If you like DSLR/OVF for fast actions, but also like EVF for other situations

Nikon 1? They are still trying to figure out how to do a good mirrorless today
I sold mine I have a Olympud E-M1II alongside my Nikons now for the reach, the evf delay is less but you still need to be aware that it is there when shooting.
 
Panning with mirrorless.

F

Yes, but a Nikon 1 V2 is not the cutting edge or mirrorless technology.
It's a example to show what happens due to evf delay, many deny that it can happen at all, they probably don't shoot sports or stand on street corners shooting with wide angle lenses, Some shoot BIF or sports and resort to add on red dot viewfinders because of the delay I have a newer mirrorless and still need to delete photo's due to the evf delay.

--
Mike.
"I say we take off and nuke the entire site from orbit, it's the only way to be sure."
 
Panning with mirrorless.

F

Yes, but a Nikon 1 V2 is not the cutting edge or mirrorless technology.
It's a example to show what happens due to evf delay, many deny that it can happen at all, they probably don't shoot sports or stand on street corners shooting with wide angle lenses, Some shoot BIF or sports and resort to add on red dot viewfinders because of the delay I have a newer mirrorless and still need to delete photo's due to the evf delay.


It doesn't mean all mirrorless cameras are as bad as each other. I have a Fuji X-T100 that is no good for following action, but my X-T2 with the battery grip and boost mode enabled is a deferent kettle of fish.



--
 
And that may well be true.

My contention is simple, and supported by control theory. Whatever you can do with lag in the system you can do better without the lag. Or, you can do more without lag than you can with lag.
True, but not all closed loop control systems with lag are unstable. Humans tracking manually with a VF also contribute lag in the system because of natural reaction time. A slow human using a OVF with zero lag can also cause an unstable system as we well know.
You are still mixing apples & oranges ...

Just like EVF-lag and mirror/shutter-lag are different problems that require different mental compensation, SO IS "REACTION" lag.
I am just talking about apples. I am talking about what constitute the closed loop control just to track and keep the object in frame, not to take the picture. Taking the picture is another operation which also involves the human reaction lag

as well as the shutter (and autofocus if applicable) lag. It is another loop which is an open loop. There are two different cases where human reaction comes into play.
Human Reaction Lag is mainly a problem w/ unpredictable/erratic subjects.
It is the combined EVF lag plus the human reaction lag and any other system lag that determine if the system is stable or unstable.
True ... but that does not change the fact that each separate/different lag's require different mental-compensation.
OK
I can understand Lee's point that OVF tracking can be easier, (if subject is PREDICTABLE and NOT ERRATIC). With "0" OVF lag, you indeed can "track" accurately and the subject can be STILL CENTERED after mirror/shutter-lag and AFTER the mirror returns down.
He talks in generalities like easy or hard. I'm talking about being a stable or unstable closed loop control systems. Only when the lag exceed a critical point does the closed loop system mathematically becomes unstable.
I suggest that ANY lag can be problematic to some degree.
With EVF-lag, the subject is ALWAYS AHEAD even before the "first" exposure. So that is already a problem because you are inherently always playing catch-up.
It's how much ahead and that depends on the EVF lag and the speed of the object. Obviously if the EVF lag is small and the object speed is slow there is no problem. There is a critical point for losing track and that is when the system becomes unstable.
Again ... I suggest that ANY lag can be problematic to some degree.
BUT ... the problem is infinitely worse AFTER the "first" exposure because the returning "live" image is (double) WAY BEHIND the original subject movement, (even IF YOUR PANNING WAS PERFECT). So the biggest problem is with that "second" (and each sequential) shot(s) of a sequence.
Each shot is independent.
BUT ... you CANNOT SHOOT AGAIN unless you "catch-up" after each sequential shot ... and the FACT is that whatever original EVF-lag there is, is then (temporarily) DOUBLED with the "returning" image after each shot.

So the original "catching-up" is complicated with the returning image after each shot.

If you cannot understand that .. you are not thinking with an OPEN mind.
Explain what situation you are talking about.
With "0" lag OVF, if you are "panning" correctly and your subject is centered when you push shutter release, there is indeed mirror/shutter-lag but the subject remains centered until after the mirror drops back down, and you can thus immediately reshoot again because the subject NEVER because un-centered, (again assuming you are correctly panning).

With LAG ... your (albeit-centered) image is ALWAYS behind the subject, and your camera may or may-NOT be actually pointed directly at subject, But your (delayed)recorded image may be still OK if it was "centered" at moment it was RECORDED. But I repeat the camera may NOT have been actually pointed at the subject any more -- maybe CLEAR SKY, (albeit the subject WAS "centered" moments before and thus a "centered" delayed image recorded).

So accepting that the camera may actually (because of the lag) have been pointing BEHIND the subject (ala clear-sky) before the shot ... even if correctly panning, AFTER the shot the camera WILL STILL be looking at CLEAR SKY.

So now a JUGGLING event much occur because you jerk the camera "forward", (to catch-up), but you STILL DON'T AGAIN SEE THE SUBJECT because of L-A-G and when it does reappear, you may now be AHEAD of the subject and have to readjust again, (thus "juggling").

So you CANNOT take ANOTHER shot until you (centered) reacquire the subject again, (but again with a DELAYED image).

This is what is referred to as "closed-loop control".
I am talking about the tracking of a single moving object to keep it framed in the viewfinder. PERIOD.
But HOW can you do that when the displayed image is DELAYED .... ???

Seeing it "centered" only means it WAS "centered" before the lag/delay. You may however NOW be pointed either before or after it.
When you have it satisfactorily framed you take a picture and start over to take another independent picture.
With a FAST-MOVING subject ... you often want to take a CONTINUOUS-SEQUENCE.

That is problematic if your subject (image) is no longer centered after the first shot.

By the time you re-acquire/center the subject again, it too often is TOO LATE.
But Lee has a bigger problem when he relies on the (slower) LCD on a (older) LV-LCD that is not optimized for minimum lag. Beginners should be aware that newer EVF are faster, (in both lag and frame-rate).
Again, only if the overall lag (EVF lag and human reaction lag) exceed other system parameters does the system become unstable, ie, you lose track of the object. So, yes, with a small EVF lag and agile person with quick reaction successful tracking of fairly fast moving objects is achievable as some have shown here.
I have agreed that LEE is over-exaggerating the problem (w/ older/slower LCD) ... but his POINT is correct.
Lee speaks from his limited experience and is hard headed when someone else show they have different experiences.
No ... he speaks from an older/slower LV on LCD, (vs a faster, more modern, EVF).

So an admitted problem is exasperated, and he refuses to admit newer EVF's have "less" of a problem, (but still a problem to some lesser extent).
As I said, these were no where near my most challenging tracking situations. Try tracking, by hand, a 9m object, in orbit, from 400km away, at night, at 4,300mm equivalent through f/21 optics while it files straight over your head and you have to manually focus.
Wow, I wonder what's harder: that or juggling 5 balls while standing on your head and hopping on one foot for a cirque du soleil audition.
Well, I can juggle 5 balls (juggling is one of my 7 hobbies), but not upside down. I'll leave the tough stuff to Anthony Gatto.
Why not automate the tracking?
It's much harder than it sounds. The only person I know who did it wrote all his own software, and still uses a gaming joystick to make real-time corrections.
 
Last edited:
Panning with mirrorless.

F

Yes, but a Nikon 1 V2 is not the cutting edge or mirrorless technology.
It's a example to show what happens due to evf delay, many deny that it can happen at all, they probably don't shoot sports or stand on street corners shooting with wide angle lenses, Some shoot BIF or sports and resort to add on red dot viewfinders because of the delay I have a newer mirrorless and still need to delete photo's due to the evf delay.
It doesn't mean all mirrorless cameras are as bad as each other. I have a Fuji X-T100 that is no good for following action, but my X-T2 with the battery grip and boost mode enabled is a deferent kettle of fish.
I never said they were But the delay exists. Look at the sample gallery for the New 90D https://www.dpreview.com/reviews/canon-eos-90d-initial-review-what-s-new-and-how-it-compares/4 look at the panned photo's of the racing cars framed with more road in front than behind. Now look at the sample galleries here https://www.dpreview.com/reviews/canon-eos-90d-initial-review-what-s-new-and-how-it-compares/4 You will see two panned shots with the M6II where the car has less road in front than behind this, is as in my shot's due to the EVF delay. Someone not used to shooting motorsport may not see this affect or deny it exists.

--
Mike.
"I say we take off and nuke the entire site from orbit, it's the only way to be sure."
 
And that may well be true.

My contention is simple, and supported by control theory. Whatever you can do with lag in the system you can do better without the lag. Or, you can do more without lag than you can with lag.
True, but not all closed loop control systems with lag are unstable. Humans tracking manually with a VF also contribute lag in the system because of natural reaction time. A slow human using a OVF with zero lag can also cause an unstable system as we well know.
You are still mixing apples & oranges ...

Just like EVF-lag and mirror/shutter-lag are different problems that require different mental compensation, SO IS "REACTION" lag.
I am just talking about apples. I am talking about what constitute the closed loop control just to track and keep the object in frame, not to take the picture. Taking the picture is another operation which also involves the human reaction lag

as well as the shutter (and autofocus if applicable) lag. It is another loop which is an open loop. There are two different cases where human reaction comes into play.
Human Reaction Lag is mainly a problem w/ unpredictable/erratic subjects.
It is the combined EVF lag plus the human reaction lag and any other system lag that determine if the system is stable or unstable.
True ... but that does not change the fact that each separate/different lag's require different mental-compensation.
OK
I can understand Lee's point that OVF tracking can be easier, (if subject is PREDICTABLE and NOT ERRATIC). With "0" OVF lag, you indeed can "track" accurately and the subject can be STILL CENTERED after mirror/shutter-lag and AFTER the mirror returns down.
He talks in generalities like easy or hard. I'm talking about being a stable or unstable closed loop control systems. Only when the lag exceed a critical point does the closed loop system mathematically becomes unstable.
I suggest that ANY lag can be problematic to some degree.
With EVF-lag, the subject is ALWAYS AHEAD even before the "first" exposure. So that is already a problem because you are inherently always playing catch-up.
It's how much ahead and that depends on the EVF lag and the speed of the object. Obviously if the EVF lag is small and the object speed is slow there is no problem. There is a critical point for losing track and that is when the system becomes unstable.
Again ... I suggest that ANY lag can be problematic to some degree.
BUT ... the problem is infinitely worse AFTER the "first" exposure because the returning "live" image is (double) WAY BEHIND the original subject movement, (even IF YOUR PANNING WAS PERFECT). So the biggest problem is with that "second" (and each sequential) shot(s) of a sequence.
Each shot is independent.
BUT ... you CANNOT SHOOT AGAIN unless you "catch-up" after each sequential shot ... and the FACT is that whatever original EVF-lag there is, is then (temporarily) DOUBLED with the "returning" image after each shot.

So the original "catching-up" is complicated with the returning image after each shot.

If you cannot understand that .. you are not thinking with an OPEN mind.
Explain what situation you are talking about.
With "0" lag OVF, if you are "panning" correctly and your subject is centered when you push shutter release, there is indeed mirror/shutter-lag but the subject remains centered until after the mirror drops back down, and you can thus immediately reshoot again because the subject NEVER because un-centered, (again assuming you are correctly panning).

With LAG ... your (albeit-centered) image is ALWAYS behind the subject, and your camera may or may-NOT be actually pointed directly at subject, But your (delayed)recorded image may be still OK if it was "centered" at moment it was RECORDED. But I repeat the camera may NOT have been actually pointed at the subject any more -- maybe CLEAR SKY, (albeit the subject WAS "centered" moments before and thus a "centered" delayed image recorded).

So accepting that the camera may actually (because of the lag) have been pointing BEHIND the subject (ala clear-sky) before the shot ... even if correctly panning, AFTER the shot the camera WILL STILL be looking at CLEAR SKY.

So now a JUGGLING event much occur because you jerk the camera "forward", (to catch-up), but you STILL DON'T AGAIN SEE THE SUBJECT because of L-A-G and when it does reappear, you may now be AHEAD of the subject and have to readjust again, (thus "juggling").

So you CANNOT take ANOTHER shot until you (centered) reacquire the subject again, (but again with a DELAYED image).

This is what is referred to as "closed-loop control".
This is what I have found with motorsport shots. and seen here
watch from 37min 14 sec. This is where he tries to keep the car centred in his panning.
I am talking about the tracking of a single moving object to keep it framed in the viewfinder. PERIOD.
But HOW can you do that when the displayed image is DELAYED .... ???

Seeing it "centered" only means it WAS "centered" before the lag/delay. You may however NOW be pointed either before or after it.
When you have it satisfactorily framed you take a picture and start over to take another independent picture.
With a FAST-MOVING subject ... you often want to take a CONTINUOUS-SEQUENCE.

That is problematic if your subject (image) is no longer centered after the first shot.

By the time you re-acquire/center the subject again, it too often is TOO LATE.
But Lee has a bigger problem when he relies on the (slower) LCD on a (older) LV-LCD that is not optimized for minimum lag. Beginners should be aware that newer EVF are faster, (in both lag and frame-rate).
Again, only if the overall lag (EVF lag and human reaction lag) exceed other system parameters does the system become unstable, ie, you lose track of the object. So, yes, with a small EVF lag and agile person with quick reaction successful tracking of fairly fast moving objects is achievable as some have shown here.
I have agreed that LEE is over-exaggerating the problem (w/ older/slower LCD) ... but his POINT is correct.
Lee speaks from his limited experience and is hard headed when someone else show they have different experiences.
No ... he speaks from an older/slower LV on LCD, (vs a faster, more modern, EVF).
As I said, these were no where near my most challenging tracking situations. Try tracking, by hand, a 9m object, in orbit, from 400km away, at night, at 4,300mm equivalent through f/21 optics while it files straight over your head and you have to manually focus.
Wow, I wonder what's harder: that or juggling 5 balls while standing on your head and hopping on one foot for a cirque du soleil audition.
Well, I can juggle 5 balls (juggling is one of my 7 hobbies), but not upside down. I'll leave the tough stuff to Anthony Gatto.
Why not automate the tracking?
It's much harder than it sounds. The only person I know who did it wrote all his own software, and still uses a gaming joystick to make real-time corrections.
 
This is certainly a good example of panning. Is it a good example of tracking?

please define the difference between panning and tracking.
Tracking is when the subject is moving around within the frame and the camera is able to lock on to the subject. The ability to track is important for shooting sports because the subject tends to change direction and there are usually other subjects in the frame you want to ignore, and for BIF because the subject often changes direction. For a tightly framed subject while panning tracking is irrelevant because it fills the frame. Mirrorless cameras are better at tracking because the on sensor PDAF points fill the frame.

For photographing a super high speed airplane tightly framed and moving at a steady speed and direction it's conceivable that a camera with an OVF might be a bit better but I think it's a moot point because with today's high res sensors it's easy to zoom out a little and crop in post. I have used both DSLRs and cameras with EVFs and I always zoom out a bit with both because I'm not skilled enough to keep a tightly cropped plane perfectly in the frame while panning.
 
Let's acknowledge a few things on both sides:
  1. It's 100% possible to track fast action with an EVF. The hitrate advantage with an OVF would be small
  2. Other factors like lens AF speed, knowledge of the subject/sport (so you know where to point the camera in the first place), and experience (panning takes a lot of practice) mean that we don't have a controlled comparison.
  3. It's not possible to make an EVF that gives a zero latency preview. There are certain unavoidable steps in generating a preview in the EVF. See https://www.dpreview.com/forums/post/63044645 for an estimate on what the latency might be like.
My personal conclusions:
  1. Factors mentioned in #2 above are almost certainly larger factors than OVF vs EVF. In general gear makes very little difference compared to the photographer, and a determined photographer will find a way to get good shots with pretty bad gear.
  2. If the latest generation of EVFs can bring lag down to 5-18 ms, we're getting close to the difference between 60 Hz and 240 Hz monitors. That's a very small difference that you probably won't notice unless you do a controlled test and see that with an OVF, the reaction-to-shot time is slightly better on average.
  3. Two shots taken 5-18 ms apart may both be equally usable. That's almost certainly the case with panning, and true much of the time with basketball (multiple shots from a 9 fps burst can be submitted with the same caption).
  4. From above, I'd be fine with using an EVF for action. It wouldn't be my first choice, but going from OVF to EVF alone certainly will not tank my keeper rate.
Now back to the original thread title: "If you like DSLR/OVF...but also like EVF" - can we focus on that instead of going back to the age old OVF vs EVF zero-sum argument? I don't think anyone loses anything with a hybrid solution.

Besides action, each finder also has advantages and disadvantages:
  • OVF gives much better battery life
  • EVF makes it more obvious when your exposure is totally wrong
  • OVF lets you use your eye's dynamic range to see the scene
  • EVF shows the correct DOF and can magnify to 1:1 for accurate MF w/fast lenses
  • and we can go on for days about why one of them is our personal preference for one situation or another.
Why not have both?
 
It won't take much for Canon to lock the mirror up for when it goes into evf mode.

The 90d has the same sensor as the M6ii. Performance should be the same using the LCD screen for both cameras.
There are many more differences between dSLR-LV and native ML camera.

Unless "TESTS" prove me wrong, (on this specific camera), there has not been any other LV that equals a ML shutter-lag. (even though Canon has had hybrid-PD sensors for some time and their lag is indeed shorter than Nikon, but not as fast as any other ML)
Imaging-resource has a lot of timing tests.

https://www.imaging-resource.com/PRODS/canon-eos-m6/canon-eos-m6A6.HTM - Canon M6 - unfortunately they don't have the M6 ii reviewed

https://www.imaging-resource.com/PRODS/canon-80d/canon-80dA6.HTM - Canon 80D. Unfortunately they don't have the 90D reviewed either.

Canon M6: 0.061s prefocused

Canon 80D: 0.057s prefocused, 0.060s prefocused in LV

So M6 vs 80D in LV are practically the same.
 

Keyboard shortcuts

Back
Top