While I'm normally not a testing geek - I'd rather just be out taking photographs - I realize it's important to become familiar with my new equipment in a relatively benign environment before I venture out into the field. In several posts here people have made reference to testing various combinations of tripods/cameras/lenses/techniques. However in searching the archives I can't find any kind of a test routine (other than the Markins whitepaper, and some esoteric quantum physics experiments using linear accelerators and photon cannons). What do you use to hone your skills and test the efficiency on actual images, rather than oscilloscopes and accelerometers.
#1. "RE: Tripod Test Regime" | In response to Reply # 0Mark V Registered since 18th Jun 2004Thu 03-Apr-08 08:03 PM | edited Thu 03-Apr-08 08:16 PM by Mark V
<Markins whitepaper, and some esoteric quantum physics experiments using linear accelerators and photon cannons.>
I can't spell that stuff so I stick with good old fashioned kites and shovels. I programed my "cyborg" to open the tripod, grab the legs about 10" below the spyder and twist. The less twist the better.
Edit: If anyone runs into a giant made from a variety of cadavers and Series 5 Gitzos ....I know nothing...nothing!
Edit 2: But, if you know where I can buy one of those lab coats with the high collar that buttons along the side, please let me know.
#2. "RE: Tripod Test Regime" | In response to Reply # 0nrothschild Registered since 25th Jul 2004Thu 03-Apr-08 10:15 PM | edited Fri 04-Apr-08 05:20 AM by jrp
A lot of that may have come from me... I shoot a standard test target with horizontal and vertical lines. I always shoot 1/10s because I have found that that is most likely the most difficult shutter speed, and if it isn't the worst, it is close enough. You can obviously test other shutter speeds and you should, just to prove that to yourself. Never trust anyone's opinion on vibration issues- not even mine
I print the targets on a laser jet (because that is my main printer) with standard 8.5x11" copy paper. I orient the paper in landscape and frame my target such that it fills the vertical (narrower) side of the frame. I try, whenever possible, to always frame the same, regardless of focal length because the result is consistent sized targets that can be compared to other tests. If you frame differently each time it is hard to compare results.
There is no magic target; I got mine off the net years ago. I like vertical and horizontal lines because once you solve the main problems, what is left is mirror slap and that is always a vertical motion that affects horizontal lines the most. The vertical lines will be sharp, even if you have significant mirror slap vibrations, except at the ends of the vertical lines. I use the vertical lines as a sanity check. If they are blurred, I take a good look at my setup and usually question the validity of the results. I used to use 45 degree lines but found that it was more difficult to assess vibrations. There is some merit to a theory that for basic lens IQ testing, 45 degree lines are best, or at least should be included in any test. Here, though, we don't care about that. I learned this the hard way; a lot of my old tests were done on 45 degree lines. My own standard test chart has targets with horizontal and vertical lines, and other targets with 45 degree lines just for a one size fits all test chart.
I always do this indoors because I have been burned with unexpected puffs of wind. Testing wind is a good idea, of course, but for basic testing of different configurations the results are not consistent and reproducible, which I think is critical in any basic testing. I never publish wind tests because I think they would lead to confusion and misunderstanding without the benefit of being there to feel the wind and do the tests.
I originally used incandescent lighting, or some adjustable halogen lights. My latest idea is to tape the copy paper to an exterior window. I get a nice exposure. the important thing is to make sure the lighting won't change throughout a test sequence because then your exposure changes, which affects the results. I like my exposure ideally at the base ISO (for least noise), and usually about 2 stops down from wide open, but depending on the lens. Better lenses can be shot 1 stop down or maybe even wide open; you have to evaluate your lenses. Never stop down more than F/11 or so unless absolutely necessary; otherwise you can get problems with diffraction (on digital). How you light the target is not important as long as that light is consistent across all tests in a sequence (resulting in consistent exposures) and you arrive at the desired exposure. Sometimes it takes some creativity.
Always shoot fully manual - manual white balance and manual exposure. I usually shoot EV +1 to EV +1.5, such that my background is about luminosity 200-220, but never higher. Never allow the background to blow because if you do you will mask vibrations, which will either blow out too or be so high on the luminosity scale that they don;t stand out from the white background. Some people think my charts are a bit dark but I find it easier to analyze. Just MHO in that regard, as in everything else in this post.
If you are comfortable shooting raw, you should do so. You may want to generate lossless images in cases where the results are confusing. Sometimes I'll blow an image up to 500-800% and count pixels of blur to try to get an objective comparison. Even high quality JPGs will confuse the count. The point isn't to measurebate, but the problem with charts is that they do not have the fine detail you see in real world images. Little differences in a chart image equate to more significant differences in "pop" and impact in real world images. Plus, if you don't know what you are looking for in the chart, it is easy to miss significant differences that become very obvious when you enlarge them and count pixels. Then you know what to look for.
My bodies have mirror up, so I shoot a reference image with a 10s mirror up delay. I then shoot an exposure delay image (with whatever built in exposure delay provided- usually about 0.4s and then an image without exposure delay. I use a remote and I either use a wireless remote or take great care not to shake the cable.
If your mirror up image is bad, then you screwed up or your support is really, really bad . The image without any exposure delay is the acid test- the image you are interested in. Interestingly, in many configurations it is the WORST way you can shoot. I can out-shoot my remote without delay by using Long Lens Technique, and sometimes with two hands on the camera (which is said to be a very bad technique), but I am using sturdy tripods and long lenses that are conducive to LLT.
It is important to review your exposure delay images to see if that mode is effective. Mirror up is great but sometimes I have time for a half second delay but not the real deal. If you don't have mirror up, you should shoot reference images in brighter light, same framing, and same ISO and aperture. Shoot at least 1/2FL, and preferably at least 1/500s or so just to overkill it.
For focus, I autofocus with visual confirmation. The thing about a test chart is that it makes focus somewhat binary- it is either dead on or not. Shooting a real world "test target" usually results in confusion over DOF. I test my body's focus as soon as I get a new one, just like I test my support. I'm confident in my own bodies AF accuracy, but that is an important preliminary step, or if you do have a problem it will be obvious in the testing because nothing will be crisp. This is why mirror up is so valuable because it eliminates the variable you are testing- the support.
I think that covers everything; I have a number of tests I have done, with notes, here. This is just my way of doing things; I've been doing this for over 3 years. I don't consider myself a test geek either, but I put a lot of $$$ into my support and before I go out in the field I want to know exactly how it will react under any real world shooting situation. Unfortunately that takes a lot of testing. Otherwise it's all guesswork and learning by reviewing blurry real world images.
I'll be happy to email you or anyone else a copy of the chart I use. Just EM or PM me with your EM address. It's a 3.7MB TIF file.
#3. "RE: Tripod Test Regime" | In response to Reply # 2Thu 03-Apr-08 10:57 PM
Well Neil, you've done it again. Thanks so much!! I think I may have a question or two, but I'll have to reread, and actually give it a shot (tomorrow when my pod arrives). This, like many of your other posts, really belongs in an indexed FAQ here. Thanks again - you saved me from having to shop for a new lab coat and a pair of monoclonal cyborgs.
#4. "RE: Tripod Test Regime" | In response to Reply # 3Smiert Spionam Registered since 26th Nov 2007Fri 04-Apr-08 12:09 AM
(Where D is the displacement field, E is the electric field, B is the magnetic-flux density, H is the magnetic-field strength, p is the free charge density and j is the free current density.)
Really, not that tough.
Alternatively, you could follow Neil's recommendations.
#7. "RE: Tripod Test Regime" | In response to Reply # 4nrothschild Registered since 25th Jul 2004Fri 04-Apr-08 03:24 PM
I think Smiert was just kidding because I am pretty sure that magnetic flux-density does not have any practical effect on stick-slip or tripod vibrations, but actually impacts white balance
Here is a very useful formula:
Pixel scale (arc seconds per pixel) = 206.265 * (sensor pixel size in microns)/Focal length
I believe the D200 (10mpx) pixel size = 6 microns
Pixel scale (10mpx DX) = 1237.59/focal length
That means that a 1237mm focal length lens would have a scale of 1 pixel = 1 arc second. A 200mm lens would have a scale of 6.19 arc seconds per pixel.
A circle is commonly divided into 360 degrees. Most of us understand the basic FOV charts describing the FOV at various focal lengths. The lens forum has a FAQ link with such a chart, expressed in degrees. One degree is divided into 60 minutes of arc and each degree is divided into 60 seconds (just like clock time). The result is that 1 arc second is 1/3600 of a degree or 1/1,296,000 of a 360 degree circle. It is arguably a very small number
"Seeing" is a term astronomers use to describe the blurriness of the atmosphere. If you point a lens up at a 45 - 90 degree angle, the limiting resolution is usually in the range of 1-5 arc seconds, depending on how you define things and the nature of the atmosphere, but usually on the high side. True 1 arc second seeing is very rare except in certain advantageous locations, such as the top of Mauna Kea, Hawaii at about 10,000 feet. I have never seen estimates of "seeing" across the ground, which would be typically applicable to us here, but I am sure from experience it is typically far less than 5 arc-seconds, except maybe at dawn.
This is why you can truly "never get long enough" with your wildlife lens. Even if you make the lens longer and heavier and way more expensive, you will quickly run into external limitations of seeing. I have never personally shot a wildlife keeper image with my 500/4P + TC-301, working at 1000mm, at distances beyond 50 feet or so. If you don't understand all these mechanics, you would be tempted to blame your optics or blame your support, but personally I believe that we are typically limited to somewhere under 1000mm when shooting long distances (beyond 20-30 feet or so). If you are thinking about throwing huge money into a 600/4 and Series 5 Gitzo, etc., etc., this is an important consideration, and why I rarely use my 500P + TC-301 for terrestrial work. It does do a nice job at my bird feeder, though, at about 30 feet.
Using my lunar imaging as an example, because the moon for all practical purposes has the same apparent size at any point in time from any place on earth. The apparent diameter of the moon varies during the month because it's orbit is an eccentric ellipse, not a circle, but software is readily available that will give you that number for any location and point in time to a very high degree of precision. For discussion purposes it averages about 31 minutes of arc (1860 arc seconds) but varies by about 15% over time (from about 29' 4" to 33' 58" over the next month or so, for example).
Using that formula I can figure out if I can cram a full moon into my D200 at 1442mm, which is one of my own options. I can also examine my images and compute the actual scale of lunar features. Lunar photographers are very interested in the smallest lunar feature resolved in an image. Lunar features average 1.8km/arc second. I have resolved lunar features down to 10km at 1442mm, but that may be more the exception than the rule, even in a given image. In the image I am thinking of, I was fairly consistent at about the 17km level. Not too bad for a full lunar image single frame shot (no stacking). Theoretically the sensor is capable of resolving sub-arc second detail but in practice it is limited by seeing, my support and optics of course. I have the means to go to 7500mm and beyond, which is this Summer's project. It is extreme stuff and takes support issues to some extreme levels. Although I haven't used it for this purpose yet, my G1410 seems to be about as sturdy as the small astronomical pier that use, and has some clearance advantages over the pier. I have to do some very precise polar alignments to fully eliminate the motion of the moon and in that configuration I have to use the M20, which takes this whole issue of stick-slip to some interesting levels because the head is angled about 38 degrees from the horizon, or 52 degrees from vertical. When I have the time and energy I try to get things aligned to within 5 minutes of arc or so, centering Polaris (the North star) in the center of the field (visually) at magnifications of about 200x using an illuminated cross hair reticle.
This leads to another interesting issue, which is "shutter slap". We all talk about mirror slap but never shutter slap (sometimes referred to as Shutter Shudder. Say that 10 times fast )
My longest lenses are compact astronomical folded mirror designs that were first produced in 1954 and are still produced today with virtually no changes, making probably the longest running production of a single optical-mechanical design. The manufacturer will restore and retrofit 1954 scopes to make them current; you have to be expert in those matters to distinguish them from a 2008 model. That is truly a "lifetime investment". Anyway, the maker has some very detailed recommendations for photographic work at focal lengths of 1400mm and beyond. They claim that even with mirror lock up, images will be blurred due to Shutter Shudder and they highly recommend using the "Hat Trick", where a piece of cardboard with about a 1" slit is placed in front of the lens, the mirror is raised, the shutter is fired and then you wait some interval for the shutter vibrations to die out before sliding the cardboard across the lens aperture. In principle, with a narrow slit, you can do exposures as short as about 1/125", but it is easier done for 1/4" and longer exposures. That is not a problem because the optics are inherently F/20 (actually T/20) or slower in terms of exposure speed and might go as low as F/60 or even slower if you have the skill and daring to attempt it.
I tried that (hat trick) recently, to test that theory, and got worse results than using standard mirror lockup. I suspect that air movement from the large piece of cardboard moving rapidly in front of the lens may have done more harm than good. That's a work in progress. I am curious, though, if anyone here has done any work at extreme focal lengths and explored this issue of Shutter Shudder. My thinking is that the scope's photographic manual was written up to 50 years ago, somewhere in the F or maybe F2 era. Shutters have come a long way since then and it may no longer be an issue. Or, maybe it is...
#9. "RE: Tripod Test Regime" | In response to Reply # 7Smiert Spionam Registered since 26th Nov 2007Fri 04-Apr-08 05:13 PM | edited Fri 04-Apr-08 05:14 PM by Smiert Spionam
"I think Smiert was just kidding because I am pretty sure that magnetic flux-density does not have any practical effect on stick-slip or tripod vibrations, but actually impacts white balance"
Dang, busted again.
Interesting discussion, though! Neil, speaking of seeing: I spent a few days in Marfa and Fort Davis out in west Texas last winter -- the night skies were absolutely unbelievable. Got a nice tour at the McDonald Observatory, too.
Tom, you'll do ok if you follow Neil's tips -- the key thing, I think, is to not get too overly impressed by your tripod's ability to capture a stable image at 1 sec. It's the tricky 1/10 range that will burn you first (with problems creeping up into shorter exposures with longer lenses).
#6. "RE: Tripod Test Regime" | In response to Reply # 0