Autofocus mechanism

[Stoshowicz]

Are you talking about the purple-ish halo around some of the edges? That's Chromatic Aberration. It occurs at intersections of high contrast and most often manifests in shades of purple or green. In this case it's purple. It's pretty easy to remove, typically, using your post processing software. Adobe Camera RAW has some sliders in the Lens Corrections module dedicated to removing CA.

....

If already familiar with what youre saying, I also know about the post-processing possibilities and their limitations, the sliders often do not remove the effect completely and tend to desaturate according to hue which can affect colors elsewhere in a photo. One can also use the PS paintbrush for color correction but that takes a bit of time to do neatly - its not a ten second fix in many cases.
Look closer please at the photo example provided. You see that the aberrant color is mostly manifest on the near wing , but not on the far wing or shoulder.. and is still present albeit more mildly about the head and bill. I was asking why it is where it is - ,as you can see for yourself, not all areas of the same contrast values have the same color problem.

 

[Stoshowicz]

What Mr. Fish said.

It's how the light makes it thru the lens and different colors that make up the same point on the subject not always getting focused on the same point in the image. Sometimes worse with digital sensors. Sometimes easy to fix and other times not.

I am guessing it is different on different edges because light reflects as different colors off of different angles of the same object and the wavelength of light is different with different colors.

There is a lot of reading out there on the causes and what can be done. Lenses are designed to reduce it and some more so than others.

So do You feel that these are reflections which occur only in situations of high contrast having to do with subject surface color ? ? Or do you feel that its a lens problem?
There may indeed be a lot written on the subject , I chose to post the question and just ask someone who has already read up on the issue.

 

[Stoshowicz]

Been studying on bird flight and freezing their variety of motions. Here is the blackbird with the PS grid turned on and set to the size of the D7100 sensor. The main grid lines represent millimeters and the lesser tenths of millimeters. So, this is the image of the bird sized to the sensor.

Shutter Travel: The bird size on sensor is roughly 5.72 mm. Shutter curtain with a travel time of 1/250 top to bottom will pass the bird in 1/681 second. Just considering roughly 1 inch of wingtip (about 1 mm on the sensor), the curtain will pass it in 1/3900 second and pass each tenth millimeter of image in 1/39,000 second.

In a shot like this the flight speed is offset by panning the camera and effectively reduced to where the body and head is perceptibly sharp. Wings are a different matter. Depending on the movement and angle of movement, the wingtips move across the sensor at a variety of speeds, but the fastest would seem to be the forward/upward movement across and parallel with the sensor plane.

Assuming a few numbers, such as 11.25" average b-bird wing length, WAG of 17.67" wingtip travel and 10 wing beats per second gives an average wingtip speed of 15.6 meters per second in real life. The bird is reduced in size by the lens and this reduces the speed relative to the shutter travel speed (noted above). The duration of the shutter speed then captures a fraction of the wingtip movement of roughly 0.10 mm. Hard to tell because the screen capture of the blackbird is not so high res, but it looks like the wingtip may have moved about that much.

Neatly described , thanks , I then surmise then that the wingblur is explainably entirely due to insufficient shutterspeed. So even at this relatively high shutterspeed its still just too slow to reliably get a crisp moving wing. It really has to be done at close range with a strobe.
Ok , so be it. Did you find any good source you can disclose on the subject?

( interesting note though- I calculated the wingtip of the duck right about the same speed,, 35mph!)

 

[Marcel]

So do You feel that these are reflections which occur only in situations of high contrast having to do with subject surface color ? ? Or do you feel that its a lens problem?
There may indeed be a lot written on the subject , I chose to post the question and just ask someone who has already read up on the issue.

I'm almost certain that the problem you're looking at is caused by Chromatic aberration (CA). What you didn't write, is which lens you were using for that shot. Some lenses are just better than others for that point.

 

[Stoshowicz]

I'm almost certain that the problem you're looking at is caused by Chromatic aberration (CA). What you didn't write, is which lens you were using for that shot. Some lenses are just better than others for that point.

Its the Nikon 70-300 VR , I know its not great for the CA issue especially wide open. The 200mm seemed to show it less. Ive got the tamron 150-600mm on order , which looks like it has the issue better controlled. Even so , being clear on what exactly is going on with the CA should still be handy to know when Im using other lenses.

 

[Marcel]

The thing you have to remember concerning VR is that the lens does NOT react to subject movement, but to camera movement. But for birds in flight, you should use a fast shutter speed and forget about using VR IMHO.

 

[Eyelight]

So do You feel that these are reflections which occur only in situations of high contrast having to do with subject surface color ? ? Or do you feel that its a lens problem?
There may indeed be a lot written on the subject , I chose to post the question and just ask someone who has already read up on the issue.

This may sound like a cop-out, but everything about an image is a combination of light, lens, sensor and processing. So, yes it is subject related and yes it is lens related and possibly a little sensor and/or processing related. Not necessarily a lens problem, but some lenses will do better than others and the same lens may do better by changing aperture.

Looking at the purple fringe on the near wing, it appears the effect may be enhanced due to the angle the light coming from above and behind the bird. Is this because there is more exposure maybe, in combination with a little blur.... maybe??? So the lens delivers this slightly blurred edge that's over-exposed compared to the pixel next door and it shows up at the sensor minus some color, so the sensor sees purple.

Can you see this in the final display of the image?

 

[Eyelight]

Neatly described , thanks , I then surmise then that the wingblur is explainably entirely due to insufficient shutterspeed. So even at this relatively high shutterspeed its still just too slow to reliably get a crisp moving wing. It really has to be done at close range with a strobe.
Ok , so be it. Did you find any good source you can disclose on the subject?

( interesting note though- I calculated the wingtip of the duck right about the same speed,, 35mph!)

I did do a little research on flight speed, wing beats and bird anatomy, but there is some guesswork and maybe a smidgeon of speculation. One thing to bear in mind is that the wing speed can vary dramatically depending on the movement.

Can't rule out DOF unless you have a small bird in a big field, which is rarely the case, becasue we zoom to the subject size and collapse the DOF on the subject.

I suspect that wingtip speed will be similar across various sizes of birds because the slower beat is offset by the longer span.

 

[Stoshowicz]

This may sound like a cop-out, but everything about an image is a combination of light, lens, sensor and processing. So, yes it is subject related and yes it is lens related and possibly a little sensor and/or processing related. Not necessarily a lens problem, but some lenses will do better than others and the same lens may do better by changing aperture.

Looking at the purple fringe on the near wing, it appears the effect may be enhanced due to the angle the light coming from above and behind the bird. Is this because there is more exposure maybe, in combination with a little blur.... maybe??? So the lens delivers this slightly blurred edge that's over-exposed compared to the pixel next door and it shows up at the sensor minus some color, so the sensor sees purple.

Can you see this in the final display of the image?

Well sure, it has something to do with camera lens and subject,
Can I see it in the final image ? sure ! You take photos , can't you see if the edges of your subject are blurry or have blue lines around it?
Once you compress and resize a photo down to one or two megapixels theres a lot you cant see.

This here is from another thread , you think this guy would notice blue lines around the subject?

"Things that stick out for mine instantly

1) You lost some of the bristles on the brush on the left. Too tightly cropped or missed in composition
2) My eyes are drawn to the upside down B.... if that was your intention, maybe turn the pen so the letter is up the right way"

I dont see why I should keep having to defend trying to produce better photos rather than being satisfied with slop.

 

[J-see]

Those yellow and purple distortions are CA indeed. When I was first using the 70-300mm a good part of the macros I did with it showed those distortions. I didn't understand what caused them until I took off the UV filter. The CA disappeared too. Yellow and purple is CA for sure. In my case the UV filter caused them but only when there was plenty of contrast.

 

[Stoshowicz]

Those yellow and purple distortions are CA indeed. When I was first using the 70-300mm a good part of the macros I did with it showed those distortions. I didn't understand what caused them until I took off the UV filter. The CA disappeared too. Yellow and purple is CA for sure. In my case the UV filter caused them but only when there was plenty of contrast.

Yep , we are all in agreement with it being a chromatic aberration. Though I have been using the 70-300 and like the lens a lot, I agree that the CA isnt controlled great on the lens. I don't even have a UV filter. Are you suggesting that the phenomenon is UV related ? or that it has something to do with polarization of the light.

 

[Horoscope Fish]

If already familiar with what youre saying, I also know about the post-processing possibilities and their limitations, the sliders often do not remove the effect completely and tend to desaturate according to hue which can affect colors elsewhere in a photo. One can also use the PS paintbrush for color correction but that takes a bit of time to do neatly - its not a ten second fix in many cases.
Look closer please at the photo example provided. You see that the aberrant color is mostly manifest on the near wing , but not on the far wing or shoulder.. and is still present albeit more mildly about the head and bill. I was asking why it is where it is - ,as you can see for yourself, not all areas of the same contrast values have the same color problem.

I never said it was "a ten second fix"; I said it was, typically speaking, easy to do.

As to your question about why the CA is occurring where it is, I can tell you a lens changes the speed of the light moving through it. As the light travels through the lens at these now differing speeds, the separate wavelengths (colors) are moving at different speeds and fall on minutely different locations on the sensor. This displacement of wavelengths on the camera sensor is what causes chromatic aberration.

...

 

[Eyelight]

Once you compress and resize a photo down to one or two megapixels theres a lot you cant see.

Precisely why I asked. You can see things in the editor or when you print an enlarged image that we cannot see online.

This here is from another thread , you think this guy would notice blue lines around the subject?

"Things that stick out for mine instantly

1) You lost some of the bristles on the brush on the left. Too tightly cropped or missed in composition
2) My eyes are drawn to the upside down B.... if that was your intention, maybe turn the pen so the letter is up the right way"

I dont see why I should keep having to defend trying to produce better photos rather than being satisfied with slop.

Lost me completely. The referenced comment appears to be about composition.

I could not see the CA until you posted the crops. Another reason I asked if they were visible in the final display of the image. Simple question; not a critique.

 

[Eyelight]

@Stoshowicz , for what it's worth, I've learned a few things while considering your questions and studying the why's and why-not's of it all. So, I appreciate the discussion.

 

[Stoshowicz]

@Stoshowicz , for what it's worth, I've learned a few things while considering your questions and studying the why's and why-not's of it all. So, I appreciate the discussion.

I gotta say as well that combining the things you guys are saying with google searches to fill it out is rather interesting.
Purple fringe is but one type of CA , If I read it correctly , purple fringe happens because of differential activation of the various components of the bayer arrays. CA also can be caused by UV, polarized specular reflection ,and infrared light , some folks seem to suggest that polarized UV filters could be of help , and others seem to say that its detrimental. -I dont get that- And one can also get some color bleed over due to the misaligned red blue and green focus distances. SO it appears that this CA thing is actually a lot more complicated than I first thought. I thought it was just a no brainer thing but its not as simple as just slapping on a filter or angling for a particular direction of light.
One thing said that one should correct the issue in camera raw by doing something with the color channels before saving as a tiff or Jpeg, because once they were converted to those formats the colors were now not combinations of the bayer array components but numerical depictions of the summed up "color" rendering. OH NO Ive been doing it wrong a long long time!
This is going to take some serious digging.

 

[J-see]

Yep , we are all in agreement with it being a chromatic aberration. Though I have been using the 70-300 and like the lens a lot, I agree that the CA isnt controlled great on the lens. I don't even have a UV filter. Are you suggesting that the phenomenon is UV related ? or that it has something to do with polarization of the light.

I only had it while using the UV on lens and taking shots in high contrast (lots of light) areas. Once I removed it, I didn't have any issues. But I never really tested if the filter magnified or caused the CA. I can take it out during the next days and shoot some high contrast shots to see if it reappears. It's related to brightness that's all I'm certain about. I've never seen it when I took shots without "sky".

In my case bugs suffered it, legs and wings had purple and yellow outlines but only during bright sunlight.

 

[Stoshowicz]

I only had it while using the UV on lens and taking shots in high contrast (lots of light) areas. Once I removed it, I didn't have any issues. But I never really tested if the filter magnified or caused the CA. I can take it out during the next days and shoot some high contrast shots to see if it reappears. It's related to brightness that's all I'm certain about. I've never seen it when I took shots without "sky".

In my case bugs suffered it, legs and wings had purple and yellow outlines but only during bright sunlight.

I spent a couple hours researching this last night. there are different kinds and causes all due to the lens handling of various wavelength light.
You and I are getting a lateral (radial) misalignment of the warm and cool tones.. which gets worse as one looks closer to the periphery moving out from the center. It has the effect of offsetting the colors which is why you see two colors- one on each side . The camera raw lens correction should work pretty good at realignment since the ' individual ' colors arent really out of focus. There are also two possible axial forms of CA, which arent what we have. With purple fringe blue is focused behind the sensor , the greens are smack on the sensor , and the reds are focused in front of the sensor. since the reds and blues are out of focus, they are both spread out as larger images than they should and since they overlap without the greens -what you see is a single color purple halo. This isnt as easily correctable since even if you blurred the green channel to the same extent ( so as to line up with the other two) your image would be softened to the same extent.

( the other axial CA shows as soft radial color bands at distance from the center of the lens like a green magenta rainbow.

I'd suspect that the UV filter could be misaligned -tilted cross-threaded or just faulty - causing the increase in lateral CA, because knocking out the UV should actually be eliminating a wavelength of light which isnt usually well handled for CA , by lenses, but may affect sensors.

 

[J-see]

Thanks for that info. I just assumed it to be CA, cursed at the UV filter before throwing it in the trash and that was that.