With all of the advancements in face detection in recent years, is there a way to also have the camera calculate the right aperture to ensure all detected faces will be in focus? For example, if there is a person close by, and also a person in the background, the camera might select F16 in order to get both subjects in focus. Combined with auto-ISO, it could be a reasonable shutter speed, too.
Canon had DEP and A-DEP modes for years in the film and DSLR eras so it is possible, but the camera would probably have to actually focus on each relevant face separately first.
I'm not a fan of the diffraction at f/16, especially on the standard sensors with <4.5µm pixel pitch. Some may find it acceptable with 6µm pixels or with deconvolution.
There are many things camera should be automating by now, including aperture calculation. Another is a precise shutter-speed calculation to avoid motion blur, based on measured motion in the scene.
Strong disagree. None of these engineers looking for solutions to problems that are 101 level photo lessons are worth the time, energy and resources necessary. It takes a few seconds to figure out motion blur and then you just stash a rough idea going forward and you’re set. Similar with DOF. These are base level skills you should acquire as a photographer.
There is a definite subset of gear heads that struggle mightily with the actual photography aspect of their cameras, so I can see the appeal for a certain group.
jlafferty wrote:
Strong disagree. None of these engineers looking for solutions to problems that are 101 level photo lessons are worth the time, energy and resources necessary. It takes a few seconds to figure out motion blur and then you just stash a rough idea going forward and you’re set. Similar with DOF. These are base level skills you should acquire as a photographer.
What about subjects that don't move at a constant speed?
jlafferty wrote:
Strong disagree. None of these engineers looking for solutions to problems that are 101 level photo lessons are worth the time, energy and resources necessary. It takes a few seconds to figure out motion blur and then you just stash a rough idea going forward and you’re set. Similar with DOF. These are base level skills you should acquire as a photographer.
In the 70s people talked about autoexposure and in the 80s autofocus like they were cheating. New technologies and new ways of thinking should be valued to move photography forward. Some species are erratic and it would be nice to have new tools to increase the keeper rate. I can have all the DOF tables, Airy disk sizes, pixel pitches, etc. in my head, but I cannot see every pixel across the entire image and apply a focus offset in hundredths of a second.
I am not opposed to technological progress when it makes sense to me.
However, as described by the OP, I would have absolutely no need for it here.
On the one hand, I have over 45 years of experience in photography and therefore a good feel for DOF at certain focal lengths. On the other hand, there is usually no leeway in terms of aperture, especially when capturing fast movement.
When very fast shutter speeds are required, the aperture range is very limited anyway.
So when we're talking about erratic species and the need to react in a matter of hundredths of a second, my equipment usually offers no leeway at all, and the usable aperture range is severely restricted so as not to send the ISO values into oblivion at fast shutter speeds.
Photography is first and foremost a craft that requires conscious work.
Where technology complements and facilitates the work, I gratefully accept it, but I don't need a nanny for photographic basics.
Snapsy's suggestion, “a precise shutter speed calculation to avoid motion blur, based on measured motion in the scene,” would be another matter.
Experience and practice help here too, but if something like this could really be determined precisely “on the fly,” it would be a nice feature that helps you not to waste any leeway with the shutter speed.
Simplification and automation would also be useful for Nikon's implementation of the focus shift feature.
While it works this way, because using it also builds up experience, it would of course be much more convenient, similar to the focus limiter, to define the desired range by setting a near and far point, and the camera would automatically determine the necessary and most effective steps/step size for the respective focus distance.
Especially if you use focus shift frequently and not always at very similar focus distances, this could save you the trouble of a test run.
I could also imagine a camera-side implementation for users who like to work with hyperfocal distance in certain situations.
While you can also use mnemonics here, like the double the distance method, precise information from the camera would be nice to have for some users.
From this perspective, I am glad that I learned photography in the 1970s, with our own little darkroom, as this means that technology helps me more than if I had left learning the basics of the photographic craft to the camera from the outset.
What does it matter? Set for their fastest speed and you’re done. Oh what, are you going to pick up 2/3rds a stop in exposure latitude 1 out of 5 frames otherwise? Are your photos going to be that dramatically better? This is a solution in search of a problem.
snapsy wrote:
What about subjects that don't move at a constant speed?
Yeah and if you keep going back you could find people saying photographers were just talentless painters. Please don’t slippery slope my position.
I’m literally saying the potential payoff - assuming this stuff would work reliably (TTL still can’t make a great photo decades in) - is so slim that current workflow is good enough.
I’d much rather dev time be directed toward raw histogram and maybe moving us to base log curves.
EB-1 wrote:
In the 70s people talked about autoexposure and in the 80s autofocus like they were cheating. New technologies and new ways of thinking should be valued to move photography forward. Some species are erratic and it would be nice to have new tools to increase the keeper rate. I can have all the DOF tables, Airy disk sizes, pixel pitches, etc. in my head, but I cannot see every pixel across the entire image and apply a focus offset in hundredths of a second.
With all of the advancements in face detection in recent years, is there a way to also have the camera calculate the right aperture to ensure all detected faces will be in focus? For example, if there is a person close by, and also a person in the background, the camera might select F16 in order to get both subjects in focus. Combined with auto-ISO, it could be a reasonable shutter speed, too.
I wouldn't be surprised if it could be done, so even less for the photographer's skill (or lack of it) to affect the photo.
What I object to is paying for it. I don't want any of that automation and I set up my Fujis to work just like my old FM2 did, except for auto iso ( iso is hardly relevant anymore if you don't have to choose it when 'loading' the camera).
But when I bought the FM2 I paid less for it than for an FE or FA, let alone an AF camera. I don't have that choice now, I have to pay for all the software engineer's timedeveloping these gimmicks.
Vento wrote:
I am not opposed to technological progress when it makes sense to me.
However, as described by the OP, I would have absolutely no need for it here.
On the one hand, I have over 45 years of experience in photography and therefore a good feel for DOF at certain focal lengths. On the other hand, there is usually no leeway in terms of aperture, especially when capturing fast movement.
When very fast shutter speeds are required, the aperture range is very limited anyway.
So when we're talking about erratic species and the need to react in a matter of hundredths of a second, my equipment usually offers no leeway at all, and the usable aperture range is severely restricted so as not to send the ISO values into oblivion at fast shutter speeds. ...Show more →
You nailed it Vento. Many presume automation is about accommodating lesser photographers but the opposite is true - it's about giving experienced shooters tools they can use to focus their attention on the artistic elements of the shot and worry less about the rote technical details. My automated shutter speed idea is exactly along those lines - it would continuously adjust the shutter speed to achieve the minimum speed required for sharp images (at a user-configurable threshold). This would provide a significant boost to image quality by maximizing exposure / minimizing noise, which can be the difference of acquiring images with useable detail vs not.
Just shoot everything at f/8 or f/11 and you’ll get all the faces in focus. If you want something that does all the thinking for you, use an iphone or android. They are smaller and do a great job of automatically taking good photos.
jlafferty wrote:
What does it matter? Set for their fastest speed and you’re done. Oh what, are you going to pick up 2/3rds a stop in exposure latitude 1 out of 5 frames otherwise? Are your photos going to be that dramatically better? This is a solution in search of a problem.
The rate of change in motion in some subjects is far greater than 2/3 stop. Consider BIF - even a bird flying at a constant speed goes through a flapping cycle with drastically-different amounts of motion depending on when you catch it in the cycle. And yes, the photos can be dramatically better, the difference between usable feather detail vs not, esp in challenging light conditions.
gyoung143 wrote:
I wouldn't be surprised if it could be done, so even less for the photographer's skill (or lack of it) to affect the photo.
What I object to is paying for it. I don't want any of that automation and I set up my Fujis to work just like my old FM2 did, except for auto iso ( iso is hardly relevant anymore if you don't have to choose it when 'loading' the camera).
But when I bought the FM2 I paid less for it than for an FE or FA, let alone an AF camera. I don't have that choice now, I have to pay for all the software engineer's timedeveloping these gimmicks.
Many think extra features like video and what's being proposed in this thread result in more expensive cameras. The opposite is true - they make your cameras less expensive, not more. Without these features the cameras become more niche products = fewer unit sales = greater amortized R&D / manufacturing setup cost = more expensive cameras.
I had considered BIF. Again, figure out the fastest their wings can flap, set shutter to that as your minimum. Build your other settings around a fixed minimum shutter. Done. I don’t shoot BIF but - bird species depending - I would start at 1/2500 or 1/3200 shutter and see what happens, adjust as needed. If it was Swifts, Barn Swallows or Hummingbirds I’d go straight to 1/8000.
If you’re proposing a setting that changes aperture and/or ISO to accommodate shutter based on metered readings, in order to generate relatively consistent exposures in shifting lighting conditions, well… I’ve got news for you…
snapsy wrote:
The rate of change in motion in some subjects is far greater than 2/3 stop. Consider BIF - even a bird flying at a constant speed goes through a flapping cycle with drastically-different amounts of motion depending on when you catch it in the cycle. And yes, the photos can be dramatically better, the difference between usable feather detail vs not, esp in challenging light conditions.
I’ll go one further, this is actually my strategy when photographing subjects moving through vastly different regions of exposure (bright highlights versus shade):
Figure out my minimum shutter. For my subjects that’s typically 1/1000, 1/2500 tops. Then, figure out what aperture I want to work at, given distance, lens, subject size, etc. let’s say that’s f/5.6 minimum. Set these and don’t touch them going forward.
Now, take a test frame with those settings fixed, in the shaded area to find your ISO. You do shade first because it’s least forgiving and you want your camera using that as a base. Then move to your bright area, and change only ISO to accommodate. The swing could be huge but it’s generally just 3-4 clicks of a dial. You don’t need a subject to do this, you can judge based on inanimate objects of relatively similar tonal value.
And that’s it. Spend the rest of your day just turning that one dial, counting to yourself “1, 2, 3, 4” each time you turn the dial. You could even set up custom User banks or whatever so it’s a single button.
Honestly I do the above but build around shutter… I’ll typically set ISO based on the shaded area at my 1/1000, f/5.6, and then pivot to the bright area and adjust shutter, let’s say it jumps to 1/5000. And that’s it, just remember those two numbers on the shutter dial. Then a cloud moves in and I panic, chuckle, and work it out anyway
Or… if you prefer, you can go shutter priority and auto-ISO with highlight protect.
