Introduction and some initial questions

[stevekale]

Hi there. First post and hopefully in the right place.

I stumbled across this forum again while searching for methods to keep bugs (in my case ephemeroptera) still while photograhing them. This is a tremendous resource and I am stunned by the quality of the images here.

I have had an interest in photography for many years (landscape and fashion) but have only recently ventured into macro and stacking in particular. Awhile ago I bought Helicon Focus - sorry I didn't even know about Zerene - and started to play around with it. Success was average and I realised I needed to learn more, control the environment more - including eliminating bug movement which brought me here. (I've raided my wife's ethyl acetate nail polish remover!) My interest in macro started with fly fishing, initially to photograph the fishing flies I was tying. From there to fishing-relevant entomology and on to photographing the bugs that the trout were eating.

My current camera is a Canon 1Ds II and I have been using the EF100mm 2.8L macro lens. Lighting has been the Canon 580EX Speedlite with a Lastolite Micro Apollo soft box hood, used off-camera and controlled with the ST-E2 transmitter.

I've come to the conclusion I need to use a rail and I'm looking at Stackshot. I admire those that have the skill - and, importantly, the time - to make their own. I hear the Stackshot gets a great rap. Right?

I have also been occasionally looking at the Canon MP-E 65mm lens but have decided I can do much better with existing lens first or should at least ask if some of my existing lens can be utilised. The only other Canon prime lens I have is the 135mm f2.0L, but I was wondering if some of my old Contax 645 lenses could be put to work:

f2.0 80mm Planar
f4.0 120mm Apo-Makro-Planar

or even my old Carl Zeiss Contax G2 lenses:

f2.0/45mm Planar
f2.8/90mm Sonnar

Any use?

Some other questions:

- can someone point me to a tool for determining depth of field for the Canon 100m L macro lens? (preferably one that I will be able to understand)

- how long a settling time do you use between shots or rather what would you recommend with the Canon camera/lens? (I use mirror lock)

BTW I had lots of issues with the 32 bit version of Helicon Focus crashing when editing. They were helpful enough to give me a 64bit license until things are sorted. (I have a Mac Pro 2 x 2.4 Ghz Quad Core Intel Xeon with 8 GB of memory.)

Anyway sorry for the ramble. I guess I am just interested in where to head to next....

Looking forward to learning from you guys!

Steve

[rjlittlefield]

Steve, welcome aboard!

StackShot is an excellent tool. I use it regularly at all magnifications from well below 1:1 up to 40X on sensor. See the range of examples from HERE to HERE.

I assume that you're speaking of the Canon EF 100mm f/2.8L IS USM Macro lens. A recent review in Advanced Photographer of "seven magnificent top-end macro lenses" ended up rating it as "Best In Test". I had been considering that lens for some time, and that review pushed me over the edge. The lens came a few days ago. I haven't done extremely rigorous testing of it yet, but on quick QA the images look awesome.

You're correct that the MP-E 65 is also a great lens. It's very hard to beat in the range of 1-5X even on benchtop, and impossible to beat for field work where you need an auto diaphragm.

However, the MP-E is also pretty expensive, and for bench work you can get almost the same image quality for much less money by reversing certain other lenses on tubes or bellows. The lens most commonly recommended for this purpose is the 50 mm f/2.8 EL Nikkor enlarging lens. The recommendation is based on its being commonly available and "known good". Numerous other lenses also work well in this mode. In your shoes I would definitely give the f2.0/45 mm Planar a shot.

Your question about depth of field deserves a bit of discussion.

You probably know that diffraction causes a fundamental tradeoff between DOF and sharpness. Stopping down gives more DOF due to reducing the geometric blur circle, but it also increases diffraction blur. At some point, the diffraction blur becomes as large as you're willing to tolerate, and that's the limit of DOF in a single shot.

Stacking allows you to get more DOF at any specified aperture, but you still need to pick an aperture that's appropriate to your application. Then you can determine a corresponding focus step that minimizes the number of frames you have to shoot, while still avoiding focus banding. By far the most reliable way to do both of these is by experiment, not calculation.

Start by shooting a set of aperture tests at the magnification you want. Look through those to determine the smallest aperture (biggest f-number) that is sharp enough to keep you happy. It's fine to go for maximum sharpness (that's what I usually do), but be aware that maximum sharpness also means largest number of frames.

After picking an aperture, then shoot a short test stack with very fine spacing, say 5X smaller than you think you might need based on DOF calculation. Process the stack several times, first doing every frame, then every 2nd frame, every 3rd, and so on, until you find the maximum separation that gives you a clean image without focus banding.

Now you know one good set of numbers: magnification, f-number, and step size. Write those down. Repeat for several different magnifications, and build yourself a table. Feel free to draw curves to interpolate between measurements -- the whole process is rough enough that interpolation won't add any significant problems.

As a guideline, here are some numbers that roughly apply to your situation. This uses a pretty small COC, corresponding to shooting for an image that's sharp when pixel-peeping. Notice that these numbers are based on maintaining a constant effective f-number, taking into account magnification. This corresponds to getting constant sharpness by opening the aperture at higher magnifications to reduce diffraction.

Code: Select all

Sensor Width (mm)              36		
CoC (mm)                     0.024
Desired Effective f-number     16			

Subject 
Width          Nominal
(mm)    Mag.   f-number   DOF (mm)	
				
200.00  0.18    13.6       23.7
168.18  0.21    13.2       16.8
141.42  0.25    12.8       11.9
118.92  0.30    12.3       8.38
100.00  0.36    11.8       5.93
 84.09  0.43    11.2       4.19
 70.71  0.51    10.6       2.96
 59.46  0.61    10.0       2.10
 50.00  0.72     9.3       1.48
 42.04  0.86     8.6       1.05
 35.36  1.02     7.9       0.74
 29.73  1.21     7.2       0.52
 25.00  1.44     6.6       0.37
 21.02  1.71     5.9       0.26
 17.68  2.04     5.3       0.19
 14.87  2.42     4.7       0.13
 12.50  2.88     4.1       0.09
 

I'm sorry to hear about your difficulties with Helicon Focus. We hear lots of comments about halos, but I think this is the first I've heard about crashes while retouching. I'm sure they'll get that fixed.

The main reason Zerene Stacker is popular around here is that it's specifically designed for dealing with deep high magnification stacks, so it has stacking methods and retouching tools that are optimized for that use. The support is also outstanding, if I may say so myself. (I'm the fellow who wrote and supports Zerene Stacker, pleased to make your acquaintance!) ZS has a 30-day free trial with no restrictions, and retouching is included at every price level. Website is http://zerenesystems.com/. It will run very well on your Mac.

I usually use 3 seconds settling time. As usual, test your own setup to see what's needed.

Hope this helps!

--Rik

[stevekale]

Thanks Rik. Yes I realised you are the developer of Zerene - my apologies for not having done a thorough search! Now I have to get HF to work. My first attempts at photographing fishing flies with all their layers of wound herl weren't very successful. I had thought it was because I wasn't getting even and fine enough frames but will have to investigate further.

I am aware of the trade off between depth of field diffraction although not at the extreme technical level it gets discussed here. Circles of Confusion etc are terms I need to brush up on. But generally I'd like to shoot for best stacked resolution and only compromise if I run out of CPU/memory (at least for controlled stuff). I "get" the need for f-stop sharpness tests although I suspect this has been done many times for this lens in various reviews etc.

I also need to get my head around how you guys talk about magnification and how it alters when I twist the focus ring although I guess this will disappear once I move to shifting the object distance. With my very basic setup (EF100mm 2.8L USM lens), I would be trying to capture the best image of, say, a Blue Wing Olive (Seratella ignita). They are about 6-8mm in body length and height. With the lens set for closest distance I have 1:1 on the sensor and the BWO fill 1/3 of the frame height. To get better magnification (in print or digital display) I have to rely on pixel density and crop down the image. Because currently subject distance is not being altered and instead I am using the focus ring, as I step from front to back I slip below 1:1 magnification.

So with that ramble in mind, how do I read your table above? I am shooting 1:1 or a bit below (constrained by lens and static subject distance) and want deep DOF across the whole compiled image (let's ignore artistic factors here for now - I recognise the benefits of shallow DOF in image composition). By "effective f-number" do you mean to capture on a composite basis the DOF that would have been there had I simply shot at that f-stop? I also don't see how to relate the magnification in your table to the choice of f-stop. I suspect I am reading across the table when I shouldn't and that the first two columns are independent of the second two but want to make sure.

BTW if I recall correctly, the Contax G2 lens, while regarded as some of the finest lenses made for 35mm film, are problematic because they can't be manually focused. Perhaps this is a problem for using them on another camera but not for bellows-macro where focus is managed by subject to lens distance?

So what bellows? (One thing I do fear about bellows and all the thread adaptors (and time) etc is that ultimately it may be cheaper to bite the bullet on the MP-E. It's not cheap but then neither was the 100mm 2.8L Macro and neither is Stackshot.) I have read here that some people recommend the Pentax M42 Auto Bellows but I assume that needs a coupling mount to fit a Canon DSLR.

There also seem to be several versions out there including

http://cgi.ebay.co.uk/Pentax-Auto-Bello ... 423wt_1033

and the "Asahi Pentax Macro Bellows II" version which has a dual rod rail rather than the single tongue and groove rail. Does it matter?

Any recommendations on how to rig up the Contax lens in reverse to the 1Ds II? You would use bellows AND a reversed lens?

Thanks and I hope these questions are too basic!

BTW I found a link to this calculator on the Stackshot website

http://eosdoc.com/jlcalc/

I couldn't get the focal length setting at 100mm to generate a Magmax of 1x with a min focusing distance of 0.3 metres.....

[ChrisLilley]

Steve, your Contax/Zeiss lenses are of very high quality and if you can use them on your Canon SLR with adapters I encourage you to do so. The register distance for Contax/Yashica mount is larger than Canon mount, so adapters are available. And clearly there is no problem with a Contax645 to Canon adapter, plenty of space there.

As to macro use, I suspect that the 120/4 apo-makro-planar would make a fine tube lens for the Nikon 10x CFI objective, giving you a high quality 6x. The 45/2 might also be interesting reversed on extension tubes or bellows.

[ChrisLilley]

I am aware of the trade off between depth of field diffraction although not at the extreme technical level it gets discussed here. Circles of Confusion etc are terms I need to brush up on.

In approximate terms, the circle of confusion is an area that is indistinguishable from being a single point. Its a way to express the smallest detail that can be seen or captured. Its complicated because there are varying assumptions about viewing distance, visual acuity, and what is 'good enough' and these give different values for CoC. Smaller values imply higher quality.

I also need to get my head around how you guys talk about magnification and how it alters when I twist the focus ring

Magnification, as discussed here, is very simple. Its the size in mm of the image on the sensor divided by the size in mm of the object. If a 2mm object forms an image 10mm across on the sensor, that is 5x.

Additional magnification when the image is developed and printed or viewed onscreen is not considered.

With my very basic setup (EF100mm 2.8L USM lens), I would be trying to capture the best image of, say, a Blue Wing Olive (Seratella ignita). They are about 6-8mm in body length and height. With the lens set for closest distance I have 1:1 on the sensor and the BWO fill 1/3 of the frame height.

.

Yes. To fill a 24x36 sensor with an object of height 8mm you would need 24/8 = 3x magnification, so your report of 1/3 of the frame height is spot on.

To get better magnification (in print or digital display) I have to rely on pixel density and crop down the image.

In the terms of purely optical magnification, such cropping does not change the magnification.

(let's ignore artistic factors here for now - I recognise the benefits of shallow DOF in image composition).

At normal shooting distances, shallow DOF can be a plus as it isolates the subject from the background (and impresses other photographers). At macro distances, shallow DOF means that only 1% or less of your subject is in focus. This is a drawback, not a benefit.

By "effective f-number" do you mean to capture on a composite basis the DOF that would have been there had I simply shot at that f-stop? I also don't see how to relate the magnification in your table to the choice of f-stop.

I suspect that what you are missing is that the actual f-stop (the actual size of the aperture, and thus the amount of light hitting the sensor and the degree of diffraction) depends on magnification. At non-macro shooting distances, this effect is very small and can be ignored. So the effective f-stop and the nominal f-stop as indicated by the aperture ring on your lens are identical (at infinity) or very similar (at 'close' shooting distances where the magnification is, say, 0.1x or 0.2x).

Representing the effective f-stop as E and the nominal F stop as F, the actual equation is E = F(1+m).

So for the situation of a telephoto lens used for portraiture at its closest focus distance, say, the magnification is likely to be around 0.15 and thus, if the nominal f-stop is f/5.6 the effective f-stop is 5.6(1+0.15) = f/6.4

Once you get to 1:1, though, the effect is significant. At 1:1 with a nominal aperture of f/8 the effective aperture is 8(1+1) = f/16 and at higher magnifications, it is overwhelming. At 10x and with a lens used at f/2.8 the effective aperture is f/30.8 which will be strongly influenced by diffraction.

Once that is clear, Rik's table is running those equations backwards - to get an effective f-stop of f/16 (so hopefully, not affected by diffraction too much)what nominal f-stop will you need at various magnifications? And what DOF will you get? As you will notice, as the magnification goes up then (to hold to a usable f/16 effective) the nominal aperture gets wider and wider and the DOF gets shallower and shallower.

[stevekale]

Okay. A little slower so that I understand the terminology. I assume you are first referring to the 645 lenses and fitting them directly to the Canon body. (I've not seen any mounts for the G2 lenses to a Canon although I did explore putting them on a Lumix 4/3 but was put off by the need for a focusing ring adaptor.) I didn't think this was possible without some very expensive adaptors which maintain the electronics to enable f-stop management - the only ones I've heard of are lens mods done by conurus.

Can you point me to some reading that will get me to understand exactly what you mean by "would make a fine tube tens"?

What adaptors would be necessary to reverse the G2 Zeiss 45/2 with a 46mm thread onto a set of bellows?

The (microscope?) Nikon lens you reference seems to be very expensive, although I may be confusing models...

[stevekale]

I suspect that what you are missing is that the actual f-stop (the actual size of the aperture, and thus the amount of light hitting the sensor and the degree of diffraction) depends on magnification. At non-macro shooting distances, this effect is very small and can be ignored. So the effective f-stop and the nominal f-stop as indicated by the aperture ring on your lens are identical (at infinity) or very similar (at 'close' shooting distances where the magnification is, say, 0.1x or 0.2x).

Representing the effective f-stop as E and the nominal F stop as F, the actual equation is E = F(1+m).

So for the situation of a telephoto lens used for portraiture at its closest focus distance, say, the magnification is likely to be around 0.15 and thus, if the nominal f-stop is f/5.6 the effective f-stop is 5.6(1+0.15) = f/6.4

Once you get to 1:1, though, the effect is significant. At 1:1 with a nominal aperture of f/8 the effective aperture is 8(1+1) = f/16 and at higher magnifications, it is overwhelming. At 10x and with a lens used at f/2.8 the effective aperture is f/30.8 which will be strongly influenced by diffraction.

Once that is clear, Rik's table is running those equations backwards - to get an effective f-stop of f/16 (so hopefully, not affected by diffraction too much)what nominal f-stop will you need at various magnifications? And what DOF will you get? As you will notice, as the magnification goes up then (to hold to a usable f/16 effective) the nominal aperture gets wider and wider and the DOF gets shallower and shallower.

You are right. Was not conscious of this. I think I get it. I'll worry about the why later. What is clear is that my attempts at stacks close to 1:1 magnification at a nominal f-stop of f5.6 and even f8 were likely in no where near fine enough increments.

[stevekale]

Any reason not to get the longer Stackshot rail? (Other than the extra $75. That's a lot cheaper than one day realising you want the longer rail and are staring at $375.)

[ChrisLilley]

Okay. A little slower so that I understand the terminology. I assume you are first referring to the 645 lenses and fitting them directly to the Canon body.

Yes.

(I've not seen any mounts for the G2 lenses to a Canon although I did explore putting them on a Lumix 4/3 but was put off by the need for a focusing ring adaptor.) I didn't think this was possible without some very expensive adaptors which maintain the electronics to enable f-stop management - the only ones I've heard of are lens mods done by conurus.

My apologies. I had missed that you were using the Contax G2 system; I had responded thinking you were using the Contax/Yashica system, whose fine Zeiss lenses can indeed be used on a modern Canon DSLR with a suitable adapter.

Can you point me to some reading that will get me to understand exactly what you mean by "would make a fine tube tens"?

Sorry for not explaining. Please read FAQ: How can I hook a microscope objective to my camera? where the use of 'infinity' objectives with a tube lens is explained. Although designed for a 200mm tube lens, good results have been reported (at lower magnifications) down to 100mm.

The (microscope?) Nikon lens you reference seems to be very expensive, although I may be confusing models...

Mine was 95 EUR (130 USD) in January of this year. Others have got them cheaper, USD 65 for example, but prices are rising due to reports on this forum of how good they are.

[ChrisLilley]

Any reason not to get the longer Stackshot rail? (Other than the extra $75. That's a lot cheaper than one day realising you want the longer rail and are staring at $375.)

Not that I can think of. The new long rail uses the same pitch of screw as the shorter one, so they have the same minimum step size. Larger size and higher weight are considerations for portable use, but not for studio use.

Be sure to get the USB version of the controller (this is why), and get the optional arca-swiss top plate as the standard one is, not to put too fine a point on it, rubbish.

[rjlittlefield]

BTW I found a link to this calculator on the Stackshot website

http://eosdoc.com/jlcalc/

I couldn't get the focal length setting at 100mm to generate a Magmax of 1x with a min focusing distance of 0.3 metres.....

As I read the thread, ChrisLilley has answered all your other questions. (Thanks, Chris!)

Your problem with the eosdoc calculator lies partly in those "unstated assumptions" that I often caution about.

When a lens manufacturer quotes the "minimum focusing distance" of a lens, they're generally talking about the total distance from sensor to subject. That's the standard specification, even though it's pretty much useless in most cases. In terms of the optics, that "minimum focusing distance" is actually the sum of three other distances: subject to first principal plane, first principal plane to second principal plane, and second principal plane to sensor. In contrast, the eosdoc calculator measures focus distance "from the front nodal point of the lens." The front nodal point is located at the first principal plane, so the number the calculator is using is actually just the first number that goes into Canon's advertised value of 300 mm for the 100 mm f/2.8L.

Once you know about this difference, it's easy enough to plug values into the calculator to make it do what you want, kinda sorta.

According to the basic lens equation that 1/f = 1/o + 1/i, a 100 mm lens will reach 1:1 when its first principal plane (front nodal point) is 200 mm away from the subject and its second principal plane (rear nodal point) is 200 mm away from the sensor.

So, plug into the eosdoc calculator the following numbers:
Focal length = 100 mm
Closest focus = 0.2 m
Focus = 0.2 m

When you do this, you'll see that the field of view becomes equal to 24 mm x 36 mm, same as your sensor size, and the magnification ratio becomes 1.

Now you have to pick a value for Circle of Confusion. The calculator defaults to 0.025 mm, pretty close to the value of 0.024 that I used.

In the DOF table, we find the line that says A_set = 8 (f/8 set on the lens). In that same line, A_eff = 16 (f/16 effective aperture) and DOF = 0.4 mm.

Now in an ideal world, the DOF computed by eosdoc's jcalc would exactly match the DOF that appears in my table. But in fact, my calculation produces the value 0.768 . (The closest line shown in my table above is for m=1.02 and f/7.9, giving a calculated DOF of 0.74.) Obviously 0.4 is not the same as 0.768, even after some pretty serious rounding.

Why does the eosdoc calculator give such a different value from mine? That's a really good question, and I don't know the answer!

The calculation I'm using is pretty simple: 2*C*f*(m+1)/(m^2), straight from the literature ["The Manual of Close-Up Photography", Lester Lefkowitz, pg. 258]. Run the numbers with C=0.024, f=8, and m=1, and you get 0.768. That number is consistent with the value of 0.96 that Lefkowitz shows on page 27 using C = 0.03 mm, it's consistent with the value of 0.1 cm computed by http://www.dofmaster.com/dofjs.html, also using C = 0.03, and it's consistent with the value of 0.00096 meters computed by http://www.bobatkins.com/photography/te ... .html#calc, also with C=0.03.

I have no idea what calculation the eosdoc calculator is doing. I thought it might be reporting one-sided DOF, but that's not it because it's very specific that with nominal focus at 200 mm, the near limit is at 199.8 mm. I thought the eosdoc calculator might be factoring in diffraction somehow, but I can't see any trace of that in the Javascript code. The eosdoc calculator also disagrees with the others by substantial amounts even at very non-macro distances, such as 10 meters at f/8 where eosdoc says 3.9 m, dofmaster says 5.04, and bobatkins says 5.036.

This is pretty frustrating. My best guess is that the eosdoc calculator is just plain buggy, but I can't take it any farther than that. I'm really not interested in debugging somebody else's code --- it's troublesome enough to do my own. The one thing that is clear is that I sure wouldn't trust the eosdoc numbers for anything more than order-of-magnitude guidance.

Of course, to be quite honest I don't trust any of these numbers for much more than order-of-magnitude guidance in the macro regime. It's just too easy to get tripped up by hidden factors, even if there's no outright error.

I hope this helps, somehow or other.

--Rik

[Will Milne]

Hi Steve and welcome.

I have done a fair amount of work in the field you describe . Fly fishing related entomology .

For field work my own approach has been to use manual lenses and extension tubes to maintain portablity when I want to get beyond 1:1 . Including a rotable 8 degree shift extension tube that allows the plane of focus to be better aligned with the subject .

In the studio a few minutes in the fridge or simply covering the bug with a dark container keeps them quiet enough to work with.

http://www.watersidemb.ca/phento1.html

I save the stacking for microscope work. There are a couple of folks here who have done hand held stacking in the field , but I have never had success with it.

hope that helps

[stevekale]

My apologies. I had missed that you were using the Contax G2 system; I had responded thinking you were using the Contax/Yashica system, whose fine Zeiss lenses can indeed be used on a modern Canon DSLR with a suitable adapter.

The Contax G2 lenses are just ones I have lying around since I switched to digital. Same with the Contax 645 lenses. One can attach a 645 lens to a Canon EOS by buying a NAM-1 adaptor and then either having Conurus adapt it for the EOS mount or by slightly modifying it and adding a Contax C/Y to EOS adaptor in which case the 645 body is needed to set aperture (not such a big deal for studio work as I have that). The problem is, the NAM-1 adaptor is expensive (c$500) and with option 1 Conurus charge a bomb.

I wouldn't mind trying to reverse the G2 45/2 but I need to figure out all the needed adaptors etc. The only 46mm reversing ring adaptors I have found are for Micro 4/3 or Miranda bodies (whatever they are). For example, if I bought a Pentax bellows that people seem to think is good, how would I mount the 45/2? Any ideas?

Mine was 95 EUR (130 USD) in January of this year. Others have got them cheaper, USD 65 for example, but prices are rising due to reports on this forum of how good they are.

Ok I was clearly looking at the wrong lens as it was over a $1300.

Thanks for the Stackshot guidance Chris and the link to the sticky on attaching microscope objectives to a camera. Which is preferred? Finite on a set of bellows (or extension tubes) or an infinite objective on lens? I would have thought the former easier. Also, wouldn't using the Canon 135mm f2.0L be better if I went the the latter route in that all metering functions etc would be retained?

Cheers Rik, got it.

Will, thanks a lot. Some nice photos on your website. I don't lug my camera gear to the river - I am unable to stop fishing long enough to sit down and take photos. I just grab a sample or two and then keep fishing!

I've tried tried the "overnight in the fridge technique" but found it unsatisfactory for stacking. Wondering how to reduce or eliminate bug movement is what brought me here.

[ChrisLilley]

Thanks for the Stackshot guidance Chris and the link to the sticky on attaching microscope objectives to a camera. Which is preferred? Finite on a set of bellows (or extension tubes) or an infinite objective on lens? I would have thought the former easier.

The fiinite approach is easier, but the infinity objectives are more flexible because they can be used with a range of tube lenses to give different magnifications. For example the Nikon 10x can be used down to 5x. the Mitutoyo 10x is reported to work well between 14x and 7x.

Also, wouldn't using the Canon 135mm f2.0L be better if I went the the latter route in that all metering functions etc would be retained?

Yes, it would be easier.

[stevekale]

I assume that's an "easier but still workable". There's a bewildering array of Nikon CFI objectives. Am I right in saying:

CFI Achromat series = okay
CFI Achromat for Phase Contrast series = marginally better than above (but not worried about non-visible spectrum)
CFI Plan Achromat Series = considerably better than the above
CFI Plan Achromat Series for Phase Contrast = just as good but likely paying for something not needed
The others are either stratospherically expensive or aren't appropriate.