Originally posted by IchabodCrane If that is your objective then nothing can be better than an OVF. As stated before, no practical EVF can render the dynamic range that a sensor can record. Will it someday? I don't venture to guess but if one does I doubt it will be linked to the number of transistors in a processor doubling every two years (now believed to be every three years). Also, no EVF can "gain up" to match a brightly sunlit scene which means there will be an additional lag that no processor can ever address -- the lag in our eyes adjusting from sunlight to whatever an EVF can pump out.
What's wrong with instant review blinkies and histograms for finding clipped highlights?
I use various methods of getting the exposure right, and want something better (for me).
For much of my photography I use manual exposure. That includes studio work and (a lot of the time) airshows and motor sports. I have a Gossen Digipro F exposure meter, but if I am going the trouble of taking a separate reading, I've found I'm better off taking a photo and looking at the instant review blinkies and histograms, as you say. (I'm told that using a separate exposure meter impresses the models! I don't think they are that naive). If you looked at my photos from an airshow shoot you would see the occasional photo of (say) public-address loud-speakers against the sky. Those were for such review purposes.
But this introduces delay. What I would like is a digital preview that I can see as I'm framing the next photo. Perhaps the next airplane is hurtling towards me. Or the model has moved from the carefully metered original position. Perhaps her arm is now too close to a softbox and has clipped. Or she has added jewelry that sparkles and has clipped. (Both real examples).
Note: I've never claimed that Moore's Law is entirely responsible for advances in EVF performance. (And Moore's Law isn't just about packing density, but also about speed). The reason why EVFs are steadily improving year by year is largely irrelevant - what surely matters is that they
are improving. I hope that isn't in dispute!
Much of my experience of the gradual replacement of mechanical parts with electronic parts over the last few decades often preceded increases in processor power, and were not necessarily driven by it. There is a sort of symbiotic relationship - someone works out how to replace mechanics with electronics in a particular subsystem, there is now a justification for more R&D in that area of electronics, features and performance are achieved that were perhaps never expected, and so on, perhaps like a feedback loop. Moor'e Law, improvements in materials science, virtually free replication costs for firmware and software, etc, carry things forward once started.
Here are what I said earlier, before the distraction of whether Moore's Law actually exists:
Post 1 Post 2