at others.

In natural images, nothing this obvious turns up in the final image, but the same problem still occurs behind the scenes. As a result, detail in some locations gets softened, while detail in other locations does not. The variation from area to area can make it pretty challenging to figure out what's going on.

This loss of sharpness during stacking can be minimized by working with images that have as many pixels as possible. Continuing the previous example, if the images were twice as large to start with, then the resampling would have to turn 200 pixels of something like black/gray/white/gray into 210 pixels of something. The maximum misalignment in this case is still 1/2 pixel, so although the stripes get degraded at some places, they do remain visible as something like dark/light/light/dark instead of going to uniform gray as in the earlier case. Again, natural images won't show anything as obvious as this contrived example, but the result is still that more pixels preserves detail better.

I'm never sure what "100% crop" means, so I did some quick calculations based on the images you posted. Working from the sizes of the first and second images as posted, I get that the second image is an actual-pixels chunk of a roughly 3000x2000 pixel image. That's interesting because as I read the specs, 3000x2000 is not close to a size that is provided natively by the 5D Mark II. As a result, I'm thinking the images might have been resized even before they got to ZS. That raises a pink flag.

The summary is that resampling softens, and the more pixels the better. Of course more pixels also consume more disk space and take more time to process, so there's a tradeoff. For highest quality, go big; for best throughput, go as small as you can while still getting adequate quality.

Two other major causes of softening are lens aberrations and diffraction. Generally speaking, as you stop down a lens you reduce aberration blurring but increase diffraction blurring. Most lenses start off being soft from aberrations when they're wide open, go to being soft from diffraction when they're stopped way down, and become sharpest at some intermediate aperture. The only reliable way to determine the sharpest aperture is by experiment. Top end specialized macro lenses are sometimes best wide open, but with most lenses the sharpest point will be 2-3 stops down from wide open. Using an inexpensive f/2.8 at f/6.3 is probably close to optimum for that lens.

Optical blurring due to lens aberrations and diffraction can be compensated to some extent by sharpening, again by USM or the equivalent. If you're familiar with the concept of MTF (modulation transfer function), what you're doing with sharpening is to push up the high-frequency end of the MTF curve so that it stays flatter for finer detail. Sharpening can't change the "cutoff frequency" where detail disappears, but it can make significant improvements below that.

As you've suggested, vibration can also cause blurring. However, I don't think you're seeing that in these images. One reason is that you're using flash illumination, so your effective exposure time is less than 1/1000 second, possibly much less if your strobes are the common type that reduce output by reducing duration. The other reason is that when I look closely at your images, the smallest point details look circular. Motion blur usually smears details in one direction but not the other, so that the finest details turn into little ovals.

One last thing to consider is illumination. The setup that you're currently using doesn't seem to include much diffusion. More diffusion would probably give better results. There are several different reasons.

One of them is that undiffused illumination produces highlights in the form of small hard-edged specular reflections. Those sorts of highlights often don't play nicely with stacking. Sometimes they move around a little as you change focus. This can cause a slightly smeared appearance of the highlight even though detail around it remains sharp. They also seem to play badly with resampling, though I have trouble describing exactly how. And they play badly with sharpening, causing visible artifacts at levels of sharpening that would otherwise make the image look better.

In addition, adding some diffusion will put more light into the shadow areas as ChrisR mentioned. It will also provide more opportunity for shiny surfaces to create reflections that show off surface texture. See HERE for an example of a complicated structure that shows texture and shape even though it's very shiny. This latter aspect is not so important for your spider, since only its eyes are shiny and they will be very smooth. But for the caterpillar, more diffuse lighting would have given a better rendering of the head.

Bottom line, I think you're seeing a combination of effects: working on the edge of what your lens can do; suffering a bit from resampling; a bit more from illumination; and a final bit by not sharpening in post-processing. I very much doubt that there is any one thing you can do to dramatically improve your results, but doing several things to attack all these aspects might make you quite a bit happier.

--Rik

thartl
Posts: 169
Joined: Wed Oct 28, 2009 8:07 pm
Location: Wyoming

Post by thartl »

WOW! I will need to re-read - again to make sure I get it all in!
. The setup that you're currently using doesn't seem to include much diffusion. More diffusion would probably give better results. There are several different reasons.


Well I am using alien bees - one B400 and two B800s - all with soft boxes and then again bouncing the flash off wall/cieling. I don't have the specs of the softboxes where I am at now- nor did I measure the distance from subject. What I found as that I wasn't getting enough light coming from the front - and I THINK that might be my problem?

I actually think you are on to something. I might dial down my shutter speed and dial down the strobe power, they were on about 1/2 wattage +or - a few. Maybe thats the problem - and maybe all the light bouncing around the room in a variety of ways is flattening images. I certainly wouldn't shoot a portrait (people,) this way.
I'm never sure what "100% crop" means, so I did some quick calculations based on the images you posted. Working from the sizes of the first and second images as posted, I get that the second image is an actual-pixels chunk of a roughly 3000x2000 pixel image. That's interesting because as I read the specs, 3000x2000 is not close to a size that is provided natively by the 5D Mark II. As a result, I'm thinking the images might have been resized even before they got to ZS. That raises a pink flag.
100% crop to me is viewing your file at actual pixels and cropping that portion out. I do not resize my photos before zs. It is LARGEest JPG files straight from the 5dmkii. Then once that is done I keep the original file (which I could post for you but don't have it here,) but crop to around 8x12 at 300 dpi for printing purposes (habit sorry.) The file above then was resized again for web. This however, I pre checked before submitting. I compared the quality of the original zs image to my cropped image and they look identical. But as I said, I do not crop before placing in ZS.
My usual guidance is to sharpen as much as you can without adding artifacts or making the image misrepresent the subject.
I don't do any sharpening - I am not a fan of it. I almost always feel there are artifacts with any sharpening. I did however have a level of underexposure on this particular shot, so I used curves and levels to adjust, and added a touch of saturation. This I thought caused some of the artifacts, but again looked at the output image original from ZS and there was little to no difference. Improper exposure will of course account for some noise........

I will re-read (and probably read again) your post and try some more later. Thanks a bunch again!

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Yeah, I write gobs of words, but I try to pack in a lot of information also. When I go back to read my own stuff, it usually takes several readings before I get it all again.

About the illumination, I'm looking at the spider's eyes. In each eye, there are two fairly small bright reflections and the rest of the eye is essentially black. What that says is that almost all the light is coming from two sources that cover a fairly small solid angle from the standpoint of the spider. Compare with the third image in MarkB's spider HERE or the second image in mine HERE. In those images, the reflections cover more of the eye and are not so bright, indicating light sources that cover a larger angle.

I don't know exactly why it is, but diffuse wide angle illumination gets more and more important at higher magnifications. I agree completely, illumination that works well for a spider or a fruit fly would be way too flat if it were just scaled up to take a human portrait. By the same token, illumination that would be great for the human portrait seems to end up too directional for tiny subjects. A lot of the high mag stacks shown in these forums were shot with essentially wrap-around diffusion consisting of half a ping-pong ball, or a styrofoam cup, or part of a yogurt container, or some tissue paper wrapped completely around the subject. As an experiment, try adding way too much diffusion and then backing off, to see what works best for you.

About the sharpening, I understand what you're saying, but I should mention that this is another area where extreme macro differs from human-scale photography. With macro at these scales, it's hard to not be working at the limits of the lens -- the fine detail you'd like to see is way out on the end of the MTF curves. When people talk about line pairs per mm at the diffraction limit of a lens, what they're talking about is detail whose contrast has been diminished to only 10%, sometimes only 1%, of what was present in the original subject. The contrast of detail that is coarser than that limit is still degraded, just not so much. Even under ideal conditions, detail twice the size of the limit loses about 50% of its contrast and detail four times the limit loses 25%. See the graph HERE. This loss of contrast for the higher frequencies translates into visual softening. Careful sharpening can restore much of the original contrast for those higher frequencies, without introducing artifacts that are too painful. So it's a balancing act -- correct the softening that was caused by optics, without introducing painful artifacts. One of the effects of more diffuse illumination is to allow better correction of the softening by allowing stronger sharpening with acceptable artifacts.

Different people prefer different styles of result, and that's as it should be. I'm just trying to explain how some of these more subtle aspects relate to each other so that you can experiment efficiently to get the result you want.

--Rik

Harold Gough
Posts: 5786
Joined: Sun Mar 09, 2008 2:17 am
Location: Reading, Berkshire, England

Post by Harold Gough »

rjlittlefield wrote:
Also - I would love to be able to do my own species identification - and my searches online are fruitless. Any pointers to lead me in the right direction? A book? Experience? Classes?
For some pointers about ID resources, see "FAQ: Where do I find ID guides and references for naming?" But please also read the philosophy written HERE and in the other posts in that topic. The bottom line is that identifying insects is challenging in the extreme. Often even a specialist cannot get closer than genus, sometimes no closer than family, without having the specimen in hand.

Perhaps the best resource at your disposal is the membership here at photomacrography.net. If you post a whole-body view, shot with a large f-number so that you get lots of DOF even if the image is not very sharp, then usually one of our entomologists can get pretty close.
Having spent my adult life needing to/curious to identify various arthropod/plant/fungal species, I agree with Rik's comments, in particular that identification to family is as much as a non-specialist, or even a specialist, can hope to do for little-known species or groups.

I would add that one image may be insufficient for identification, as this usually requires viewing the specimen from various angles. At the very least, include the whole specimen, not just a 'head and shoulders'
My images are a medium for sharing some of my experiences: they are not me.

AndrewC
Posts: 1436
Joined: Thu Feb 14, 2008 10:05 am
Location: Belgium
Contact:

Post by AndrewC »

thartl wrote: I don't do any sharpening - I am not a fan of it. ...
You might want to revisit that position, as 99.9% of the world does sharpen digital images. You have to. This is a good intro as to why you need to sharpen. There are many, many words out there about the do's and don'ts but very few about the "not at all"

http://www.bythom.com/sharpening.htm

What do you convert your images with ? Are you sure there is no sharpening happening ? Once you resize you need to sharpen again.


rgds,

Andrew

Cyclops
Posts: 3084
Joined: Sat Aug 05, 2006 5:18 pm
Location: North East of England
Contact:

Post by Cyclops »

thartl wrote:P.S. - I am not sure what species it is, and if you all have a website you use to look these up I'd be open to suggestions!
Very nice shots, I'm a bit of a spider fan myself, and this one looks like a wolf spider (Lycosidae) but I cant tell how big the spider is.
Personally I like the second picture, its fabulously done and doesnt need stacking!

I know of a few spider websites to check out
http://s714.photobucket.com/albums/ww14 ... s/Hunting/
This comes from the UK Wildlife site:
http://greatbritishwildlife.tripod.com/id231.html

Also theres
http://www.eurospiders.com/
http://arachnophiliac.info/burrow/spiders.htm
You could always join this spider forumfor info
http://www.arachnophiles.co.uk/forum/
Also theres my own spioder blog here
http://spidersamongstus.blogspot.com/
Canon 5D and 30D | Canon IXUS 265HS | Cosina 100mm f3.5 macro | EF 75-300 f4.5-5.6 USM III | EF 50 f1.8 II | Slik 88 tripod | Apex Practicioner monocular microscope

Post Reply Previous topicNext topic