Pages: [1] 2 3   Go Down

Author Topic: What COULD a camera be in 2018?  (Read 6294 times)

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
What COULD a camera be in 2018?
« on: July 31, 2018, 12:05:24 pm »

Hi All,

We've moved from a camera being a light-tight box with some light sensitive gloop at the back to being a super-computer equipped with massive processing power and a whole heap of additional sensors (eg gyroscopes, GPS).

But in many ways we're taking little advantage of all that tech in terms of what we expect the camera to DO.

Sure, the ability to treat the supercomputer-with-sensors like it's a film camera from the 1950's is good in some ways. One definitely wants the ergonomics of picture taking to take a back seat a lot of the time.

But there are also a lot of things the camera could handle for us which are far from universal, and probably some things which they should do that no-one has implemented yet.

A few examples which I think should be there as options for all "pro" level cameras. Many cameras do some of these. As far as I know, none of them do all:

- The "don't fire the shutter until the vibrations have died down" triggering mode (I think someone said PhaseOne? have this).

- Proper live view and complete camera control from a remote screen, ideally wirelessly, but failing that via USB-C. The wifi implementations I've used on pro cameras are a bit naff compared with what a DJI Mavic Pro provides, for example.

- Intervalometer functionality properly integrated. One shouldn't need apps or dangling bits of wire to do a timelapse!

- Focus stacking. Maybe not assembling the actual composite, I don't know how difficult that is, but reliably acquiring the necessary images in camera with minimum fuss. I should be able to touch screen to show the two extreme focus points, say "go" and have the camera fire off a focus series in the same way as auto bracket lets you take an EV series.

- ETTR exposure mode (using a pre-shot to judge exposure if necessary). "Get the maximum exposure you can get whilst allowing no more than x% of pixels in any colour channel to clip" - and mark the metadata to say how that corresponds to the camera's best guess for "correct" exposure so when you open the image in post a suitable guess at the dial-back of exposure has been put in for you already.

- Pentax's night sky tracking using IBIS and GPS. If you're at all interested in astrophotography it's almost enough to tempt one to buy a Pentax just for this feature.

- In camera panorama stitching.

- In camera HDR (at least for simple cases- with the option to save the stack as well).

- Pixel shift (I don't use it myself, because I shoot people primarily. But since every camera should have IBIS, there's no reason not to have it).

- Uploadable picture profiles. This is coming in in the cine world (where they're called LUTs). The use case for stills would be to get out of camera JPEGs with your fave wacky post processing corrections and colour curves applied in camera, rather than having to choose from a relatively small range of in camera options via a limited control interface. Fuji film simulation on steroids; load up your camera with your own personal looks library.

- In camera frame stacking, for long exposures without ND filters and with much better signal to noise.

- A third option for in-camera output. RAW should be exactly that- the information from the sensor with calibration metadata but no processing. JPEGs for backwards capability. Then some lossless high-bit-depth file format for composite images and images processed in camera. Why should we be limited to having Fuji-like film simulations and image stacks in 8-bit lossy format? It's not appropriate to export them in RAW- that's not what it is for. But something that lets you use the camera's processing as a starting point and fine tune in post would be very useful. For example, an HDR stack image might well benefit from saving RGB channels separately and having pointers to the original RAW files in the metadata, and would certainly benefit from 16 bits per channel or more- why do we squeeze this down to a lossy 8-bit JPEG?


What else COULD the camera be?

What stuff do you do in post that the camera could facilitate much better than at present?


I'd also like to see some serious investigation into ergonomics and automation.


For example, Sony's auto ISO plus minimum shutter speed preference plus exposure compensation dial plus IBIS have completely changed the way I shoot in available light. I select my working aperture, tell the camera that I want a faster shutter speed one stop faster than 1/F as the minimum (so for a 50mm lens I get a min shutter speed of 1/100th of a second). The camera selects the lowest ISO necessary to get that shot, and the IBIS means that even at 42 megapixels my hit rate for critical sharpness is excellent. And if the exposure seems a bit over-cooked or under-exposed, I can control that with just one physical dial (the exposure compensation).

I agree with the online consensus that broadly, Sony ergonomics are not as nice as CaNikcon. But this one they're got just right, and I find it hard going back to a camera that doesn't work that way now.

What other ergonomic improvements might we reasonably ask for?

I think we've not really cracked the ergonomics for focussing yet. The closest I've come to perfection is a Canon 1Dx Mark II using live view. The functionality of Sony's eye focus is great, but I don't think we've quite hit the most effective way of controlling it yet. Tap-to-focus is the obvious way to go with touch screens, but I personally don't think we've got it right for viewfinder use yet. Maybe a return of Canon's look-to-focus? Joysticks and dpads are OK but still feel one step removed from the process compared with tap-to-focus on the monitor. For me it would be great to set eye focus, then just look at the appropriate person on screen and have the camera use that to drive it's selection of which face (and maybe which eye) to focus on, for example.

Global electronic shutter with super-fast flash sync speeds (without needing leaf shutters) is something else I think would really be a killer feature (I know a lot of people are thinking about how best to do that- global electronic shutter is also very desirable for video).


What do you think? What else could a camera be, now that it's a supercomputer and a bunch of sensors in a miracle box, rather than silver halide?

Cheers, Hywel











Logged

KLaban

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2451
    • Keith Laban Photography
Re: What COULD a camera be in 2018?
« Reply #1 on: July 31, 2018, 12:25:50 pm »

My cameras do little for me which is why I bought them, allowing me to retain control and make decisions.

32BT

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3095
    • Pictures
Re: What COULD a camera be in 2018?
« Reply #2 on: July 31, 2018, 12:38:19 pm »

My cameras do little for me which is why I bought them, allowing me to retain control and make decisions.

Yes, perhaps the problem is exactly that the camera is indeed becoming a supercomputer. On the one hand this is nice, because you can do more in camera and spend less time in post, on the other hand the camera has become a plethora of buttons, most of which get in the way of what you really want it to do.

So, perhaps a more modular design would be appreciated. Why is the lcd and evf fixed to the device? I totally imagine that the evf of the future is more like a wearable "bluetooth" device a la google glasses. Similarly the lcd. Or that functions like picture mode or movie mode are available in a different grip one can attach. Want minimal buttons? Purchase the minimalist grip. Want extra moviebuttons? Purchase the movie grip. Etc.
Logged
Regards,
~ O ~
If you can stomach it: pictures

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
Re: What COULD a camera be in 2018?
« Reply #3 on: July 31, 2018, 12:54:34 pm »

My cameras do little for me which is why I bought them, allowing me to retain control and make decisions.

I'm all for retaining control and making decisions. I'd just like to be able to take the decisions at a higher level of abstraction.

Take the example of shooting an available light portrait using Sony auto-ISO functionality.

I want to control the look of the image and the overall exposure.

The look I control primarily via aperture- so I choose to select that directly. (And by choosing which lens I put on the camera in the first place of course).

The overall exposure- I want to get a good ETTR exposure without blowing any colour channels. Nothing stops me from setting shutter speed and ISO manually if I so desire, but that's not actually what I want to be thinking about shooting sets of photos in changing natural light. I want to get the good ETTR exposure without going so low in shutter speed that I'm liable to incur camera shake, and without turning the ISO up any higher than it needs to go.

I can either do that by twiddling two interdependent camera controls (shutter and ISO), take a test shot, look at the histogram, apply a mental exposure correction, choose whether to dial it in via ISO or shutter speed, and take the shot.

Or I can keep an eye on the histogram, let the camera combine the two interdependent controls into just one (exposure compensation) - under my command, because I can set the parameters for how it chooses minimum shutter speed and maximum ISO. Then I just need to twiddle one dial, not two, to produce the same ballet of interconnected exposure control.

Once I've dialled in the compensation for the scene I rarely need to change it, even as the exposure value changes over several stops as the intensity of the sunlight ebbs and flows. It's significantly quicker and doesn't interfere so much with my shooting "flow".

I would ALWAYS advocate all cameras have a "act like a dumb 1950s film camera" mode.

But shooting a set of 100 pics of a model outdoors in dappled sunlight with the sunlight coming and going in a British summer- I really appreciate the higher level abstraction and automation. I shoot faster, get more correctly exposed shots, and therefore make more good photos (and at the end of the day make more money too).

Likewise, manual focus is all fine and dandy, and mirrorless/live view to zoom in and get precise manual focus shot by shot is great. I can still do that with pretty much all modern systems, although admittedly fly-by-wire is a bit of a compromise.

But what do I actually DO when shooting a set of photos of a model? Maybe 8 out of 10 shots, I want the closer eye to the camera in sharp focus. I can do that manually, focussing by hand shot by shot. Nothing stops me, although relying on both model and photographer to stay stock still when shooting with an 85 mm f/1.4 lens wide open is pretty challenging. (Judging it without live view/EVF and 1:1 zoom is beyond the capabilities of my eyesight and always has been).

I can turn on continuous autofocus with eye focus and get a much greater hit rate, and consequently take more good shots in less time. And therefore make more money from a day's shoot.

I'm still in control. If the shot needs to be focussed on her lips or her shoes or whatever, I can still do that. The camera provides controls to help me do so. I think we might be helped by better ergonomics in telling the camera AF where to focus, but I simply don't buy any argument against AF at this point so long as the "dumb 1950's film camera" manual mode is still there. I'm deciding what to shoot and where to focus. The camera is just providing me with a higher-level abstraction of how to control that than a ring connected to a helicoid physically moving lens groups while I squint through the viewfinder.

I care about getting the right thing in focus. The more the technology can help me do that, the better.


My question is what other possibilities are there? What other "advanced modes" controlling picture taking at higher levels of abstraction might we not really have dreamed up yet?


I totally agree that the ergonomics for all these "advanced modes" need a serious rethink, as opgr says. I'm trying to envisage what such a system might look like, what controls it might have, what might actually be useful to us rather than providing "just another button on the camera".


Cheers, Hywel

« Last Edit: July 31, 2018, 01:01:49 pm by HywelPhillips »
Logged

chez

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2501
Re: What COULD a camera be in 2018?
« Reply #4 on: July 31, 2018, 01:35:12 pm »

Yes, perhaps the problem is exactly that the camera is indeed becoming a supercomputer. On the one hand this is nice, because you can do more in camera and spend less time in post, on the other hand the camera has become a plethora of buttons, most of which get in the way of what you really want it to do.

So, perhaps a more modular design would be appreciated. Why is the lcd and evf fixed to the device? I totally imagine that the evf of the future is more like a wearable "bluetooth" device a la google glasses. Similarly the lcd. Or that functions like picture mode or movie mode are available in a different grip one can attach. Want minimal buttons? Purchase the minimalist grip. Want extra moviebuttons? Purchase the movie grip. Etc.

Get ready to shell out big bucks for modularization. I use my camera for multiple applications where shooting landscape I only want very minimal buttons, but when shooting travel, I like AF and all its options...and then when shooting sports, I want all the speed related features...and I always intermix some video...so I need the video options.

Personally that would be a nightmare...sort of like buying a car back in the 80's where everything was an option.

As far as wearing an EVF like google glasses...we already have idiots walking into traffic as they are texting...I could just see the mess as everyone is wearing their EVF and not paying attention what's around them.

My one want is to have the camera manufactures open up the firmware and provide access to the camera for 3rd party apps...like the iPhone.
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: What COULD a camera be in 2018?
« Reply #5 on: July 31, 2018, 01:47:29 pm »

For what it’s worth, one of the ways I like to use auto-ISO is in M mode which allows me to set both aperture and speed and let the camera adjust the ISO while taking into account whatever exposure compensation I set. This works very well, but exposure remains based on the camera meter.

It would indeed be neat to do this based on an exposure assessment done from raw histogram with adjustable degree of ETTR.

Without underestimating the potential of technology nor the value of discussing it, the more I photograph though, the more I think that planning, awarness, light, vision and timing make 90% of an image and I wouldn’t consider an AI algo to be able to contribute to making my images better.

Cheers,
Bernard
« Last Edit: July 31, 2018, 02:02:38 pm by BernardLanguillier »
Logged

32BT

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3095
    • Pictures
Re: What COULD a camera be in 2018?
« Reply #6 on: July 31, 2018, 01:51:30 pm »

Get ready to shell out big bucks for modularization. I use my camera for multiple applications where shooting landscape I only want very minimal buttons, but when shooting travel, I like AF and all its options...and then when shooting sports, I want all the speed related features...and I always intermix some video...so I need the video options.

Personally that would be a nightmare...sort of like buying a car back in the 80's where everything was an option.

As far as wearing an EVF like google glasses...we already have idiots walking into traffic as they are texting...I could just see the mess as everyone is wearing their EVF and not paying attention what's around them.

My one want is to have the camera manufactures open up the firmware and provide access to the camera for 3rd party apps...like the iPhone.


Then purchase the jack-of-all-trades-master-of-none grip.

The google glasses method has been used by helicopter pilots for ages, vastly different then people with their heads burried in a screen. You don't have to have the glasses engaged when not actually taking a picture.

Didn't Sony have the Sony apps solution that was supposed to become open to thirdparty options at some point, but has been removed from recent offerings? Not sure why though, it certainly makes sense if you consider the camera as a computing platform.
Logged
Regards,
~ O ~
If you can stomach it: pictures

KLaban

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2451
    • Keith Laban Photography
Re: What COULD a camera be in 2018?
« Reply #7 on: July 31, 2018, 02:05:06 pm »

...I would ALWAYS advocate all cameras have a "act like a dumb 1950s film camera" mode...

Thanks but I'll take basic modifiable settings over camera modes, dumb or smart.

I'll admit that perhaps it's just me, but experience has taught me that the more the camera does for me the more my work suffers.
« Last Edit: July 31, 2018, 02:23:05 pm by KLaban »
Logged

petermfiore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2705
    • Peter Fiore Fine Art
Re: What COULD a camera be in 2018?
« Reply #8 on: July 31, 2018, 02:26:05 pm »

Thanks but I'll take basic modifiable settings over camera modes, dumb or smart.

I'll admit that perhaps it's just me, but experience has taught me that the more the camera does for me the more my work suffers.

It's not just you...I agree, basic manual never goes out of style. If you know what your doing.

Peter

KLaban

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2451
    • Keith Laban Photography
Re: What COULD a camera be in 2018?
« Reply #9 on: July 31, 2018, 02:38:37 pm »

Peter, perhaps the cameras of the future will be so perfectly automated that they won't allow us to make a bad image.

Now there's a nightmare.

petermfiore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2705
    • Peter Fiore Fine Art
Re: What COULD a camera be in 2018?
« Reply #10 on: July 31, 2018, 02:41:42 pm »

Peter, perhaps the cameras of the future will be so perfectly automated that they won't allow us to make a bad image.

Now there's a nightmare.

Oh God forbid. I relish the awkward, clumsy and odd in all pictures. Including my work in camera and Canvas. Imagine perfect paintings?

Peter
« Last Edit: July 31, 2018, 02:46:24 pm by petermfiore »
Logged

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
Re: What COULD a camera be in 2018?
« Reply #11 on: July 31, 2018, 02:48:40 pm »

Thanks but I'll take basic modifiable settings over camera modes, dumb or smart.

I'll admit that perhaps it's just me, but experience has taught me that the more complex and automated the camera the more my work suffers.

By "mode", I mean the ability to operate the camera as if it were a 1950's film camera, with the physical controls thereof.

1950s era film cameras were that way for good reason- it is the necessary and minimal set of controls for manual picture taking. 

But we now have other options and opportunities which in the correct circumstances go from nice optional extras to "so useful you'll hardly find a pro who doesn't use it". The poster child for which is autofocus- I doubt you'll find many people on the sidelines at a sports game with manual focus lenses. Eye focus AF is already heading that way for people photographers.

I'm advocating adding options on control dials and in menus, like the "Auto" selection for ISO which facilitates using the Sony auto exposure mode I talked about above. How do we make them unobtrusive when their use is not desired, but easy to control when they are desired?

I feel like there are better ways to control the advanced functionality than a million buttons on the camera. Customisable buttons (with little LED screen labels, ideally) are one way to go. In "1950's film camera mode" these could even be deactivated entirely- don't do anything, don't light up, no confusion.

As an example, control via the rear touch screens is pretty good as direct control interfaces go. It's direct, tactile- you poke your finger at the thing on screen you want to control. There's no easier way to select what you want to be in focus than poking it with a digit and the camera making it so. There are current limitations to do with brightness, glare, limited size and articulation, etc. but there's a reason the touchscreen interface has taken the world by storm. It allows you to present the most relevant control inputs to the current task without all the physical clutter.

One could literally have just the physical controls and nothing else. But can we make it so a different photographer- or the same one taking a different sort of photo with the same hardware- could switch on whatever they need for the job at hand? How do we supplement the interface for when it really helps (eg eye AF) without drowning in complexity?



I think we should consider what abstractions are most useful, photographically, and how best to control them- so we can enhance and extend the functionality of the 1950s film camera interface in the way that the high level abstractions help rather than hinder.


Which is why I think we've still not cracked control of focus when looking through the viewfinder, for example. It's still too many steps to shift the focus point and lock it where we want it.


Cheers, Hywel
Logged

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
Re: What COULD a camera be in 2018?
« Reply #12 on: July 31, 2018, 03:04:06 pm »

It's not just you...I agree, basic manual never goes out of style. If you know what your doing.

Peter

Basic manual does go out of style when there is a reliable and better alternative.

Case in point-autofocus for fast-moving subjects, quintessentially sports and wildlife. 

Sure, you CAN focus manually, and one wants to still be able to for some situations (like prefocussing where you know a car or a bird is going to be).

But the autofocus on a top-flight dSLR does a better and more reliable job in many situations and I strongly suspect that most photographers working with moving subjects now rely heavily on it, as I do when photographing people.


Similarly, writing computer programs in machine code is still the purest "meeting the metal" way to get a computer to do what you want. If you're really good, you may even be able to make the most performant code that way. But the arrival of optimising compilers meant that there was just no reason to do it in. In fact you'd be mad to- it's better to deal with your code at a higher level of abstraction. You get more done, more reliably, in less time. Machine code has gone out of style in all but a few specialised situations because writing high-level code is a more appropriate abstraction. You'd definitely not contemplate machine coding a quick web app or a macro to rename some files.

But... you still need a programmer to write the program. To decide what to do and how to do it. You're just dealing with the tasks at a higher level of abstraction.

The compiler just facilitates that. You still need to be a good programmer to make a good job of it.

Like "lock on THIS and keep it in focus" is a higher level of abstraction than manually twiddling the focus ring. You still need to be a good photographer to take photos where the most artistically appropriate and important thing is in focus. You need to be able to control that system, sure- that needs good technology and good user interface ergonomics.

But the truth is that few human beings can accurately focus a long telephoto lens or an 85 mm f/1.4 lens wide open on a moving subject as quickly or as accurately as a modern camera's autofocus can. So for many applications, manual focus has totally gone out of style. The main limitation now is controlling the initial lock-on, which is one of the things I think could do with a re-examination so we can just indicate to the automated system what SHOULD be in focus, and let it take care of the mechanics of keeping it so. 

  Cheers, Hywel

 

« Last Edit: July 31, 2018, 03:19:16 pm by HywelPhillips »
Logged

petermfiore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2705
    • Peter Fiore Fine Art
Re: What COULD a camera be in 2018?
« Reply #13 on: July 31, 2018, 03:18:45 pm »

Basic manual does go out of style when there is a reliable and better alternative.

As far as moving subjects go...zone focus works perfectly. DOF is a great way to go.

Peter

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
Re: What COULD a camera be in 2018?
« Reply #14 on: July 31, 2018, 03:19:56 pm »

As far as moving subjects go...zone focus works perfectly.

Peter

Really? With an 85 mm f/1.4 lens wide open and a moving model?

Hywel
Logged

petermfiore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2705
    • Peter Fiore Fine Art
Re: What COULD a camera be in 2018?
« Reply #15 on: July 31, 2018, 03:23:34 pm »

Really? With an 85 mm f/1.4 lens wide open and a moving model?

Hywel

For my kind of work...yes. Zone focusing works fine. Your original question "What COULD a camera be in 2018?" For me Simple settings.

Peter

HywelPhillips

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 64
Re: What COULD a camera be in 2018?
« Reply #16 on: July 31, 2018, 03:27:21 pm »

For my kind of work...yes. Zone focusing works fine. Your original question "What COULD a camera be in 2018?" For me Simple settings.

Peter

OK, I will rephrase my original post, having clarified in my mind what I'm getting at! ;)

What higher-level abstractions which have heretofore not been widely or universally implemented in digital cameras do you think would be useful and worthwhile?

If the answer is "none, I like my cameras with simple settings", you are cordially invited to be happy with what you've got!  :)

Best regards,

  Hywel Phillips

Logged

chez

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2501
Re: What COULD a camera be in 2018?
« Reply #17 on: July 31, 2018, 04:36:42 pm »

As far as moving subjects go...zone focus works perfectly. DOF is a great way to go.

Peter

What happens when you are shooting moving subjects using a very narrow DOF? This is where eye focus tracking kills zone focus.
Logged

chez

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2501
Re: What COULD a camera be in 2018?
« Reply #18 on: July 31, 2018, 04:40:28 pm »

For my kind of work...yes. Zone focusing works fine. Your original question "What COULD a camera be in 2018?" For me Simple settings.

Peter

Since you are not really using many of the advanced features in the camera...what is so complex after you initially setup the parameters. I know my TV has a lot of colour calibration and sound setup...but I did it once and now I just turn it on / off and change channels...all the complexity is not required anymore.
Logged

petermfiore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2705
    • Peter Fiore Fine Art
Re: What COULD a camera be in 2018?
« Reply #19 on: July 31, 2018, 05:12:10 pm »

Since you are not really using many of the advanced features in the camera...what is so complex after you initially setup the parameters. I know my TV has a lot of colour calibration and sound setup...but I did it once and now I just turn it on / off and change channels...all the complexity is not required anymore.

Nothing is complex... I prefer a basic approach. For street photography this works well for me.

Peter 
Pages: [1] 2 3   Go Up