how do i get a clear picture of osprey flying

[photographerinoregon]

i cannont for the life of me get a nice clear pictures of a osprey flying. i have a canon 50d along with a canon 100-400 f/4.5-5.6 lens for some reason i cant get the eyes to come out more clear and the rest of the bird. attached is a photo non edited to see what you people think i can do to make a nice clear picture. hope someone can help. and if i shoot on av or tv mode the photos come out way to dark also can someone help me with that too i am using spot metering on all this by the way and shutter at about 1/1000 sec and apeture at 5.6 and iso at 800. please help


kevin:confused::confused:

[sparker1]

Get closer.

[Scottes]

Birds in flight, cloudy days, and the 100-400mm make for a bad mix. Try again on a sunny day, and stop down a little (f/6.3 or better, f/7.1).

I have a few thousand shots on ospreys, many with that lens but on a 10D or 20D. Last weekend I shot on a bright sunny day, 50D + 400mm L prime. 300+ shots, 3 keepers, many more in focus, but many, many shots that did not satisfy my criteria for sharpness.

Again, try on a sunny day - preferably in the early or late light so the sun lights up the underside of the bird.

[tdodd]

I also shoot with a 50D and 100-400 and, to be honest, I have a devil of a job getting great shots in anything but the best of light, especially if the bird is small in the frame. My cull rate is massive and I really should learn to not even bother when conditions are against me, but still I keep trying. Here are some things to consider....

- Compared to the 400/5.6 prime, the AF on the 100-400 is not that quick and the lens is (with some possible exceptions) not as sharp (or so the word on the 'net tells me). Viewed at normal reproduction sizes there should be no problem - the lens is just fine. Heavy cropping will reveal any flaws in AF accuracy and lens sharpness.

- Viewed at 100%, noise is clearly visible from my 50D at 400 ISO. Any NR to smooth the sky will also smooth the detail on the bird, unless you can apply selective NR using masks etc.. At 800 ISO, with a heavy crop, I think your chances for great IQ are low.

- The 1/focal length guideline for shutter speed when shooting hand held with 35mm cameras needs to be adjusted to 1/640 for a 400mm lens on a crop body. That assumes you are at least averagely good at holding the camera steady and are able to relax and settle into a stable position. I'm not sure that tracking a bird high above gives you the maximum stability. Shutter speed should be increased to compensate. At 1/1000 you might be OK, but, the rule was intended for reproductions at "normal" levels of magnification, such as a 10x8 print. When you view at 100% you are generating a virtual image of somewhere around 40"-50" on the long side, depending on your monitor. That's a magnification 4X or 5X greater than the guideline allows for. If you want the image to be sharp at the pixel level you'll probably need a shutter speed of around 1/4000. This is not only to avoid camera shake but also subject motion. If your subject move across the frame by just one pixel during capture then you will not have much success with heavy cropping.

To put this another way, if you resize a full frame from a 50D to 800x533 pixels then you have a pixel magnification factor of about 1/6 (800/4752 = 5.94). Thus, even if you had shake or blur that smeared across, say, three pixels at capture, the blur would (probably) not show up at all when resized to 800x533. At 50% viewing it may still not look too bad, but at 100% viewing it would be mush. To add insult to injury, because the noise is within the pixels themselves you end up with the noise looking sharp while the actual subject is blurred.

- You're shooting the shadow side of the bird against a backlight. Your exposure for the sky looks pretty good - very close to the right hand edge of the histogram but not blown. Unfortunately, your bird is underexposed. Also, the light that is on the bird is very flat and producing minimal contrast. Contrast contributes greatly towards the perception of sharpness. I've never tried the technique myself but you should probably try fill flash for shots like this, perhaps with a Better Beamer as well.

Take a look at the Arthur Morris video here - http://www.youtube.com/watch?v=l3xJ-wUXti8. It's about the 40D but look at the techniques - big primes on massive tripods with gimbal heads, or a hand held 70-200, but note the use of flash and the Better Beamer to throw light onto the subject. Also note how the subject fills the frame. At the end of the day it's all about the light. If you don't have the light you don't have the picture. Also, if you don't have the light then (heavy) cropping is off the agenda.

In good light, here is one of my more successful efforts at long range with the bird small in the frame. I know it's not an osprey but I don't think that matters especially. The significant difference is that I have the light in my favour and I can use more satisfactory combinations of ISO, aperture and shutter speed. I'm afraid I find it hard to achieve even this level of IQ consistently with my setup, when the bird is small in the frame. This was shot at 400mm, 1/2000, f/7.1, 400 ISO.

Full frame, no edits:
http://lh4.ggpht.com/_4_R8TkwT74w/SjDALZkyVNI/AAAAAAABGOo/6I-5MOr6Ej8/s800/20090607_131356_7261_LR.jpg

50% crop, no edits :
http://lh3.ggpht.com/_4_R8TkwT74w/SjDAN1AfAyI/AAAAAAABGOs/jTqrrGCSJ1Q/s800/20090607_131356_7261_LR-2.jpg

100% crop, edited for sharpness and NR in Lightroom :
http://lh3.ggpht.com/_4_R8TkwT74w/SjDAPnb0o_I/AAAAAAABGOw/9_xdYmVcS94/s800/20090607_131356_7261_LR-3.jpg

Even at 400 ISO noise is killing IQ in the crops. The captures may look a fraction underexposed but they are correct for the incident light (sunny 16) and necessary to maintain highlight detail in swans, which I was shooting at the same time with the same exposure. Example - this is a 50% crop, with the lens at 235mm and the same exposure as above :

http://lh4.ggpht.com/_4_R8TkwT74w/SiwSrH_PSVI/AAAAAAABGFc/PuRXT-9q-Ug/s800/20090607_130948_7231_LR.jpg

Lastly, if you want an example of a bird shot against the sky, here is one from the same day, again at the same exposure settings (1/2000, f/7.1, 400 ISO), and back at 400mm again....

Full image, no edits :
http://lh6.ggpht.com/_4_R8TkwT74w/SjDKsF-lVdI/AAAAAAABGPM/WKIcxn1GuKM/s800/20090607_131701_7290_LR.jpg

100% crop, some sharpening but no NR :
http://lh5.ggpht.com/_4_R8TkwT74w/SjDKtsIeNqI/AAAAAAABGPQ/NuJn9x41ab4/s800/20090607_131701_7290_LR-2.jpg

[hollis_f]

- Viewed at 100%, noise is clearly visible from my 50D at 400 ISO.

So don't look at your images at 100%. Unless you plan on printing them 6 foot wide!

[tdodd]

So don't look at your images at 100%. Unless you plan on printing them 6 foot wide!
Nobody is suggesting printing at 6', but in my opinion the size of the osprey in that first image is too small to make a "picture" for printing and could do with cropping to 1/4 of the frame to achieve a reasonable composition. At a stroke the available pixels from the file would have been cut to just 2376x1584. For printing at 300 PPI that would get you a print of approx 7x5 or 8x6. That's not much to hang on the wall. So let's say we really want a 12x8 print of this osprey and can live with printing at 200 PPI.

My monitor has a pixel density of 133 PPI. Thus to view the image cropped as proposed, at print size, I want the image to be around 12x133 pixels on the long dimension = 1596 pixels. Well, with only 2376 pixels to play with after cropping I would need to view at 1596/2376*100% = 67% to generate a physical 12" image on my screen. It seems like viewing at 50% or even 100% is not so preposterous after all.

The fact is, if the subject is small in the frame and you need to crop significantly to create the composition you desire, you do need to be concerned about the quality of individual pixels. So, what's the answer? Well, either you keep the ISO low and the exposure to the right so that you have good quality pixels that you can crop, or you make the subject sufficiently large in the frame that you can avoid cropping. I find that high ISOs and heavy cropping with a subject that contains fine detail (bird feathers) is a recipe for disappointment. So, fill the frame, get good light, or go shoot something other than BIF.

p.s. I don't print my stuff but I do display it on my 1920x1080p 40" HDTV. Thus, if I don't end up with at least 3840x2160 pixels after cropping then I am looking at viewing my images at >50% magnification as a matter of routine. With a 1D3 or 40D that leaves little room for cropping before I am exceeding 50% magnification. To labour the point to death, if that had been my osprey shot, but taken on a 40D or 1D3, then I would have cropped it to 1920x1080 for display on my TV. That is viewing pixels at 1:1. The individual pixels would have needed to be sharp, and largely noise free, in order for the image to be worth displaying.

[tdodd]

Let's put it another way. Here's a shot taken this afternoon in conditions of bright sunshine. One might imagine that this is similar to a shot of the underside of a BIF. EXIF is 390mm, 1/1250, f/5.6, 400 ISO. It was shot raw and pumped out of Lightroom with no edits....

Full frame :
http://lh5.ggpht.com/_4_R8TkwT74w/SjEyvLAbAZI/AAAAAAABGPY/bv5ZK8l8Quc/s800/20090611_172258_7786_LR.jpg

100% crop :
http://lh3.ggpht.com/_4_R8TkwT74w/SjEy_U_k_1I/AAAAAAABGPg/yttflgB9WiY/s800/20090611_172258_7786_LR.jpg

The full frame image looks great - smooth sky, plenty of clarity, reasonably sharp. But, in the crop, you can see the noise is significant, both in the sky and underneath the chopper. Now, with this example, NR can be used to tidy the noise up with very little detriment to the important details of the image.

100% crop with Neat Image NR :

http://lh6.ggpht.com/_4_R8TkwT74w/SjE2tbxfl2I/AAAAAAABGQE/pzskA1grPbw/s800/20090611_172258_7786_LR_filtered.jpg

But what if this was a bird, with fine detail in the plumage? The simple truth is that noise would have already destroyed that fine detail. The noise is simply too significant with respect to fine details within the scene. You can smooth the noise out but wave goodbye to the detail at the same time. It does not matter whether you view at 100%, 50%, 33%, 25% etc.. The fine detail will have been consumed. I'm sure 100 ISO would improve the results but, truthfully, for BIF, the answer is to fill the frame or at least light your subject well and keep your ISO low.

[mikeivan]

Kevin,
Lots of good suggestions herein. In my experience, the 50D will allow a pretty drastic crop and still sharpen up OK.
What I would do:
1) Shoot Av, stop down to 7.1, evaluative meter, use EC=+2/3 (I have limited success with spot metering, YMMV)
2) ISO 400 if you can, 800 if you must, 1/320 s or faster (IS is your friend)
3) If you can get closer, do whatever it takes to do so
4) Check that lens for front/back focus and correct with MA, if required

[Scottes]

Kevin,
Lots of good suggestions herein. In my experience, the 50D will allow a pretty drastic crop and still sharpen up OK.

Yep.

~25% Crop - this is 3.9 megapixels, cropped from the 50D's 15.1 megapixels
Canon 50D + 400mm L prime, f/6.3, a/1600, ISO 500, center-weighted, EC +2/3
Neat Image cleaned the noise up perfectly, and a little creative sharpening finished it off.

http://www.itsanadventure.com/postimages/Osprey_101625.jpg

[mikeivan]

Yep.

~25% Crop - this is 3.9 megapixels, cropped from the 50D's 15.1 megapixels


Scottes:
Very nice. I get confused with crop terminology. Why isn't your image a 75% crop? By that I mean, you discarded (or cropped) 75% of the pixels. So is it what is left or what has gone away or is there a standard way to think about it? I can see that 50% crop works either way and 100% crop is pixel for pixel but in between, I am not clear. Thanks for any help, not trying to hijack a good thread.

[tdodd]

Nice shot, Scottes. Maybe it's me, maybe it's my kit, but digging through my archives I have not found anything from my 50D and 100-400 (160 images in my library out of probably a couple of thousand shots with this combo) to match your image for sharpness. I know you used the prime here but even so, your example leaves me frustrated with my own results. I keep trying to figure out where the problem lies but so far I haven't pinpointed an obvious weak point. I mean, even static subjects don't come up as sharp as your osprey, even at similar crop/resizes. I guess I'll keep on looking.

[Scottes]

Mikeivan - There's no standard about "25% crop" which is why I spelled it out! :D

Tim - The light makes one hell of a difference when using the 100-400. The 400mm prime is a little bit sharper - in the field. Lab tests say the prime is definitely sharper, but when used under their optimal conditions in the field then I think they're close.


The image below is just the first sharp osprey I could find. To give you an idea, I skipped over 17 not-so-sharp images before finding this.

Canon 10D, 100-400mm, 1/500 at f/7.1, ISO 200. That's actually a hell of a lot less light than my shot above - about 3 stops less light. This is 6 megapixels cropped to 0.75 megapixels, 926 x 823. This is a 100% crop, no resizing done. I did post-process it in Lightroom for exposure, then in Photoshop for noise removal, creative sharpening, and contrast.

http://www.itsanadventure.com/postimages/Osprey_4850.jpg

No, I'm not bragging on the following: This was the 4850th photo I ever took, which means that I probably have about 3000 frames of birds, all with the 100-400mm. This is the 18th osprey I ever shot (did I mention the first 17 weren't too sharp?) I learned that lens up one side and down the other (have you noticed my sig?) and then went to a turkey shoot. I shot 700 pics of ospreys that day.

I will tell you that this can be done. I was a rank amateur when I took this shot. (Still an amateur, but now with 30,000 frames of birds.) This means that you can do it. I'm no superstar, that's for sure.


Now here's all my secrets for taking good bird shots with the 100-400:

1) Go where the birds are. If pretty ones aren't around, shoot sea gulls! Practice, practice, practice!

2) Go when the light is. Avoid low light with the 100-400 for birds in flight.

3) Shoot where the background isn't. Don't shoot against a busy background - the 100-400 doesn't like that (unless you can hold the center focus point on it's eyeball while it's flying around). If that focus point falls off the eye it will hunt for something else to focus on, and it's slow to AF, so it's frustrating as hell. Clear skies rule.

4) Practice, practice, practice!

5) Learn the bird. Watch what it does - a lot - and learn to anticipate when it's going to do something.


Another shot from that day, 160 shutter-clicks later. This is 1.5 megapixels, so about 25% of the frame. This is one of my very favorite osprey shots, even at 1.5 mp.

http://www.itsanadventure.com/postimages/CRW_4994_800.jpg

[tdodd]

Thanks for sharing, Scottes. I do agree, light rules! Another thing about busy backgrounds, apart from AF confusion, is that I do find the bokeh from the 100-400 to be absolutely horrible. It's very fuzzy and "tizzy", unlike the creamy smooth bokeh from my 70-200/2.8IS.

I was actually trying out different techniques today, with my 50D and 100-400, and even tried One Shot AF and all points active for birds that were flying across my path. I managed to get my best swallow shot to date by using that technique, but still not good enough to be a keeper. What surprised me was that even with a bird placed over the centre point, although somewhat small in the frame, the camera would often lock on to clouds in the background with outer focus points. The points on the left and right bottom diagonals seemed especially eager to lock onto the background. I'm not even talking about the edges between white clouds and blue sky here - just the transition from greyish clouds to whiteish clouds. I also know that, while a taped 1.4X and 100-400 will not AF with the centre point, the outer diagonals work quite well. This leads me to think that the 50D centre point is actually quite poor when it comes to using the 100-400 (or any f/5.6 lens?), and really not up to the job of AI Servo tracking at all. In fact, my sharpest images shot with the 50D and 100-400 were shot with either One Shot or Live AF. AI Servo results seem almost universally poor with my kit. It's not even the challenge of the moving subject. That helicopter shot I posted earlier should simply not have been that soft. It's a mess. Do you have any thoughts on that?

[Scottes]

Looking at your heron shot, you weren't on the focus point. Well, the very tip of the beak is on it - maybe that's not enough?

With the helicopter, the center AF is on the left wheel - a great contrast spot, so that should be dead-on. 1/2000 should eliminate any motion blur even if that chopper were flying at 100mph. This is less than an inch of movement while the shutter was open. But this would be 10 feet of movement during the 50D's 0.063 second shutter lag. Where you shooting AI Servo?

The bird below the swan is also dead on for the AF point. But I don't think that shot is so bad - it's hard to tell since there's nothing that *should* be ultra-sharp in that shot. it's just a lot of feather edges.


When I first got my 20D I felt that it's AF was off. But it was perfect on a tripod with mirror lockup and a cable release. So I I researched a weekend of shooting by checking every photo with DPP, which will show the AF point used. I checked the settings - ISO, aperture and shutter speed. From those I figured out how much light there was. I also realized that IS was off the entire weekend. Ouch.

The entire exercise told me that every shot that *should* have been in focus *was* in focus. Every shot that wasn't in focus had a reason.

The post is an interesting read, even 4.5 years later.
http://photography-on-the.net/forum/showthread.php?t=54646


I do agree with you about the 100-400's bokeh. Again, avoid busy backgrounds. If nothing to blur then you won't ever notice it. :D

Seriously, this is another point of using the right lens for the job at hand. Granted it's easy for me to say that since I have the 100-400 and 400 prime, so I have a choice.

BTW, here's a review I did comparing the 100-400 to the prime:
http://photography-on-the.net/forum/showthread.php?t=43436
It has some interesting insight into both lenses, and may help you understand the 100-400 better.


It might help to test the lens. A bright day and a stop sign will work. 1/2000, f/8 and shoot. Try a tripod and mirror lockup and a cable release (or self-timer) to remove your human flaws from the situation.

If you're crazy - like me - check a bunch of shots in DPP and check that the AF point is where it should be.

If necessary, set up to change the microadjustment for the lens. A very nice feature that will "fix" the focusing of the lens.

[tdodd]

Thanks again, Scottes. I have read the thread you linked to and I understand fully what you are saying. Yes, I do get good/great results on a tripod, with One Shot focus and Live (contrast detect) focus, but things seem to go downhill whenever I use AI Servo. The camera just does not seem to lock on with confidence and decisiveness. This is in contrast to my 1D3, which delivers far better results, but just not many pixels on the subject. Seemingly my findings are at odds with those of some well respected birders who seem to rave about the 50D's AF performance. I've tried IS on and off for BIF but it doesn't seem to change my results.

As for focus point aiming accuracy, we read everywhere, and can see within Canon's white papers, and also from field results, that the actual AF sensors are larger than the focus point indicators would suggest - about three times wider and taller. I know my camera can/will grab focus on something not covered by the focus point indicator at all, but simply close to it. Thus I may be just a little cavalier with my focus point aiming, but still on target enough for the camera appear, through the viewfinder, to secure focus.

Regarding the chopper shot, I doubt the thing was moving at anything like 100mph - probably more like 50mph. Since it was passing overhead and not coming directly towards me the actual closing speed on my position would have been significantly less - maybe 10mph or so. I was in AI Servo mode, with Mode 2 IS engaged, and I was tracking/panning the helicopter (or trying to) for an extended shutter burst, until I filled the buffer. There are sharper shots in the sequence, starting from the fifth frame (that's when I began firing the long burst rather than individual shots), but the one I posted is representative and the results do go downhill again later in the sequence. With the broad coverage of the AF sensor point I can see no reason for there to be a focus error at all. So if we assume the focus is perfect (is that a fair assumption?) then the problem is all down to movement - somewhere - but in a panned sequence, at 1/1250, with Mode 2 IS enabled, I had hoped for better results.

I have posted an album of the chopper sequence, as 100% crops with the focus point indicated. I wonder if you could take a look and offer your opinion as to whether my aiming accuracy, poor panning or perhaps poor shutter button technique is the primary cause of the low IQ of these shots....

http://picasaweb.google.co.uk/EezyTiger/Chopper#slideshow/5346354241741911362

The thing is, is this a large subject moving in a predictable path and yet I still can't get the results I would hope for. The image is too soft and so marred by noise, that if this was a BIF the results would be unuseable. Either the bird would have to remain so small in the composition as to be of dubious merit, or simply look like a mess of blurring and smearing, with no feather detail of any kind.

I really hope it is all my fault, because then I have something to work on, but if it is the kit then I may as well give up shooting with this combination, for BIF at least, unless I am far, far closer to my subject.

Anyway, I'm off to recheck my static results and AF calibration - again. I look forward to any insights you might offer.

Cheers,
Tim.

[Scottes]

I have no more insights on the chopper shots. I would have to think that they should have been sharper. The burst shots are better, but bursts definitely help getting at least 1 keeper from the series. (Though not so much with birds in flight, since they often move too fast.)

However, I have to admit that I have been in the same boat - shots that should be sharp simply aren't. But I've gotten this with 3 cameras (10D, 20D, and 50D) and 3 birding lenses (100-400, 400 prime, 500 f/4). Could it have been all 3 cameras? All 3 lenses? After all my experience?

There have been many times when I think my sharp shots are just luck. That's when I lust after a 1D body. And why I have never tried any of my friends' 1D bodies - I don't want to know how much better they are, because then I'll buy one, and that could end in divorce. :D


I have done a LOT of shooting with Cyberdynesystems, who always has a couple of 1D bodies and L lenses when we shoot. We've discussed this a lot - he listens to my whining fairly well. I am of the opinion that the 1D AF is far, far superior compared to any of the prosumer bodies I use. I now lust after one, and am convinced that I need one. Let's hope that my wife understands come September when the new 1D hopefully comes out.


One thing that might be of interest. I recently went through all my shots, rating them to finally build a Smugmug gallery of my best stuff. Out of 39,000 shots I've taken, I have a total of 37 shots that rate 5 stars in all categories (subject, composition, sharpness, etc.) That's 1 out of every thousand. And that ain't bad at all, really.

So for now, I am trying to note the good ones, not the ones that don't meet my expectations. If I go out shooting for a day and get 1 great shot, and/or 2 good ones, then it's a very successful day. As much as I'd like every shot to be sharp and properly exposed, well, that's just not realistic, I guess.

[BradM]

With years of experience with the 100-400mm the shots of the helicopter posted are just what I would expect as well as those of Scottes' Osprey images.

The primary difference between the two are the nearness of the subject but as important is the aperture. The 100-400mm just does not come into it's own until it is stopped down a bit, 5.6 just doesn't pull up detail, by f/7.1 the lens starts to shine and f/8 is usually the sweet spot on the number of them I have used.

And the noise comments about the 50D (and slightly applicable to other xxD bodies as well) is about right, I don't use my 50D for wildlife stuff because the 40D performs better at higher ISO's. However the use of noise reduction software doesn't mean that fine detail has to go away, it only means that either better technques or tools are needed to remove any noise than is currently being used. All of the following images have had at least one pass of Neat Image to remove chroma noise, most have had two passes, and fine detail is still there.

While anymore I am usually shooting the 1D3 w/ the 500mm, the 40 has the 100-400mm mounted for other circumstances that pop up. All of these were shot with the 100-400mm and various bodies from the 30D, 40D, maybe a 50D or 1Dmk2 and ISO's from 400 to 800 I believe.

http://i7.photobucket.com/albums/y259/Bradklr/hummerorb10.jpg

http://i7.photobucket.com/albums/y259/Bradklr/nightheron-1.jpg

http://i7.photobucket.com/albums/y259/Bradklr/swansfir.jpg

http://i7.photobucket.com/albums/y259/Bradklr/mallardhenflight.jpg

http://i7.photobucket.com/albums/y259/Bradklr/pelican10-1.jpg

http://i7.photobucket.com/albums/y259/Bradklr/redtailflt4.jpg

http://i7.photobucket.com/albums/y259/Bradklr/redtailfltsm.jpg

http://i7.photobucket.com/albums/y259/Bradklr/roughlegged12.jpg

But I only shoot wildlife in AI servo, 99% of the time on center point only and always at ISO's of 400 or higher. But the technique when using servo needs one to adapt to the systems strengths and avoid the weaknesses. One of the best ways to do this is get the focus off of the shutter button and use the * or the AF button.

You can acquire the subject early and track it in, as the servo system is predictive the more time of views of the subject you can give the algorithm to calculate the better I have found it works. If you lose focus then off and on the focus button again, sometimes it requires this to done multiple times but it does work for me.

To improve bif shots, one must first get closer, counting on big files to crop into is a losing game.

Next, exposure has to be right to reduce noise at any ISO setting that robs detail. You can shoot at high ISO's and still have little issue with noise if the image is exposed where it needs to be.

Learn the strengths of the gear in hand. If you want the fine detail you have to shoot where the lens captures that detail and adjust other parameters to be able to do so. And practice, practice and practice some more.

[Scottes]

...the lens starts to shine and f/8 is usually the sweet spot...

...shoot wildlife in AI servo, 99% of the time on center point only and always at ISO's of 400 or higher.

...get the focus off of the shutter button and use the * or the AF button.

You can acquire the subject early and track it in... If you lose focus then off and on the focus button again...

To improve bif shots, one must first get closer, counting on big files to crop into is a losing game.

Next, exposure has to be right to reduce noise at any ISO setting that robs detail.

Learn the strengths of the gear in hand.

And practice, practice and practice some more.
So many good points, Brad. I had to quote for emphasis, as I have found them to be true 100%.

[tdodd]

Thanks for your input once again, Scottes, and you Brad :)

I have, once again, double checked my AF microadjustment. As far as I can work it it is as good as it gets as it is. I have compared sharpness (focus) using Live AF, One Shot and AI Servo on the same fixed subject from a tripod, using tethered Live View. Live AF does just have the edge, and I cannot match it by twiddling with AF microadjustment. The results with AI Servo on the same fixed target are distinctly softer. Not horrendous, but not as good as One Shot.

Anyway, confident that my AF is set up as well as can be expected, I once again tried a few more distant targets. One example was a TV aerial at a distance of perhaps 150m. I have clear line of sight to it and no background distractions. Although the camera will lock focus without any problem the resulting images viewed at 100% are hopeless. It does not matter which focus mode I use. Everything is cack. A light plane flying slowly by, even further away, was absolutely diabolical.

Surely my gear can't be this bad, and surely when I apply all my concentration to steadying the camera I can't be this hopeless. I know the lens can deliver results at distance because my moon shots are not horrific, for a zoom. So what is going on? Well, the Live View image reveals something quite interesting about the more distant targets. Quite simply they are wobbling and morphing. It is a sunny, but not especially warm day (air temp = 20/68 in the shade), but I can only conclude that there is enough heat coming off the roof tops and ground to make the air shimmer and destroy the optical path to more distant targets.

I've uploaded a video of the phenomenon to YouTube. The video is only handheld but the shimmering/wobbling/morphing of the subject is quite evident. This is not something that is visible to the naked eye but the camera picks it up easily at 400mm and 100% Live View magnification. Just watch this - I try AFing a few times within the seqence but it doesn't do a bit of good....

http://www.youtube.com/watch?v=MsOSBy8P7eY

The same effect is not evident when pointing the camera at a target 5m away instead of 150m or 300m. I would be interested in your opinions. Is my analysis correct, do you think? It certainly helps explain my disappointing results, but it makes me wonder how anybody manages decent photos at a distance in warmer climates. I imagine a long lens would only magnify the problem in the same way that pixel peeping does.

p.s. While using tethered Live View at 100% I did also try stopping down and using Live DOF Preview, which maintains the exposure. I detected no perceptably change in IQ between f/5.6. f/6. 3, f/7.1 or f/8. Things did seem to sharpen up a bit at f/9, which is kind of weird when you understand that f/7.1 is supposedly the safe limit before diffraction starts softening things again. I also do use back button focus 99.9% of the time.

[Scottes]

Checking Brad's shots again...

Look at the light in those shots. Every shot has perfect lighting. Knowing Brad's work, this is not an accident. Get there by dawn.

Note the backgrounds in all shots. Only the swan shot gives a slight chance to foul the 100-400, and even that is distant enough. Again, not an accident.

All of those shots show planning and preparation. These conditions are not circumstances.

[Scottes]

The same effect is not evident when pointing the camera at a target 5m away instead of 150m or 300m.
First - and I mean this nicely - what the hell are you doing shooting at something 150m away? A condor would be 10 pixels wide at that distance! Why not test at the typical shooting distance? My osprey shots range in the 10-20m range. For small birds - finches and stuff - it's 3 or 4 meters.

I can only conclude that there is enough heat coming off the roof tops and ground to make the air shimmer and destroy the optical path to more distant targets.
Makes sense. Another reason to be out early before the sun starts to warm things too much.

I have compared sharpness (focus) using Live AF, One Shot and AI Servo on the same fixed subject from a tripod
So we have a statement here that says that everything works perfectly under perfect conditions.

Then make sure you are shooting in perfect conditions!

When I shoot birds, I am almost always out there by dawn. I usually quit by 10 AM, unless the subjects, lighting, and temperature (heat shimmering) warrant it. I get more keepers when the sun is below the bird than when it's above the bird.


Mind you, I still agree with you on many points about the "imperfections" in the camera, AF and lens. Work around them. Put yourself into the situation that provides the most perfect conditions possible.

[tdodd]

Looking at the contrast (or not) and the exposure settings for Brad's shots it looks to me like he is shooting in everything from bright sunshine to "brighter than bright sunshine!" to overcast to shade and all the shots look fab, so I don't think it is just the light. The bokeh also looks nothing like any 100-400 bokeh I've seen. Nice and smooth and creamy - not jagged gashes.

In terms of picking my times, access is a bit limited. My local RSPB reserve doesn't open its doors until 4.75 hours after sunrise and shuts them around 4.25 hours before sunset. I am also restricted in terms of destance and direction by fencing and multiple water channels with no crossing points. My local country parks don't offer that much of interest either and they don't open until 3.25 hours after sunrise. Basically I have to work with what I have, which may be why I'm largely on a hiding to nothing :(

On rare moments when I can get (much) closer I do not have a problem. This was shot with my 50D and 85mm/1.8 prime. I might have had my Kenko 1.4X fitted as well but as the lens does not report there is no info to confirm or deny that in the EXIF.

85mm (or 120mm?), 1/2500, f/4, 200 ISO :

http://lh3.ggpht.com/_4_R8TkwT74w/SdyLqpHkREI/AAAAAAABCSQ/UCQxVRqM-t0/s800/20090401_153747_5657_LR.jpg

100% crop :
http://lh5.ggpht.com/_4_R8TkwT74w/Sit0EV9YKAI/AAAAAAABGE8/hwaMrYqzM90/s800/20090401_153747_5657_LR.jpg

My expectation might be unrealistic, but that's what I want to see from my 100-400 L grade lens.

[tdodd]

First - and I mean this nicely - what the hell are you doing shooting at something 150m away? A condor would be 10 pixels wide at that distance! Why not test at the typical shooting distance? My osprey shots range in the 10-20m range. For small birds - finches and stuff - it's 3 or 4 meters..
I calibrated at approx 20m and checked again at 30m. The example at 150m was only to see whether calibration went way off at greater distances. e.g. the chopper was (I guess) around 200m above me and it was garbage. So I wanted to see whether the camera will focus accurately at those greater distances. If you don't mind me saying, there are other things in this world to shoot apart from birds at 20m or less - motorsport or airshows, for example. I would like my camera to be able to deal with such things too. I hope that is OK with you :)

I suppose the biggest lesson to learn from all this is to stop wasting my time shooting birds in my local environment, because the odds of anything remotely interesting (not ducks, swans, geese or gulls) coming within 20m are about nil. It doesn't stop me trying though.

[Scottes]

...so I don't think it is just the light.
I just can't agree with you here. On all but the first and last shots I can tell that the light was perfect, or close to it, and Brad made the best use of it.

My expectation might be unrealistic, but that's what I want to see from my 100-400 L grade lens.
The 85mm f/1.8 that close with plenty of light, compared to the 100-400 under lesser conditions? (From what I've seen so far.)

I think this just says that you need to make the conditions better for yourself.

[Scottes]

If you don't mind me saying, there are other things in this world to shoot apart from birds at 20m or less - motorsport or airshows, for example.
Planes and car details, compared to feather details... You'll be OK. I just hope that you don't expect to read the parking sticker on the windshield of a car doing 200 kph. :D

I suppose the biggest lesson to learn from all this is to stop wasting my time shooting birds in my local environment, because the odds of anything remotely interesting (not ducks, swans, geese or gulls) coming within 20m are about nil.
For the less-than-interesting birds, practice on them! But... LOL, been there, done that! I am bored to tears shooting the local fauna. I've shot them all so much that they are no longer interesting. I make the occasional weekend trip - always within driving distance - to get a bit more diversity.

You can get damn near anyplace in England or northern France with 6 hours of driving. Go find a place or event that will guarantee you several hundred shots in a day. You'll get a ton of practice, and a keeper or two.

[tdodd]

Thanks again, Scottes. I know you are right but it is not easy for me to arrange for those perfect conditions to all materialise at once. I have to make do with what circumstances offer me.

Hopefully Kevin has picked up some useful pointers from this thread. I know I have. Thanks again.

[Scottes]

I have to make do with what circumstances offer me.
No, you don't. Do you think I am THAT much better off than you, here in New England? My weather isn't THAT much better than yours. My closest birding spots (besides 1 mediocre place) are 1.5 hours away. The ospreys are a 3.5 hour drive, and they fly like this only 3 weeks out of the year. I get about 4 months of bird shooting weather each year. Crap, it mid-June and I've had 3 chances to go birding.

When I go birding, I check the weather report, I'm up at 3:30 AM, and I'm on location before the sun rises over the horizon. Sure, your local parks aren't open that early, but what do you have within driving distance? London to seashore isn't that far, and driving before dawn ensures a quicker trip.

If you want to bird, you'll bird.

[hollis_f]

I suppose the biggest lesson to learn from all this is to stop wasting my time shooting birds in my local environment, because the odds of anything remotely interesting (not ducks, swans, geese or gulls) coming within 20m are about nil. It doesn't stop me trying though.

Well, this...

http://www.frankhollis.com/temp/Godwit%20001.jpg

And this...

http://www.frankhollis.com/temp/SnowBunting.jpg
were taken within a two-hour drive of you.

While these two...

http://www.frankhollis.com/temp/Sanderling.jpg

http://www.frankhollis.com/temp/Kingfisher.jpg

Are less than an hour's drive.

[CyberDyneSystems]

Great thread, nothing to add that hasn't been said.. and said well.

[BradM]

I calibrated at approx 20m and checked again at 30m. The example at 150m was only to see whether calibration went way off at greater distances. e.g. the chopper was (I guess) around 200m above me and it was garbage. So I wanted to see whether the camera will focus accurately at those greater distances. If you don't mind me saying, there are other things in this world to shoot apart from birds at 20m or less - motorsport or airshows, for example. I would like my camera to be able to deal with such things too. I hope that is OK with you :)

I suppose the biggest lesson to learn from all this is to stop wasting my time shooting birds in my local environment, because the odds of anything remotely interesting (not ducks, swans, geese or gulls) coming within 20m are about nil. It doesn't stop me trying though.

Shooting something like a TV aerial at a 150m is just asking for issues, not only can focus be off but atmospheric issues wil arise, whether heat haze or cooler air, or pollen, or humidity/moisture in the air will all detract from image quality.

It certainly is no test if equipment is operating correctly, like Scottes' said you shoot or test at the distances you want to capture images. And if you are limited to 100m for subjects, it isn't the gear that has issues it is the distance.

Too many people believe a big lens means you can stretch your shooting range out, it doesn't work that way if you want detailed images you need to have has many pixels of the subject filling the frame to do so, and that means shooting close. If you were doing it with a 300mm, a 400 means you can get about 25% further away, but most would perfer to get closer I know I do.

The only shot I shared above that exceeded maybe 20m was the swans, everything was under that. And I don't care where I go in the US and Canada I have found that I find subjects under that distance.

It is never a drive out and shoot affair. Shooting wildlife takes more than equipment and camera skills; it takes time for planning where the subjects might be (what they eat, where they live, what their schedule is), planning on how to approach the subject and where the light is coming from, and then it takes time of just sitting and waiting.

The biggest issue for many is that they just don't spend the time to find the subjects and wait for them. All of the raptor shots above were all under 15 meters because I knew where they hunted regularly and I went and sat and waited on them to show up. Some Rough Legged hawk shots took a week of traveling a total of 3 hours each day and 200 miles round trip to get the shots I wanted. That is not an uncommon occurrence for those that take this seriously.

Wildlife photography is just not an instant gratification situation if you want close and detailed images. It is work and many find they have better or more important things they need to do with the time that is required to do this stuff.

Also on the heat issue, it doesn't have to be just on a sunny day or even a warm day. I often shoot from my car in the winter at subjects a near distances and just having the warm air spill out of the car and in front of the lens can affect the accuracy of the AF system. For the best detail it is important to aware of these types of issues that will contribute to less than stellar images and to take steps to mitigate the issues. Like in the above turn the heater off, or open the sunroof and let the heat rise directly and not out the window I am shooting from.

I just can't agree with you here. On all but the first and last shots I can tell that the light was perfect, or close to it, and Brad made the best use of it.


The 85mm f/1.8 that close with plenty of light, compared to the 100-400 under lesser conditions? (From what I've seen so far.)

I think this just says that you need to make the conditions better for yourself.

Thanks the light isn't often perfect here is the PNW, but you make do with what you have. That is why my ISO lives at 400 and above. But I do set myself up to take advantage of the direction of light and where I "expect" the subject to be or come from.

But sometimes you just get lucky, like the female mallard shot that was early in the morning and we jumped that duck off a creek. She decided to fly into the sun crossing my location. I happened to have a friends 100-400mm on my camera that he thought was a soft copy. Think this shot proved to him that the lens image quality wasn't an issue. For some reason I can't check the exif from opanda of my shots but I believe the ISO there was 800.

As almost an aside, the softer backgrounds than one might expect is two reasons, first trying to set myself up for the shot I want and also post production. Using noise reduction will soften a background and not strip detail, these shots are good examples of that, particularly the hummer or the swans.

[Nighthound]

A pleasure to read this thread. I can't add anything that hasn't already been stated so well. Great information for many skill levels.

Brad, you are an invaluable resource in this hobby and always an inspiration.