Can DOF & bokeh be faked in-camera? (future)

Home · Register · Search · Software · Hosting · Software · Join Upload & Sell

Moderated by: Fred Miranda

Pre-order new Rokinon 16mm f/2.0 ED AS UMC lens

Pre-order new Sigma 18-35mm f/1.8 DC HSM lens

Bower 35mm f/1.4 price dropped to $379

FM Forums \| Alternative Gear & Lenses \| Join Upload & Sell
end

Can DOF & bokeh be faked in-camera? (future)

trale
Offline
•
Upload & Sell: Off

p.1 #1 · Can DOF & bokeh be faked in-camera? (future)

Over the course of browsing many excellent image threads in this forum, it comes as no surprise that many of the most striking images come from shooting fast lenses at max apertures to obtain the best bokeh, subject isolation, and DOF effects.

Of course such lenses (f/1.2, f/1.4) lenses are exotic and expensive.

So I'm wondering if in the future, might we see bodies capable to emulating pleasant DOF and bokeh via in-camera processing, so that one can achieve these effects with much cheaper f/2.8 or f/4.0 lenses.

Post-process blurring and bokeh is possible via Photoshop and plugins, but it is an work-intensive manual process. I'm guessing that it actually would be much easier for a camera body do to - since it is easily capable of determining the distance and focus of subjects.

Just as a quick example, Sony cameras have focus-peaking that displays what part of a scene is in focus and how much they are in focus. Taking this focus-data further, it seems to me that it would be easy for the camera to blur more the parts of the scene that are not in focus, to exaggerate their bokeh and blur, and result in a picture that looks like it was taken with a f/1.2 aperture.

And as the computation and data-processing power of cameras grow (at the pace of advancement of general computer tech), it seems this would be easily achievable in the near future.

Basically if this comes to pass (and it works well enough), we could just walk around with a general good 24mm-135mm F/4.0 lens and end up with pictures that seem to come from 24mm-135mm f/1.2, negating the need for a collection of expensive primes.

Any thoughts on this?

Dec 27, 2012 at 11:08 PM

p666
Offline
•
Upload & Sell: Off

p.1 #2 · Can DOF & bokeh be faked in-camera? (future)

Well you're not going to get the same light gathering ability of a wider aperture lens, but in theory seems plausible. Still seems a little prone to error, as it will only be an estimate from a two dimensional sensor. Would like to see the science behind it.

Dec 27, 2012 at 11:21 PM

deadwolfbones
Offline
• • • •
Upload & Sell: On

p.1 #3 · Can DOF & bokeh be faked in-camera? (future)

Several cameras already try to do this (the Pentax Q probably most prominent among them). It's not very good thus far.

Dec 28, 2012 at 03:58 AM

Bifurcator
Offline
• • • • •
Upload & Sell: On

p.1 #4 · Can DOF & bokeh be faked in-camera? (future)

It's very possible with the speed of current camera processors!

The are two ways I can see such a thing being implemented.

One way (for people shots at least) it would need to do a slightly more detailed face detection type object recognition and then blur around it. That's a 2D blur tho and usually doesn't look as good. I think this is the Pentax Q way mentioned above. It isolates the focus peaked elements and progressively blurs the other parts.

The other way would be for the camera to sample Z. This is already being done in prototypes in any of several different ways but the most interesting to me are the "Living Image" cameras of which Lytro already even has a product. In order to bend a lytro camera to your evil purposes here all it would take is some processing before the assembly of the LI image. This shouldn't be too hard. Heck, some chaotic types are already playing with the LI format and assembling their own - in a browser no less.

So, yeah, it's currently doable right now. But I don't think any camera companies are interested. Probably because not enough people even know what it is - let alone what DOF is and how it works.

Dec 28, 2012 at 04:15 AM

AhamB
Offline
• • • •
Upload & Sell: On

p.1 #5 · Can DOF & bokeh be faked in-camera? (future)

Something slightly related, but impressive: http://vimeo.com/43442146

Perhaps a camera could gather some data about the depth of a scene by moving through the focus range and analysing how things blur to use that in a more advanced way to artificially reduce the DOF.

Bifurcator wrote:
So, yeah, it's currently doable right now. But I don't think any camera companies are interested. Probably because not enough people even know what it is - let alone what DOF is and how it works.

Not to mention that they most won't want to lose sales of their expensive fast lenses. I can see such a function coming on compacts perhaps, but not high-end cameras.

Dec 28, 2012 at 04:49 AM

justruss
Offline
• • • •
Upload & Sell: Off

p.1 #6 · Can DOF & bokeh be faked in-camera? (future)

trale wrote:
Over the course of browsing many excellent image threads in this forum, it comes as no surprise that many of the most striking images come from shooting fast lenses at max apertures to obtain the best bokeh, subject isolation, and DOF effects.

Any thoughts on this?

I take issue with the premise. I don't think the ultra-thin dof images here or on the other boards are particularly striking. They scream amateurish to me-- a pursuit of what is technically possible mainly to highlight what is technically possible as different from the masses and their cheap, unsophisticated kit glass. I see ultra thin dof-- as a popular phenomenon-- as a status symbol.

There are some GREAT ultra thin dof images here and elsewhere. But I see stronger medium to near-full dof images than ultra thin dof images on the whole-- here and elsewhere. I think because most people seem to see ultra thin dof and subject isolation as sufficient for a compelling image. But it is not. And, BTW, for many fast lenses the smoothest bokeh is stopped down a bit, not wide open; optical considerations, subject-camera-background relationships and aperture blade shape play a bigger role here than ultra-wide aperture.

From this standpoint, I don't think the feature of in-camera, simulated thin dof is going to have legs. The main advantage to super wide angle lenses is light gathering ability. Expect better high ISO and IS to combat this main front.

You can add simulated thin dof in a much more sophisticated way in post if you must, on a big screen, with ever faster processing and more user input.

Dec 28, 2012 at 01:50 PM

Bifurcator
Offline
• • • • •
Upload & Sell: On

p.1 #7 · Can DOF & bokeh be faked in-camera? (future)

Hmm,

I just see it as another tool or technique. Some are trying to get good at using it while others actually are.

That's it... I don't think any deeper into it than that.

Dec 28, 2012 at 02:49 PM

Bernie
Offline
• • • •
Upload & Sell: On

p.1 #8 · Can DOF & bokeh be faked in-camera? (future)

With a generation of people coming from the ranks of P&S and smart phones / cameras, the reaction to a lot of art photography is "Why isn't the whole picture in focus?"

Maybe the role of Lytro is the opposite of what you suggested. Recreate in-focus planes for focus stacking to get more of the image in focus. And gain some of the sensor's resolution back at the same time.

But then there's nothing that says a camera can't provide a matrix of distance information which could be used in PP for creating better realistic bokeh.

Dec 28, 2012 at 03:12 PM

Mescalamba
Offline
• • • •
Upload & Sell: Off

p.1 #9 · Can DOF & bokeh be faked in-camera? (future)

Good artificial bokeh is near impossible thing. Probably only Lytro could do something with it, cause you need to software to actually see depth of scene.

That and almost everything in lens does have some effect on bokeh, so to mimic it is really not possible.

And Im glad its that way.

Otherwise f1.4 lens really arent expensive.. it depends what you want, but decent 50mm f1.4 is quite cheap..

Dec 29, 2012 at 12:38 AM

end
FM Forums \| Alternative Gear & Lenses \| Join Upload & Sell end