D7100 Inconsistent Focus Tracking Errors

[Eyelight]

Thank you for having a look at this for me i appreciate it and feel welcomed by all on this site, i have been a long-ish time lurker and read all the DX forums but then decided to join and now i am glad i have, anyway -
Do you think it would help to up the shutter speed to 1250-1500 esp as i crop a fair bit? and is it logical that the reason i did not have these troubles with either my D200 or D90 with my 55-300ED VR was down to the resolution difference of 8/12 compared to 24 or 16 if i use the crop mode in the D7100>
I had read that the higher resolution of these high power sensors puts a lot of stress on the lens and the users ability to get the best out of them, and that our technique needs to improve compared to the lower older sensors like my D90/D200.
Does any of that make sense?
Also sometimes i used to get better focus with grab shots where i fired as soon as a plane was in the frame roughly where i wanted it without me following or panning?

P.

Faster shutter speed would help eliminate any possible shake and help some with the planes motion. Good panning technique is essential because the pan speed reduces the need for shutter speed. If you are panning from straight ahead by raising both elbows, I can't imagine that would be always smooth.

Given the same conditions more pixels will be blurred on the D7100 than the D90200 (just did that for fun), but the same area would be blurred on the D90200. You would just be losing detail on the D7100 that you never had on the D90200. In other words, it's a pretty darn small detail that you would be losing, but it does exist.

Kinda sorta makes sense that you would need better technique to achieve finer detail for a given camera and if the camera suddenly improved its detail capturing ability (swapped the D90200 for the D7100), your technique could be lacking.

Grabs shots are luck of the draw with perhaps the advantage of not thinking too long.

 

[Paganman2]

Faster shutter speed would help eliminate any possible shake and help some with the planes motion. Good panning technique is essential because the pan speed reduces the need for shutter speed. If you are panning from straight ahead by raising both elbows, I can't imagine that would be always smooth.

Given the same conditions more pixels will be blurred on the D7100 than the D90200 (just did that for fun), but the same area would be blurred on the D90200. You would just be losing detail on the D7100 that you never had on the D90200. In other words, it's a pretty darn small detail that you would be losing, but it does exist.

Kinda sorta makes sense that you would need better technique to achieve finer detail for a given camera and if the camera suddenly improved its detail capturing ability (swapped the D90200 for the D7100), your technique could be lacking.

Grabs shots are luck of the draw with perhaps the advantage of not thinking too long.

I am just trying to get my head around if there is an equation for working out a required shutter speed for a given lens length x the distance away from camera x the speed of the objects movement x the amount of crop of the picture for viewing, as i understand the larger you crop and see more of an object, the more errors- focus/movement show up?
So if there was a way of selecting a safe shutter speed for this it would be a great help.

Great Avatar and name by the way - very spiritual.

P.

 

[J-see]

I am just trying to get my head around if there is an equation for working out a required shutter speed for a given lens length x the distance away from camera x the speed of the objects movement x the amount of crop of the picture for viewing, as i understand the larger you crop and see more of an object, the more errors- focus/movement show up?
So if there was a way of selecting a safe shutter speed for this it would be a great help.

Great Avatar and name by the way - very spiritual.

P.

It's pretty hard to calculate the required shutter unless the distance and speed is always identical. You basically need to know how far the plane travels on your sensor in a second and then it's that 1 second divided by the amount of pixels the subject travels. That will be a perfect freeze shutter-wise.

You can guess based upon what you see in the viewfinder and then you know the minimum shutter required. If it travels 1/4th of the sensor horizontally in that second, you'd need at least 1/1500s. If you crop less, you can get away with some pixel blur.

By panning, you can lower the required shutter since you lower the amount of pixels traveled during the shot.

 

[Paganman2]

It's pretty hard to calculate the required shutter unless the distance and speed is always identical. You basically need to know how far the plane travels on your sensor in a second and then it's that 1 second divided by the amount of pixels the subject travels. That will be a perfect freeze shutter-wise.

You can guess based upon what you see in the viewfinder and then you know the minimum shutter required. If it travels 1/4th of the sensor horizontally in that second, you'd need at least 1/1500s. If you crop less, you can get away with some pixel blur.

By panning, you can lower the required shutter since you lower the amount of pixels traveled during the shot.

Very interesting, i was just trying to work that out from a plane i shot from memory - i think it flew diagonally from corner to corner in about 4 seconds, then i wanted to crop it 50% would that require double the shutter speed you mentioned, so something like 1/3000sec?

P.

 

[J-see]

Very interesting, i was just trying to work that out from a plane i shot from memory - i think it flew diagonally from corner to corner in about 4 seconds, then i wanted to crop it 50% would that require double the shutter speed you mentioned, so something like 1/3000sec?

P.

On a 6k*4k sensor, the diagonal is 7211 pixels. That's 7211/4 = about 1/1800s for pixel freeze if you crop to 100%. When 50% you should get away with half of that, roughly.

Keep in mind not all cams see the full picture in their viewfinder.

 

[Paganman2]

On a 6k*4k sensor, the diagonal is 7211 pixels. That's 7211/4 = about 1/1800s for pixel freeze if you crop to 100%. When 50% you should get away with half of that, roughly.

Keep in mind not all cams see the full picture in their viewfinder.

This is very useful information i will write it down.
Does lens reciprocal length come into this eg - the 1/1 rule for speed v lens length taking into acount my 300mm lens on my d7100?

P.

 

[J-see]

This is very useful information i will write it down.
Does lens reciprocal length come into this eg - the 1/1 rule for speed v lens length taking into acount my 300mm lens on my d7100?

P.

Basically nothing matters but the pixels on the sensor and the amount it travels during a second. That's your required shutter speed for a perfect freeze.

You need to know how long a second lasts which requires some practice if you don't. We learned that as kids by counting the seconds between lightning and thunder to calculate the distance.

Once you control those two values, it's simply getting good at guessing. You can use the focal points as indicators since they have a fixed position on the sensor.

The larger your subject (when zooming), the faster it'll travel and the more you crop, the more correct your shutter has to be.


To add: this is under the assumption you don't add shake during exposure. When you move the lens in the wrong direction during, you can affect the distance traveled. At fast shutters it'll not make much difference but at slower shutters it does.

 

[Eyelight]

When panning it is the difference between pan speed and subject speed that is required to freeze the motion. If the pan speed perfectly matches the subject speed, then a fairly slow shutter can be used; even slower than required for a handheld shot at a static subject.

It's the roughness of the panning that needs to be overcome, similar to shake for a static subject. I'd say 2 times the focal length for a smooth well matched panning motion and faster anytime the panning cannot be smooth and well matched. Looking at your good shots will tell you quite a bit, both with the right speeds and confirming the ideas about whats going on.

 

[Paganman2]

Basically nothing matters but the pixels on the sensor and the amount it travels during a second. That's your required shutter speed for a perfect freeze.

You need to know how long a second lasts which requires some practice if you don't. We learned that as kids by counting the seconds between lightning and thunder to calculate the distance.

Once you control those two values, it's simply getting good at guessing. You can use the focal points as indicators since they have a fixed position on the sensor.

The larger your subject (when zooming), the faster it'll travel and the more you crop, the more correct your shutter has to be.

So i guess - if the planes i see and shoot fly at nearly the same speed with exception to other more random planes, then i can get a rough guide to set shutter speed by, and going by 100% crop i simply deduct downwards accordingly?

P.

 

[J-see]

So i guess - if the planes i see and shoot fly at nearly the same speed with exception to other more random planes, then i can get a rough guide to set shutter speed by, and going by 100% crop i simply deduct downwards accordingly?

P.

If you're shooting close to an airport, direction and speed will be fairly similar for most and when calculating one, you can use that as a starting point for the other. But like Eyelight said, if you pan, it's different.

When I say 100% crop, it's when you look at a shot that is displayed at the same size as the pixels it contains. If you'd crop something to let's say 1200*800 pixels and save it as that size, you require a perfect pixel freeze. But if you crop to 2400*1600 and save that as a 1200*800 you can get away with a blurred pixel. Then it may travel 2 pixels during that same shutter and you'd likely not notice in the finished shot.

Remember: this is not exact science since we guess the time and distance so you have to find out what works best for you.

 

[Paganman2]

If you're shooting close to an airport, direction and speed will be fairly similar for most and when calculating one, you can use that as a starting point for the other. But like Eyelight said, if you pan, it's different.

When I say 100% crop, it's when you look at a shot that is displayed at the same size as the pixels it contains. If you'd crop something to let's say 1200*800 pixels and save it as that size, you require a perfect pixel freeze. But if you crop to 2400*1600 and save that as a 1200*800 you can get away with a blurred pixel. Then it may travel 2 pixels during that same shutter and you'd likely not notice in the finished shot.

Remember: this is not exact science since we guess the time and distance so you have to find out what works best for you.

This is what i was thinking - where i shoot the aircraft are on speed governed airways so i can set aircraft that travel across my view on that, with my cropping - i tend to use the 1.3 crop mode in camera and then in post i crop the pics about another 50% this gives me the detail i want but still leaves me with a workable file, i then save and view on my 1300mp screen on my laptop.

P.

 

[J-see]

This is what i was thinking - where i shoot the aircraft are on speed governed airways so i can set aircraft that travel across my view on that, with my cropping - i tend to use the 1.3 crop mode in camera and then in post i crop the pics about another 50% this gives me the detail i want but still leaves me with a workable file, i then save and view on my 1300mp screen on my laptop.

P.

I'm not familiar with the D7100 crop mode so you'll have to check the exact pixels it uses.

Here's an example of pixel blur and scaling down. The shot I used was about twice the size and I saved it as half its size. The one I saved normal, the other I first applied a "gaussian blur" to simulate one pixel blur and then saved it to half its size.





If cropped to double the size as what you save the final shot, you can get away with half the shutter required for pixel freeze. There will be a difference but without anything to compare it to, you'll hardly notice.

 

[Paganman2]

It would be interesting to see what other owners can do with aircraft pictures like i shoot, its a challenge to me and the object of my game is to read off the registration serial written under the wings on most of these commercial planes, i have done it a few times with my gear this is mt bench mark and what i am striving to achieve with all my aircraft shots at the same distance, its no different under the right "Seeing" conditions to reading a car plate from a mile away with a 300mm lens + cropping of course.

P.

 

[Paganman2]

I'm not familiar with the D7100 crop mode so you'll have to check the exact pixels it uses.

Here's an example of pixel blur and scaling down. The shot I used was about twice the size and I saved it as half its size. The one I saved normal, the other I first applied a "gaussian blur" to simulate one pixel blur and then saved it to half its size.

View attachment 142237

View attachment 142238

If cropped to double the size as what you save the final shot, you can get away with half the shutter required for pixel freeze. There will be a difference but without anything to compare it to, you'll hardly notice.

Those are both nice but i can see the extra detail in the top one it seems to have more contrast as well, thank you so much for helping me with my new acquisition at times i felt like i have a lemon and my D90 would have done better.

P.

 

[J-see]

Actually, the top one is the one with blur. If you look at the name, you see it has a 2 added. It also is smaller in size.

Quality can be subjective.



Btw, some more on focusing on a plane moving towards you. If you focus with tracking lock, it is entirely possible the focus does not move forward when the plane does since that's what this lock-on does; it ignores distance differences for a specific duration.

It's to prevent focus from jumping to bushes or trees a target might suddenly and momentarily fly behind. In your case, the plane moves forward but to the system, this can as easily be interpreted a such a sudden distance change and thus be ignored in regards to focus adjustments. Especially when you focus on the nose of a plane which has quite a length difference from the middle to the end.

Shots from below it should not be a problem. If your D7100 has settings to change the duration, toy with it for these shots and set it shorter if you have focus issues in such a case, or even disable tracking lock if needed.

 

[Paganman2]

Actually, the top one is the one with blur. If you look at the name, you see it has a 2 added. It also is smaller in size.

Quality can be subjective.








































































Btw, some more on focusing on a plane moving towards you. If you focus with tracking lock, it is entirely possible the focus does not move forward when the plane does since that's what this lock-on does; it ignores distance differences for a specific duration.

It's to prevent focus from jumping to bushes or trees a target might suddenly and momentarily fly behind. In your case, the plane moves forward but to the system, this can as easily be interpreted a such a sudden distance change and thus be ignored in regards to focus adjustments. Especially when you focus on the nose of a plane which has quite a length difference from the middle to the end.

Shots from below it should not be a problem. If your D7100 has settings to change the duration, toy with it for these shots and set it shorter if you have focus issues in such a case, or even disable tracking lock if needed.

Great advice and info - Thank you, i have been trying it out today and i disabled the Lock on feature, the AF-C mode is still being a bit twitchy and not wanting to achieve focus confirmation, so i switched to AF-A but still with the Lock on feature off, and it does seem more stable as i was getting the confirmation light on and nailing focus from what i could see on the rear screen in playback with the zoom.
Perhaps it was the Lock on that was interfering all along with the focus tracking - predictive etc.

P.

 

[J-see]

I encounter this problem often when shooting birds. I shoot a lot at the canal here and at the other side are trees. If for some reason I mess up and the focus locks on the trees in the distance, I can keep the focal point on the bird as as perfect as possible, it won't focus until the duration of the lock I set has passed.

One thing I learned by now is not to push the AF button until I am certain I have my subject where I want it. I usually track it with my lens not concerning myself with it being out of focus and only when it's in the range I want, I will activate auto-focus. That way I avoid an incorrect lock or my system having to catch up with the bird.

It will work the same for airplanes I think. With auto-focus, too early is never that beneficial.

 

[Paganman2]

I encounter this problem often when shooting birds. I shoot a lot at the canal here and at the other side are trees. If for some reason I mess up and the focus locks on the trees in the distance, I can keep the focal point on the bird as as perfect as possible, it won't focus until the duration of the lock I set has passed.

One thing I learned by now is not to push the AF button until I am certain I have my subject where I want it. I usually track it with my lens not concerning myself with it being out of focus and only when it's in the range I want, I will activate auto-focus. That way I avoid an incorrect lock or my system having to catch up with the bird.

It will work the same for airplanes I think. With auto-focus, too early is never that beneficial.

Do you have any opinions about the differences between AF-A and AF-C? as i do seem to have more success with AF-A it just seems to settle the focus down a bit, i understand it is a balance between AF-S and AF-C and so some may not like it due to its auto feeling and letting the camera have more control over what it thinks is static or moving, but i do seem to have more in focus shots with it than AF-C.

P.

 

[J-see]

Do you have any opinions about the differences between AF-A and AF-C? as i do seem to have more success with AF-A it just seems to settle the focus down a bit, i understand it is a balance between AF-S and AF-C and so some may not like it due to its auto feeling and letting the camera have more control over what it thinks is static or moving, but i do seem to have more in focus shots with it than AF-C.

P.

I never use AF-A so I have no idea how well it works. I shoot back-button focus. If I activate, I'm in AF-C mode with either group or d9 and when I release, my focus remains as is. If my target is stationary, I usually switch the AF-C to single point mode, lock focus and release my button.

I'm personally not fan of this sort of auto-modes. There are things the cam is better at than us but decision-making isn't one of those.

 

[Paganman2]

Here is one with AF-A