Poll: in focus or out of focus

[arash_hazeghi]

I am curious to know what you think about this photo as presented



There are three choices, pick one :)

[DanielTMiller]

Looks in focus to me, nice shot.

[Axel Hildebrandt]

I think it is in focus but the shutter speed is a bit low.

[Jan Wegener]

the web image looks quite ok, but you can almost make every image look good on web, even if it's way OOF....

So I do not really see the purpose of this question. :)

Looks like the focus is on, but it might be a bit soft due to a low shutter speed.

[arash_hazeghi]

the web image looks quite ok, but you can almost make every image look good on web, even if it's way OOF....

So I do not really see the purpose of this question. :)

Looks like the focus is on, but it might be a bit soft due to a low shutter speed.

Jan the purpose of this question is a $100 bet I made with someone ;) and there is "not sure" choice too if you think it is not clear if the original was in focus. Do you think a shot that is way oof can be recoeverd by sharpening?

[Stuart Hill]

Hi Arash,

Difficult to tell at this size. Think all the areas that are out of focus helps cheat our eyes in to thinking the 'sharp' areas are sharper than they actually are, if that makes sense?

regards.
Stu.

[Doug Brown]

I wouldn't say it's critically sharp, but it is salvageable.

[arash_hazeghi]

Thanks guys, I just need 30 votes before I can win :) or lose :eek:

Also keep in mind DOF at 700mm f/5.6 :D

[Jay Gould]

Arash, the problem with the choices is not only are they not the same :p, they are critically different! :eek:

While you ask if it is "in focus" you also ask if it is "out of critical focus".

What is "critical focus"? For a bird it is the eye and the immediate surrounding area.

Comparing this image to some of your other images, this image is "out of critical focus" - it is a bit soft around the eye - certainly recoverable I would think.

[Charles Glatzer]

I consider this out not to be critically sharp. In focus, perhaps. Either way the overall image is soft, as presented.

Agree with Jim below.

Chas

[Jim Neiger]

In focus and sharp are two different things. The focus seems to be ok, but I think there is some slight motion blur. I answered "In foucs".

[John Chardine]

The plane starting at the top of the head running down the back of the bird with a DoF about the width of the body looks in focus to me, as presented. With the head angle the way it is, the eye may be just at the front edge of this DoF.

[Mital Patel]

as what the dof says along with the focal length i vote this in focus for your bet.

[Ed Cordes]

I voted "in focus". I think the DOF and motion blur are acting up. On my work monitor the point of best focus seems to be the top of the neck just behind the head. So, the eye sufferes from lack of DOF. IMHO

[David Smith]

I would vote for "in focus" however I think you need a little more DOF to get the wing tips..

Daver

[Nick Palmieri]

I went with in-focus some DOF issues and some motion blur...certainly acceptable focus on the head

[arash_hazeghi]

Thanks guys I won the bet :)

The original photograph is out of focus, with focus being a few feet closer than the bird, the body is soft but within the soft body wing tips are even softer due to motion, I exploited this fact and sharpened up the head in post :D

Here is the original shot, focus is 4-5 ft closer than the bird.



Jan is right, unless full size image is shown it is not possible to assess focus or sharpness, this is the point I wanted to make to my friend :) sure he will agree now :D

Thanks for participating!

[John Chardine]

How can you tell from the image that the bird is not flying directly above the in-focus grass, or at least within the DoF you obtained? As DoF is determined as well by distance to subject, which is unclear here, you may have had plenty.

[arash_hazeghi]

How can you tell from the image that the bird is not flying directly above the in-focus grass, or at least within the DoF you obtained? As DoF is determined as well by distance to subject, which is unclear here, you may have had plenty.

John I can tell because I was there and I took the photo :)
You can see the squirrel tail in the lower left corner, the squirrel was way closer than harrier. You can easily tell this by comparing the relative sizes.

I have never seen a harrier attack a squirrel but this was a hot chase at least, after this harrier banked and chased the squirrel for sometime, unfortunately I could not get any pics :( I got one good shot before the chase I will post in the avian forum.

Here is another one

[Axel Hildebrandt]

Why don't you post an unprocessed vertical 100% crop with bird and grass underneath. Are you saying that the critical focus is in front of the near wing?

[arash_hazeghi]

Why don't you post an unprocessed vertical 100% crop with bird and grass underneath. Are you saying that the critical focus is in front of the near wing?

Yes Axel that is where it is.

I deleted all the oof RAWs last night, just kept a small jpeg for this thread-sorry. I will post the good one in the avian forum.

Any ways, I just showed the poll results to my friend and he agreed. Won the bet :)

I forgot to mention I used nearest neighbor method when down-sampling the original file to harden the edges and enhance the visual sharpness. I used PS NR to clean up the noise that resulted. I also enhanced the soft eye with the brush tool :D

Thank you.

[arash_hazeghi]

I posted one keeper in the avian forum, notice the entire harrier is is sharp focus despite f/5.6

http://www.birdphotographers.net/for...853#post444853

Hope you like this image :)

[John Chardine]

John I can tell because I was there and I took the photo :)
You can see the squirrel tail in the lower left corner, the squirrel was way closer than harrier. You can easily tell this by comparing the relative sizes.

I have never seen a harrier attack a squirrel but this was a hot chase at least, after this harrier banked and chased the squirrel for sometime, unfortunately I could not get any pics :( I got one good shot before the chase I will post in the avian forum.

Here is another one

Sorry Arash but I remain unconvinced. In this one, where BTW, the bird could have moved in relation to the OP, the in-focus grass appears to lay to the front, underneath and to the rear of the bird so figure that one out.

We biologists are a skeptical lot. It goes with the territory.

[arash_hazeghi]

Sorry Arash but I remain unconvinced. In this one, where BTW, the bird could have moved in relation to the OP, the in-focus grass appears to lay to the front, underneath and to the rear of the bird so figure that one out.

We biologists are a skeptical lot. It goes with the territory.

John I apologize but my goal was not and is not to convince you about a picture that I took. I did not post this to "debug" anything, I wanted to poll the average opinion and win a bet which I did.



Thanks

[Ed Cordes]

Arash, I think this was a good exercise. It certainly shows how subjective these things can be. The photographers here are among the best there are and we can't all agree on how to judge an image. BTW, your post of the "good" image in the other thread is super. I wish I could find a location to get close enough to make an image like these. Hawks around here seem to fly as soon as they see a lens.

[arash_hazeghi]

Arash, I think this was a good exercise. It certainly shows how subjective these things can be. The photographers here are among the best there are and we can't all agree on how to judge an image. BTW, your post of the "good" image in the other thread is super. I wish I could find a location to get close enough to make an image like these. Hawks around here seem to fly as soon as they see a lens.

Thanks Ed,

Hawks here (NorCal) are relatively easy to approach, they know no one's gonna mess up with them ;). Harriers are not easy to approach what I do is I use my binoculars first to spot their flight path, they fly in large circles, then I try to position myself in a location where I have higher chances of getting them against a nice BG and wait until they come to me :D Of course it can be very frustrating and sometimes I go home empty handed but if you have patience you will get something :) good luck

[Ed Cordes]

Arash, thanks for the suggestions. There is a place a few hours North of me where in the Spring Red Tails hunt in a field. I really need to go early and wait. Thanks again.

[Jeff Donald]

Critical focus is the eye, and yes the eye is critically sharp and in focus. However, because of the focal length of the lens, distance to subject and aperture, you ran out of DOF and the wing tips are OOF.

[Jay Gould]

Critical focus is the eye, and yes the eye is critically sharp and in focus. However, because of the focal length of the lens, distance to subject and aperture, you ran out of DOF and the wing tips are OOF.

I guess that is what makes horse racing; I do not find the eye critically sharp and in focus even after the extra sharpening.

[arash_hazeghi]

Critical focus is the eye, and yes the eye is critically sharp and in focus. However, because of the focal length of the lens, distance to subject and aperture, you ran out of DOF and the wing tips are OOF.

Hi Jeff,
missed the answer! I fixed the eye in photoshop :D (Eye doctor as Artie says) you can get the whole bird in dof from this distance checkout the other shot, this one was even closer and sharp tip to tip;) the photo above in this thread is just front focus. PS does wonders!


http://www.birdphotographers.net/for...ad.php?t=56993

[Jeff Donald]

Far wing tip is still out of focus in the shot you linked to.

[arash_hazeghi]

Far wing tip is still out of focus in the shot you linked to.

that's motion blur I think (1/1000), near one is sharp, this isn't full frame btw, it is about 40% of RAW.

[Jeff Donald]

Motion should affect the far tip as much as the near tip, I think you ran out of DOF. I also find the bright halo around the bird distracting, too much sharpening?

[arash_hazeghi]

Motion should affect the far tip as much as the near tip, I think you ran out of DOF. I also find the bright halo around the bird distracting, too much sharpening?

It depends on the motion, I have to check the RAW again but it is totally possible that one wing has more motion blur than then the other, have seen ti many times.

The halo is because of slight overexposure, I had originally metered for darker grass but harrier changed course and didn't have enough time to adjust. I will try to tone it down more, not much sharpening applied to this one, smart sharpen (80,0.5)...Not such a brilliant image technically just posted for Axel, I still like the wing position and harrier's HA, in my book expression, pose and BG comes before technical perfection, but that's just me.

[David Smith]

Maybe someone can explain something to me.. In your full frame shot, which is including the histogram and focusing sensors, why is the focusing sensor on the Hawk's back illuminated if that is not the focus point and second, how could the camera focus on the grass when there is no focusing point anywhere near the grass that you say was the focus point..:confused::confused:

Dave

[arash_hazeghi]

Maybe someone can explain something to me.. In your full frame shot, which is including the histogram and focusing sensors, why is the focusing sensor on the Hawk's back illuminated if that is not the focus point and second, how could the camera focus on the grass when there is no focusing point anywhere near the grass that you say was the focus point..:confused::confused:

Dave

The red frame is where the camera is supposed to focus on, but the camera can focus at any distance based on the solution its servo algorithm converges to. This distance maybe quite different from actual object distance sometimes.

[Jay Gould]

The red frame is where the camera is supposed to focus on, but the camera can focus at any distance based on the solution its servo algorithm converges to. This distance maybe quite different from actual object distance sometimes.

Now you have lost me too! I thought that the red box that we look at after the shot is taken is in fact where the camera focused. Now you are suggesting that the red box is not necessarily where the camera in fact focused.

[arash_hazeghi]

Now you have lost me too! I thought that the red box that we look at after the shot is taken is in fact where the camera focused. Now you are suggesting that the red box is not necessarily where the camera in fact focused.

Jay you need to read a bit about how AF system works. The red box indicates the active AF sensor when the shot was taken, the camera does what it can to focus on the feature within the box that has high contrast. However like any other system AF has imperfections and weaknesses, it can miscalculate distance or speed and therefore actual focus plane can be shifted forward or backward relative to the intended object. The red box just shows the active AF sensor.

[David Smith]

Jay:

I'm with you... Maybe a Canon Rep. can come in here and give us a quick explanation of something that is apparently confusing some of us..

Dave

[Ed Cordes]

Jay:

I'm with you... Maybe a Canon Rep. can come in here and give us a quick explanation of something that is apparently confusing some of us..

Dave

I am not a Canon Rep, but I think what Arash is saying is that when you point the selected AF point at the spot you want to AF on, and it is moving, the camera's computer must determine what it is you intend to focus on and then based on the camera's analysis of the spot's movement run a calculation and predict where that spot will be milliseconds in the future. So, as fast as modern electronics are, it is possible for the subject to move quickly in a direction that was not predicted by the computer. This probably doesn't occur often, but it does occur. At least that's my take on it.

[Roger Clark]

I'll add may take on it.

First, the AF sensor indicator is just that, an indicator. The actual AF spot maybe slightly misaligned and is usually larger the the red rectangle.

If you are in AI servo mode, as focus changes, the AF sensor is indicating a change in the phase from the light coming through two apertures. The light that hits the sensor is subject to noise (mostly that due to light itself) so the intensity distribution of light on the phase detectors may change due to that noise, and that will lead to a different velocity than reality.

Next, the subject may be accelerating by changing direction toward or away from the camera.

Finally, as you track the subject and you move around on the subject, you a causing the AF sensor to look at different distances, thus changes in velocity.

The AF tracking system projects a constant velocity (it is possible to project acceleration but much more prone to errors, so I don't think that is done; would need Canon to say for sure). Iether way, a constant velocity or a constant acceleration profile is subject to error projecting into the future because neither may be constant.

The bottom line is that when tracking moving subjects, it is important to know not oly which focus point was active, but where the focus point was in the previous time intervals when the velocity was determined. For example in Arash's images, in evaluating a focus error, it is important to see where the focus points were in the previous ~3 frames if shooting continuously.

It is quite important tracking moving subjects to track them smoothly. Jerking movements confuse the AF computations and lead to greater errors in projected focal point. Image stabilization is a large factor here, not just for the image when you trip the shutter, but also for smoothing the tracking to give more accurate AF and AF predictions. At high shutter speeds some people turn off IS, but that might harm AF. Note there are some reports that on older IS systems, high shutter speeds harms image quality.

Roger

[Charles Glatzer]

Now you have lost me too! I thought that the red box that we look at after the shot is taken is in fact where the camera focused. Now you are suggesting that the red box is not necessarily where the camera in fact focused.

In One-shot... if you manually select a AF point, lock focus, and thereafter recompose the AF point will move off target as you reposition the subject within the viewfinder. In this case the "red box" will not represent the original point of focus, only the AF point position selected in the viewfinder.

In Servo... as far as I know the "red box" will represent the point of focus when depressing the shutter. But, I am open to enlightenment.

Best,

Chas

[Jim Neiger]

In One-shot... if you manually select a AF point, lock focus, and thereafter recompose the AF point will move off target as you reposition the subject within the viewfinder. In this case the "red box" will not represent the original point of focus, only the AF point position selected in the viewfinder.

In Servo... as far as I know the "red box" will represent the point of focus when depressing the shutter. But, I am open to enlightenment.

Best,

Chas

You can use AF lock when in AI Servo and it will lock the focus while the AF Lock button is depressed. I use this method often since it allows me to recompose in camera for still subjects while remining in AI Servo and ready for action. This is really needed when using only the center AF point. In this case the red box will not represent where the camera was focused even though you are in AI Servo.

[Jim Neiger]

Another cause for the focal plane differing from the position of the active focus point is the tracking sensitivity delay. For example: If the focus was on the grass in fron of the bird and then you position the focus point on the bird and immediately shoot, the delay may not have expired and the focus may still be on the grass. The delay may be overridden by bumping the focus. So when the camera was focused on the grass bumping the focus back to the bird would have allowed the shot right away with the focus on the bird as intended.

[Charles Glatzer]

You can use AF lock when in AI Servo and it will lock the focus while the AF Lock button is depressed. I use this method often since it allows me to recompose in camera for still subjects while remining in AI Servo and ready for action. This is really needed when using only the center AF point. In this case the red box will not represent where the camera was focused even though you are in AI Servo.

Jim,

I should have elaborated a bit more. I use the * to lock focus when in AI Servo all the time, as you describe above. As you know you can also choose to set the back button to activate focus in either mode , and deactivate focus by simply taking your finger off the button.

My point was that when tracking in Servo the "red box" will as far as I know represent the point of active focus at the time of shutter depression

Best Amigo,

Chas

[Roger Clark]

My point was that when tracking in Servo the "red box" will as far as I know represent the point of active focus at the time of shutter depression

Chas,
I agree, but the active focus is not necessarily where the true focus is actually at, nor where it is projected to be at the time the shutter actually fires, due to the things I raised in my post above.

Roger

[Charles Glatzer]

Another cause for the focal plane differing from the position of the active focus point is the tracking sensitivity delay. For example: If the focus was on the grass in fron of the bird and then you position the focus point on the bird and immediately shoot, the delay may not have expired and the focus may still be on the grass. The delay may be overridden by bumping the focus. So when the camera was focused on the grass bumping the focus back to the bird would have allowed the shot right away with the focus on the bird as intended.

So, the "red box" only shows the active AF point, not the point of active focus. I have never used the "red box" as I find it irrelevant (and now even more so) the image is or is not in focus.
Thanks for the info, greatly appreciated.

Chas

[arash_hazeghi]

Canon "AI servo" AF system uses predictive focus tracking, the error signal that is read by the AF sensor which is proportional to the phase difference between the split images is recorded continuously and then analyzed by a separate processor in 7D camera. the camera then uses curve fitting or statistical correlation to calculate subject velocity vector and predict its trajectory.

If trajectory is random as opposed a simple line or curve (like the flight path of a harrier that banks and intermittently stops) camera cannot extrapolate the trajectory correctly so when servo is pulsed and error signal is read again, the trajectory solution will not converge and camera will focus at a random distance.

In good light and high contrast situation like the example I posted above, there is not much noise in the AF sensor's reading but the camera cannot predict trajectory and has focused on a different plane no matter how long it has been tracking the subject before, in the above case the camera had been tracking the subject for 6-7 frames 2 of which were in focus, this is a few frames before:



Bumping the shutter button causes the camera to reset the tracking and try to pick the subject again. I have been using this technique since the 20D and still use it with 5DMKII, for some reason it doesn't seem to work very well with 7D, in particular bumping in this case usually leads to focus being locked on the BG and no amount of bumping will pick the BIF unless BG changes.

The other issue is servo design, in order to make AF drive fast, canon have chosen to use a slightly under-damped servo which overshoots, this is why the initial frame in a sequence of shots is usually OOF despite the fact that servo drive is fast enough to grab the same shot in single shot AF. You may be able to test this at home, set your camera to AI-servo mode and point at a stationary object with medium contrast then half-press and hold the shutter button, then carefully look at the distance window on top of the lens, sometimes you will see small oscillations, this instability causes slightly OOF photos within a sequence. I wish to see Canon change this approach and trade some speed for stability. Nikon for examples use overlapped-servo operation that provides intermittent feedback to increase AF precision.

This is explained in this article http://www.nikon.com/about/technolog.../caf/index.htm which schematically shows the operation of CAM3500FX module.

Here is another article which explains the concept and operational basics' of Canon's AI-servo

http://photonotes.org/other/ai-servo.html

very basic stuff but hope it is useful to some.

[arash_hazeghi]

So, the "red box" only shows the active AF point, not the point of active focus. I have never used the "red box" as I find it irrelevant (and now even more so) the image is or is not in focus.
Thanks for the info, greatly appreciated.

Chas

That's correct, red box is the active sensor at the time shutter was depressed. However it is sometimes useful for diagnosing AF problems, for example checking to see if active AF sensor or its expansions were on the intended subject.

[John Chardine]

Have a look at this thread I started:

http://www.birdphotographers.net/for...924#post445924

Looks like the expansions don't show unless you have the C.Fn set to "expand to all 45"