Optical Viewfinder RIP

[BJL]

I do not want EVF period.

An aid in that pre-visualization is an optical viewfinder or groundglass. It shows us the REAL and ACTUAL scene, altered only by the glass. An EVF shows instead an
Best regards,
Anders

To me "pre-visualization" means a preview of my results, including previewing the sharpness of in-focus parts and the extent of out-of-focus effects. The secondary image scattered off ground glass (or plastic) overstates DOF at low f-stops, and is too low in resolution to check fine details or precise focus. So _in these respects_, a good EVF can give a far more reliable preview.

That still leaves the issue of accurate rendering of contrast, and lag between what the EVF shows and what the sensor is seeing at the same moment, so I will add contrast/DR preview to my list of requirements for an EVF that unequivocally surpasses an OVF.

But before we reach that "better in every possible way" status, we are in the familiar "old vs new" debate, like film vs digital, where each alternative has advantages, and people take sides partly on the basis of their personal priorities. I for example rarely photograph moving subjects, and when I do it is with panning, so that the subject is not moving much in the VF, and I deal with DR by careful exposure settings, not on the basis of what I see on the viewfinder? So _for me_ a big, bright (even in low light) zoomable EVF image of good resolution that can accurately preview DOF is clearly preferable. And non-TTL OVFs with parallax errors close-up (I like macros) and no focus information at all are completely off my list.

P.S. why is this thread in the medium format forum? That is the last place the EVFs will take over!

[Enda Cavanagh]

I hope it won't be matching my vision or we're all doomed!

The next generation of electronic viewfinders will match human vision.   


See http://www.microoled.net/news/news-from-microoled

Cheers,
Miles

[bcooter]

I've heard the term "decisive moment" since I was an Art Director and later a photographer.  

"uhh, ooooh, aaahhhh, he/she is so marvelous at capturing the "decisive moment".  

I hear it today and most of it is just BS, because if every photographer was perfect at capturing the "decisive moment" they would only shoot one frame a session and be done with it and nobody would ever retouch a blinked eye, a crooked finger, hair that looked like a blown scarf, fix a crooked smile, remove a crew member's hand,  the list goes on.

In fact if you want to make a boring photograph, especially one that involves people, make a perfect photograph.  Move every finger, direct every strand of hair, the lips exact, the body in an exact way and the humanity will go flying off the image.

When I shoot a 35mm camera I get more "decisive moments" because I can shoot faster with less lag time.   When I shoot medium format, (yea, yea I know it's more "thoughtful") which means it's slower and less active which makes easier decisive moments but definitely not more.

So what's the difference between 1 frame every 3 seconds or 1 frame a second or 30 frames a second if you get the right shot?

If you don't think photographers constantly miss the decisive moment, then shoot a RED like a still camera for 30 seconds and click through the frames.  You'd be amazed at how many decisive moments there are.

And . . . it doesn't change your input or creativity one bit, it adds to it.  Regardless of fps you still compose, direct, light, capture (I said capture rather than record, so not to piss anyone off).

I really don't have a dog in this fight, because I use what people make and every camera has an advantage, every camera has a liability.  The goal to use the advantages to their fullest and try to make the liabilities an asset.

In fact I believe in both optical and evfs, as I use both and still cling to those old contax' with the crappy small viewfinder . . . because I just like the camera.

Regardless of the above, regardless of what I think I am positive of this;

Things change and either you go forward, are you get passed by.

That's not the nature of photography, that's the nature of the world.

IMO

BC

[hjulenissen]

Probably all arguments are mentioned by now, but anyways:
1. The OVF of a large professional DSLR is not the same as a cheap Canon/Nikon one. For most users it is not a choice between a large, bright OVF and a cheap EVF, but between a cheap OVF and a cheap EVF. Cheap electronics tend to improve a lot faster over time than cheap glass does.

2. EVF means full-time sensor grabbing, probably more battery drain and heat generated, may mean different compromises in the sensor itself, and may mean that you might as well go with CDAF (or the unusual solutions by Nikon and Sony). A seemingly isolated change leads to large changes elsewhere.

3. EVF means helpful overlays, no doubt. (and annoying ones). I think Guillermo has some nice ideas that probably will be scrapped by the iPad-swinging interaction-designers :-)

4. Some talks about "seeing the scene as it really is, not after distorted by the jpeg engine". Funny that through-the-lense viewfinder is a good thing, while through-the-sensor is a bad thing. I tend to think that having a conservative visual feedback on how the sensor clips white and black, and how noisy it is, can be a good thing. Same with color limitations and WB. If you really want to see how the scene really is, you need to open your other eye.

5. Latency will be a problem for those who have little time. I am sure that a $4000 sports-monster mirror-less could optimize the sensor/processing/EVF a lot more than current designs, but probably never close to the speed of light through 10cm of glass and air.

I would be very excited to see mirror-less technology applied to 35mm or MF sensors. Perhaps Panasonic would do like Pentax, take their existing software/hardware solution, put it into a larger sensor housing, add some processing power. If nothing else, to prove that the concepts works and for the publicity.

-h

[LKaven]

BC, the "decisive moment" has as much to do with the photographer's "decision" as it does with the elements of the scene being "decisive."  H C-B was not as skilled a philosopher as he was a photographer, but he was onto something fundamental about bringing inspiration to the point of action.  Surely, you are right, and there is not just a single such moment.  This is to say that you are also correct, but H C-B walks away unscathed.

It does seem that the era of "pre-roll" is nearly upon us.  In audio recording, pre-roll is a feature of a recorder whereby it is always recording and retaining 30 seconds worth of material.  Each time you press the "record" button, the recorder has the previous 30 seconds of material already in the can.  Invaluable.  Correspondingly, there is no reason for the camera to await your shutter release before taking some pictures on its own.  I believe the Nikon 1 does something like this.  And in the era of fast image pipelines (EXPEED3 and beyond), why not?  I agree with you, there are a lot more moments in there than one supposes.

BTW, some of the models I've seen, the nerdier girls, give so many interesting looks per second, that I think I'd need a rapid-capture device to sort it all out.  Working at a slower speed, it's be too easy to dismiss some of these gems as awkward, when I think they are the true fashion aliens that a good photographer wants.  Have you ever felt this way?

[BJL]

I would be very excited to see mirror-less technology applied to 35mm or MF sensors. Perhaps Panasonic ...

Good points about the issues still to be dealt with. Note though that lag does need to exactly equal that of an OVF (about a nanosecond), but just for the difference to be imperceptible (about 10 milliseconds?).

On speculation about an EVF-equipped camera in one of the traditional film formats, one common guess is that the first such camera will be a Sony SLT, since Sony seems to be moving on from traditional DSLRs. But my guess is that the SLT approach is a temporary solution to the challenge of combining EVF with good, fast, flexible auto-focus, and that it will be surpassed soon enough by either
1. In-sensor PDAF/CDAF hybrid, as seen first in the Fujifilm F300 EXR and more recently in the Nikon One models,
or
2. Making CDAF good enough, which seems to be the Panasonic plan.
Of those, it might be interesting if Fujifilm decides to try again with a medium format sensor; this time with live view and Hybrid AF.

[ErikKaffehr]

Hi,

I used to be shooting show jumping, with SLR and film. Neither the system or the photographers was fast enough to react, you needed to predict the decisive moment and realease the shutter like 0.08 seconds before. With electronics the display may lag a bit more but the response may be a bit faster as you don't need to wait for the mirror to move out from the optical path.

Best regards
Erik

BC, the "decisive moment" has as much to do with the photographer's "decision" as it does with the elements of the scene being "decisive."  H C-B was not as skilled a philosopher as he was a photographer, but he was onto something fundamental about bringing inspiration to the point of action.  Surely, you are right, and there is not just a single such moment.  This is to say that you are also correct, but H C-B walks away unscathed.

It does seem that the era of "pre-roll" is nearly upon us.  In audio recording, pre-roll is a feature of a recorder whereby it is always recording and retaining 30 seconds worth of material.  Each time you press the "record" button, the recorder has the previous 30 seconds of material already in the can.  Invaluable.  Correspondingly, there is no reason for the camera to await your shutter release before taking some pictures on its own.  I believe the Nikon 1 does something like this.  And in the era of fast image pipelines (EXPEED3 and beyond), why not?  I agree with you, there are a lot more moments in there than one supposes.

BTW, some of the models I've seen, the nerdier girls, give so many interesting looks per second, that I think I'd need a rapid-capture device to sort it all out.  Working at a slower speed, it's be too easy to dismiss some of these gems as awkward, when I think they are the true fashion aliens that a good photographer wants.  Have you ever felt this way?

[bcooter]

snip

It does seem that the era of "pre-roll" is nearly upon us.  

If the best EVF came out today and every camera company adopted it, that doesn't mean you can't use legacy cameras and shoot to a piece of ground plastic to your hearts content.

What comes with an EVF hopefully will come with different formats.

It's kind of silly that we look at the "medium format" segment of digital and think it's actually medium format given that only a small percentage of backs/cameras are actually full 645 frame and 645 was the smallest of all medium formats.

I'd love to see a square that we could crop in viewfinder to any format we wished. 

Actually going from 2:1, 16x9, 24x36 to 645 in my view of the electronic world, 24x36 is the medium format and probably the most hobbled and less creative format of all.

I like a 4:3 format, kind of like 16x9, love 2 to 1 and 24x36 falls somewhere in between.

I guess what I'm saying is I'd love to see real innovation.  I was hoping RED would go forward with a larger format sensor if only to allow us to shoot a panoramic look, or square or vertical all within the same camera.

Of course I want to decide beforehand what format fits the scene, or the creative brief, but having that option would be amazing . . . at least for my work.

Since most medium format digital cameras only shoot close to 1.2 or 1.4 seconds a frame, I'd love to push the button and get one and a half seconds of images.  I know it will make for a more natural image.

As far as hitting the decisive moment, when I do it I'm genius, when I miss it my heart sinks.  In fact with mirror based cameras the shot I don't see is the good one, the shot I do see through the viewfinder is the one we missed.

As far as shooting models or actors, so many times on set while were placing lights, or preparing the camera I'll look over and see the most amazing looks.  Every day I walk over and say . . . what you just did, do it again, but let me have a lens pointed at you.  The great on camera talent can repeat it with perfection, the really great on camera talent can expand upon it.

Still, I know that making cameras must be hard, must be expensive and must have a market to make a profit, but I'd love to see a complete rethink of how we work, what we work with and how the tools respond.

We're still in the shadows of film cameras and though I'm not an engineer, I assume we don't have to be.



IMO

BC

[hjulenissen]

Good points about the issues still to be dealt with. Note though that lag does need to exactly equal that of an OVF (about a nanosecond), but just for the difference to be imperceptible (about 10 milliseconds?).

If televisions are an indication, 10ms is hard to achieve, even in "gaming" mode. Seems that 1-3 frames (at 60fps) is usually added for whatever reason (HDMI buffers, image processing,lcd panel rise-time...)

Now, doing dedicated camera electronics might offer other opportunities.

10ms is occasionally quoted as a threshold for electronic musical instruments latency. Guess that photographers have no stricter timing requirements than drummers?

-h

[eronald]

And that, in a nutshell, is my definition of photography hell!

Rob C

Hey Rob,

 Video camera operator is the job still photographers get when age takes their reflexes away

Edmund

[BobDavid]

Hey Rob,

 Video camera operator is the job still photographers get when age takes their reflexes away

Edmund

Pithy!

[LKaven]

Video camera operator is the job still photographers get when age takes their reflexes away

...and their spines are still healthy enough to sacrifice.

[BJL]

My 10ms guess corresponds to 100fps, and only video resolution sub-sampling is needed (never mind a little moiré in the VF, especialy if it only in a high-speed mode for action photography), and there are already some compacts from Casio for example with faster video rates than that. My guess is that the same increase in read-out and processing speeds will also make CDAF fast enough and good enough within a couple of Moore's Law performance doubling generations, meaning about three years.

[Anders_HK]

To me "pre-visualization" means a preview of my results, including previewing the sharpness of in-focus parts and the extent of out-of-focus effects. The secondary image scattered off ground glass (or plastic) overstates DOF at low f-stops, and is too low in resolution to check fine details or precise focus. So _in these respects_, a good EVF can give a far more reliable preview.

I would argument that what you describe is post-visualization (not pre-visualization), since at time for choosing aperture, with a trained photographic mind and eye, you should already know what will be correct aperture for the desired DOF. Pre-visualisation refers to the decision for making the important exposure decision prior to pressing a shutter or looking at a preview of the image. Simply speaking, I believe the eye should be at the subject and not at any display of it, whether at back of camera or EVF.

As one of posters above mentioned H C-B, it is a good example that the decision for a photograph should be in anticipation of the correct/decisive moment. Thus any live view will fail at capturing as decisive moment (no matter how many frames per second your camera can shoot).

Exception from the above is a scene that is static and in which the light does not change, in which case shooting tethered is an option because it gives you access to all processing means for your RAW file at same time of preview and for making the exposure decision.

Best regards
Anders

[BJL]

... with a trained photographic mind and eye, you should already know what will be correct aperture for the desired DOF. ... I believe the eye should be at the subject and not at any display of it, whether at back of camera or EVF.

... H C-B, it is a good example ... any live view will fail at capturing as decisive moment ...

Exception from the above is a scene that is static and in which the light does not change, in which case shooting tethered is an option ...

This is sounding a lot like "digital will never be as good as film" traditionalism and not much like a balanced, rational comparison of the alternatives. And frankly, the suggestion that one can always avoid an EVF with stationary or slow moving subjects by tethering is ludicrous, as if that only happens in a studio or with a tripod. Most of my stationary subjects are reached with a lot of walking and I confess, most are photographed hand-held.

[Ch-Jaeger]

EVF will come no matter what, because it's cheaper for the manufacturer in the long term. Manufacturers don't care for decisive moments.

I have always enjoyed looking at a ground glass or a 6x6 matte of a MF camera. I still think this is superior to any LCD-Liveview, although it is essentially the same thing only one is electronic. I don't know why that is and I'm not that old to be emotionally connected to either of these techniques...

--
Christoph

[hjulenissen]

As one of posters above mentioned H C-B, it is a good example that the decision for a photograph should be in anticipation of the correct/decisive moment. Thus any live view will fail at capturing as decisive moment (no matter how many frames per second your camera can shoot).

If the sensor grabs at full resolution all of the time, then no matter what the latency is, you will experience a delayed version in the EVF, and have equal oportunities to grab the "decisive moment" by clicking at the right time according to the EVF.

Assuming that even though your view may be 100ms delayed, the grab that ends up on flash is also delayed by the same amount.
-h

[John R Smith]

There is something absolutely magical about the 6x6 cm image which appears when you pop-up the WLF on a Rollei or a Hasselblad or any MF SLR. I felt this very strongly the first time I used one (an old Pilot 6, quickly followed by a Yashica 124). The fact that the image is reversed left-to-right strangely adds to the effect. We gaze enraptured at this little square picture which represents our chunk of reality, and which we can capture, fix, make permanent at the touch of a button. The whole experience is nothing like peering through a 35mm viewfinder (or any prism, for that matter). And when you have a good lens and screen (an acutematte on a Hasselblad) the quality of the image is superb. Why on earth would I want to trade that for an EVF?

John

[ErikKaffehr]

Hi,

I absolutely hated the viewfinder image on the Hasselblad 500C (which I never owned) and the Pentax 67. I also had a waist level finder for the Pentax, but I never used it without the loupe. So, I guess my experience is different. I had the meterless prism on the Pentax 67, as I was a spot meter addict and felt no need for light meter. Quite possible that the prism with light meter was better.

Later Hasselblad 500 had a groundglass made by Minolta, they may have been better than my initial experience.

I'm not saying that MF viewfinders are bad, just that the couple I have used were bad.

Best regards
Erik

There is something absolutely magical about the 6x6 cm image which appears when you pop-up the WLF on a Rollei or a Hasselblad or any MF SLR. I felt this very strongly the first time I used one (an old Pilot 6, quickly followed by a Yashica 124). The fact that the image is reversed left-to-right strangely adds to the effect. We gaze enraptured at this little square picture which represents our chunk of reality, and which we can capture, fix, make permanent at the touch of a button. The whole experience is nothing like peering through a 35mm viewfinder (or any prism, for that matter). And when you have a good lens and screen (an acutematte on a Hasselblad) the quality of the image is superb. Why on earth would I want to trade that for an EVF?

John

[LKaven]

If the sensor grabs at full resolution all of the time, then no matter what the latency is, you will experience a delayed version in the EVF, and have equal oportunities to grab the "decisive moment" by clicking at the right time according to the EVF.

Assuming that even though your view may be 100ms delayed, the grab that ends up on flash is also delayed by the same amount.
-h

As soon as the pipeline (reset->expose->read->render) becomes fast enough, the "pre-roll" option could become a viable tool.

Imagine if a single shutter press would record three images taken in the period of 1/2 second /before/ the shutter button was pressed, as well as (nominally) 2 more in the 1/4 second following?  You could get one of the moments you just missed this way. 

My biggest problem with the EVF is the moments that it omits by necessity.  You are getting only a fraction of the events that are occurring continuously in real-time.