Shooting birds (in flight)

[DionM]

What type of shutter speed is "fast enough"?

I did some shots today and the ones at 1/640 or 1/800 (@ 300mm) seemed to be quite blurry. at 1/4000th they are nice and sharp ...

Eg this one @ 1/800th (with lots of USM, probably too much).
http://fotos-files-forums.net/fotos/Birds/IMG_0565_crop_frame.jpg

vs this one at 1/4000th
http://fotos-files-forums.net/fotos/Birds/IMG_0583_crop_frame.jpg

The first to me looks terrible (even ignoring USM artefacts) whilst I am pretty happy with the second.

So my question is was I just really shakey, are the birds really that fast, is my lens AF not very good (100-300 4.5 to 5.6 USM lens; not the L; but not the 75-300) or am I just imagining things?

Taken with a 20D in AI Servo mode.

I've never really done bird photos before, but its something I am exploring now that I have a digital and can shoot away and experiment.

[Scottes]

The first looks more like it's slightly out of focus, but that's hard to tell in such a small image. Check the original and zoom in to 100% and check the face and the wing-tips. Slow shutter will usually show either ghosting or motion trails. If you see either in the wing-tips but not the face then it's most likely OoF. If you see either in both spots then it's most likely slow shutter. But it still can be hard to tell.

But the "correct" shutter speed depends on the distance to the bird and it's size, the speed of the bird, and the bird's movement (flapping fast or slow). It's very hard to judge, really. So if you wish to freeze them then 1/1250 will usually work fine and 1/2000 almost all of the time (but not on hummingbirds in flight!) but it still depends on the factors mentioned above. For most bird shots if I see that I'm getting faster than 1/2000 then I'll stop down the aperture to get more depth of field.

Also, it depends on what you like. I like to freeze the bird body but leave the wing-tips blurred. 1/4000 will generally completely freeze the bird and I don't think this looks natural. I like the implied motion when the tips are still blurred by movement.

Lastly, birds in flight are one of the toughest subjects to shoot. You didn't pick an easy subject, so don't get discouraged easily.


Oh, for 300mm then 1/600 or faster should be more than enough to eliminate shake by either you or the camera. But that still depends on you - good camera-holding and shooting technique can help this a lot.

[DionM]

Scottes,

Thanks for that. I looked in at 100% and yes the whole bird is blurry. Maybe it is speed of the bird (they were playing in the wind). As the shots were done at 300mm I am looking for about 1/500 as a minimum for the basic "rule of thumb". So maybe 1/640 is a little too close; especially as I am moving, and there was a bit of wind etc? I've also heard that the rule of thumb is also a rule of thumb that only applies for non-cropped 6x4s ... for 100% crops etc off digital you will still see some blur ... any weight to that sort of comment? I'm new to the whole crop it and sharpen it area in digital photography :)

I think next time I will set shutter priority for around 1/2000 and see what happens. I don't mind totally frozen shots but a little motion on the wingtips wouldn't go astray, as you mention.

My camera holding may not be up to scratch, I am noticing the weight of the 20D vs my old EOS300 (Rebel2k) film camera. I should also start using a monopod too, maybe.

[Scottes]

Blurry or ghosted? Check the trailing ends of the wings and tail. If it's blurry then it's OoF - if you see ghosting it was bird speed.

100% crops will definitely show more blur but realize that you may just be looking at a lack of sharpening, especially if you shoot RAW which won't have sharpening applied.

Camera-holding and panning techniques are very important - they're worth practicing. A monopod can make a world of difference but restricts your movements so it can be a trade-off for birds in flight. I use one all the time, but rarely for birds in flight. But you can get one for $30 to see if it works for you. You certainly won't be wasting the $30 as you'll get plenty of use out of it even if you never use it for flight shots.

As a note I always shoot birds in Av - wide open almost always - and keep an eye on the shutter speed. If it gets too low then I jack the ISO up a notch. But with Av mode you'll always be getting the fastest shutter possible in the current lighting. Just check the shutter from time to time to make sure that the light isn't dropping off - the human eye won't notice a 1-stop drop in light but the camera will.

[DionM]

Yeah it was blur rather than ghosting.

And yes, I always shoot RAW.

I have a monopod - but only use it rarely. Must use it more often. My trouble is my height - at 6'8" none really suit me ...

Thanks for the tips - I shall practice more :)

[Scottes]

You might want to look into a harness - like those that parade flag-carriers wear. Slip the end of the monopod into the harness and you now don't have to worry so much about the length of the monopod. That would certainly help the monopod situation considering your height. It won't be ground-stabilized but it will certainly help anyway. Just an idea.

[maderito]

The first looks more like it's slightly out of focus, but that's hard to tell in such a small image. Check the original and zoom in to 100% and check the face and the wing-tips. Slow shutter will usually show either ghosting or motion trails. If you see either in the wing-tips but not the face then it's most likely OoF. If you see either in both spots then it's most likely slow shutter. But it still can be hard to tell.

Excellent thread!

Scottes - Can you define (or show examples) of the difference between "ghosting" and "motion trails" and their relationship to shutter speed and focus?

How does your strategy distinguish between blurriness related to speed, camera shake, and focus? Is it more generally applicable to shooting other moving objects (e.g. cars, humans, etc.)?

[DionM]

Here are the 100 percent crops with no sharpening or anything from the above photos.

[Scottes]

Scottes - Can you define (or show examples) of the difference between "ghosting" and "motion trails" and their relationship to shutter speed and focus?

Oh, crap. You actually expect me to prove that I know what I'm talking about? I fake most of this you know....
:)

Let me see what I can find here... I generally delete the stuff that's OoF so it might take a bit...


OK, here's one for motion blur.

Full image:
http://www.itsanadventure.com/postimages/Osprey_4994l.jpg (http://www.itsanadventure.com/postimages/Osprey_4994l.jpg)

Blown up wing tips. Resize 300% I think but maybe 400%.
http://www.itsanadventure.com/postimages/Wingtips_4994.jpg
Note, with the help of the arrows, that all blur is in one direction. (Well 2 directions if you count the down-stroke but you know what I mean.) Note that you see a distinct line next to some ghosting. Note that there is no ghosting at the very tip of the feather (black arrow).

This is motion blur.

Blow-up of the head 200%:
http://www.itsanadventure.com/postimages/Head_4994.jpg
Fairly sharp, no real blurriness except the wingtips again.

This is what I call perfect. The body is in focus and sharp (enough) but the tips of the wings show motion-blur, implying motion. 1/1000 at f/8 ISO 400, by the way.


Here's a 100% crop of one that's completely out of focus (well, it's not that bad, but bad enough). You can see that it's generally soft, and there's no real detail anywhere. (Remember that the one above was resized up.) Compare the overall softness to the obvious motion blur in the wingtips.
http://www.itsanadventure.com/postimages/OoF_4890.jpg

Note also that you see a white line trailing the right wing. This is more like chromatic aberration - well, the purple near the tip of the beak is more like chromatic aberration, really. But the white is due to an imperfect lens along a high-contrast line - it's not due to motion. The ghosting shows motion.

With motion blur you'll still see some distinct line along with ghosting. When totally OoF it's just soft and no lines of detail are evident. f/9 at 1/400 ISO 200.

[Scottes]

Now take a look at DionM's 100% crops. On the first pic, note the very tip of the wing. Note the lack of a distinct line or ghosting. It seems like the blur is in all directions. This is out of focus.

On the second image we can still some details. Note where the neck meets the breast - you can still see some detail lines in the white - an area that would quickly lose details if it were out of focus. Keeping the subtle details in the whites is a great indicator of focus. This was definitely in focus - maybe it isn't the sharpest lens but it's in focus.

Everything in the second pic is fairly sharp, though some finer details are lost in noise. (Was this at ISO 800? ) There's no ghosting or blurriness anywhere. Note in particular the sharp edges of the white spots near the tip of the wing. They're still sharp, and this part of the bird should have been moving at a high speed. It's been frozen in time by a very high shutter speed. (The off-color line on the upper part of the body is high contrast mixed with a little bit of CA.)

Also note that the wing is partially covering the face, indicating that it was very most likely in mid-flap - most birds will lift the wings when gliding. Compare this to the first pic - the wing position indicates gliding. Since the first gull doesn't show motion blur at the wingtips this strengthens the thoughts of gliding. Also, I know the shutter speed so I'm cheating a bit here. Wing position isn't fully indicative of gliding versus flapping but it's pretty accurate. Knowing the shutter speed and the bird habits helps.

Gliding birds can handle lower shutter speeds, especially when combined with decent panning technique.

[Scottes]

Scottes - Can you define (or show examples) of the difference between "ghosting" and "motion trails" and their relationship to shutter speed and focus?

I guess that I never really answered this question... I consider ghosting and motion trails to be basically the same thing, with a subtle difference. Maybe.

Motion trails are ghosting that occurs at the trailing edge of a moving object. They'll generally fade to nothingness. With the wingtip in the first Osprey shot there's still a bit of a distinction in the ghosting since the wing paused very briefly and then came back down through the same space - double-ghosting somewhat. This bird was basically hovering so it has very little side movement, just flap movement. The ghosting would be less distinct in a bird that was flying AND flapping.

Imagine taking some slow-shutter pics of a metronome. If you confined the pic to the bottom of the arc you'd see motion trails - a blur that fades to nothing. But if you took the pic at the end of the arc you'd get motion trails but then the pendulum tip would swing back through the motion trail.

So basically what I call ghosting and motion blur are really the same thing.


As to shutter speed... it all depends on the speed of the movement. A pic of a hummingbird - even at something like 1/2000 - will still show a LOT of blur. But a big old, slow-flapping heron will barely show wing blur at 1/200 (as long ast he panning was done well).


Out-of-focus shots will just look soft with no distinct lines. I think we're all used to seeing THAT. :-)

[Tom W]

Scott - thanks for the information. I've been working on honing my flight image technique, sometimes successfully and other times not. My own experience provides good pointers, but your experiences are providing even more.

[DionM]

Thanks for the info Scottes ...

Yes the lens isn't the sharpest - 100-300 USM which does get a little soft at 300mm end. Also shooting wide open. ISO was 400 on the second image ... I too am a little bit puzzled about the noise in the image ? Seems quite bad really ? Background was muddy sand/muddy water (mudflats).

20D, sRGB etc.

Here's the full EXIF:
Make - Canon
Model - Canon EOS 20D
Orientation - Top left
XResolution - 350
YResolution - 350
ResolutionUnit - Inch
ExifOffset - 200
ExposureTime - 1/4000 seconds
FNumber - 5.60
ExposureProgram - Aperture priority
ISOSpeedRatings - 400
ExifVersion - 0221
DateTimeOriginal - 2005:01:22 17:30:00
DateTimeDigitized - 2005:01:22 17:30:00
ShutterSpeedValue - 1/4000 seconds
ApertureValue - F 5.60
ExposureBiasValue - 0.00
Flash - Not fired, compulsory flash mode
FocalLength - 300 mm
UserComment -
FlashPixVersion - 0100
ColorSpace - Uncalibrated
ExifImageWidth - 652
ExifImageHeight - 351
FocalPlaneXResolution - 3959.32
FocalPlaneYResolution - 3959.32
FocalPlaneResolutionUnit - Inch
CustomRendered - Normal process
ExposureMode - Auto
WhiteBalance - Auto
SceneCaptureType - Standard

[Tom W]

DionM - how was your histogram? Was the image originally underexposed? That would contribute to noise.

[DionM]

DionM - how was your histogram? Was the image originally underexposed? That would contribute to noise.

No the histogram was quite good. Shot in RAW. That's why it is puzzling ...

[Scottes]

Underexposure will certainly contribute to the noise but the noise is in the background which is all lower midtones. Lower midtones show noise the worst. My osprey was also at ISO 400 but the sky show very very little noise since the blue sky is more like lower highlights or upper midtones.


My theories from my experience and just my opinion... nothing scientific....

Take the histogram and imagine it broken into 5 stops from left to right being dark to bright. Anything to the left is black - efectively no light - which doesn't show noise, as you can see in the wingtip. The left most fifth has only 128 levels of info - you're still not likely to see much noise here. The 2nd fifth is where the most noise will appear, and that's about where you background seems to be. There's enough levels to have distinguishable noise and there's enough light to "create" noise when the signal is amplified. The middle fifth (the grey on the wings) is a bit better but noise is still fairly visible. By the 4th fifth (the darker whites on the breast) there's enough light so noise isn't so apparent here, and by the 5th fifth (the bright whites on the bird) the image is usually properly exposed (or exposed enough) so that there's so very little noise that it's not noticable.

Details also help to hide noise, and your background has none so the noise is apparent. The gray feathers have some detail so the noise isn't glaring but it's there when you look for it. The whites (dark and right) don't have much details but by this time there's not much noise to begin with.


A touch of noise reduction would help this image tremendously. Check out Noiseware for a freebie, but I think you can get a free version of Neat Image,too. Since there's so much noise in the background any noise reduction will wipe that out easily and will blur/muddle into a nice background. You won't even have to set it too strong so there's very little fear of losing details.

[maderito]

Scottes -

Excellent presentation -- and thanks for answering my questions. :):)

I'm going to look through some of my pics of humanoids in motion to see if I can apply these principals. I may be getting back to you.

The real challenge for me is distinguishing camera shake from OOF. Maybe we'll talk about that next time.

[Scottes]

Camera shake will usually be in a distinct direction or two or three. Look at the edges of the ghosting. Especially at the points of extreme motion, like fingertips in a person walking (if they swing their arms a bit) or the heel of the foot that's moving.

Of course, 400m at 1/8 second blows this theory out of the water. But a little bit of camera shake should show small ghost lines, not overall softness.