[pcarfan]

KungPow, it still doesn't make sense, as the automatic exposure compensation is there when the bulit-in light meter goes wrong - They have put in all the technology they can to meter a scene, and get the proper exposure (much more advanced than the TV). The exposure compensation is there, when all these technology to detect the scene goes wrong, you can manually correct it.

I would how ever agree, that The K-7 will have better metering.

[mattdm]

Of all the cameras I've used, the light meter can only be trusted to adjust the exposure to 18% on average. Even complex matrix systems do this. So a matrix metered snow scene will still under expose. How could the camera know that the scene was bright?

So, that's the logical flaw. In fact, all the camera knows is that the scene is bright. It "knows" that in the direct sun in the middle of summer, and it "knows" it when shooting a field covered in snow. Either way, the sensor reads "wow, it's bright". What it can't know is whether the subject "should be" inherently bright (like snow) or just happens to be really bright looking but really should be exposed to look normal. Adding another sensor just adds more "yup, it sure is bright", and no more knowledge about the scene.

Matrix metering attempts to add artificial intelligence, by using multiple sensors to compare different parts of the frame to then guess what the subject is — for example, dark center with bright background is probably a backlit subject, so expose to get the center at 18% even if the background blows out. Having a second whole-frame or ambient light sensor isn't going to add much more information.

However, one could imagine a matrix metering system that seems a lot smarter simply by giving it more to work with. The current one has 16 areas, and presumably only works on luminosity. Instead, the entire live-view sensor could be read and a small full-color image thumbnail constructed. This thumbnail could be compared against a much, much larger database and the best match used. Flash memory is cheap — put half a gigabyte of possible images in there and a fast hash algorithm to match 'em up, and there you go.

[KungPOW]

KungPow, it still doesn't make sense, as the automatic exposure compensation is there when the bulit-in light meter goes wrong - They have put in all the technology they can to meter a scene, and get the proper exposure (much more advanced than the TV). The exposure compensation is there, when all these technology to detect the scene goes wrong, you can manually correct it.

I would how ever agree, that The K-7 will have better metering.

All I was trying to propose is that maybe Pentax found a way to solve a common problem with all metering systems, and the solution was as simple as twin meters.

I only came up with this because of the picture Ned posted shows the SV with an external meter.

I am sure that what I have proposed is not what Pentax has done. The simple explination for the "eye" on the K-7 is that it is a focus assist light. Sometimes I run with my ideas, and forget the "box".

But I think the question I ask is still valid:

How do you design a metering system that can tell the difference between a bright snow scene and an open pit coal mine?

[Class A]

"I Want You (She's So Heavy)"
...
"Don't Let Me Down".

Nice Medley!!!

Of course, finally we'll all say "I Should Have Know Better".

[PentaxPoke]

Some folks are really overthinking this thing (as Ned implied). It is a "FEEL" thing. Not some new technology or camera feature related to what you are seeing in the picture.

Using Ned's words:

"Small detail"
"Something that can be felt and appreciated"

[mattdm]

How do you design a metering system that can tell the difference between a bright snow scene and an open pit coal mine?

It can already. The difficulty is telling the difference between a coal mine and a gravel pit, or between snow scene and an empty parking lot. And as I said, the answer is: smarter matrix metering. Make more information available to the system both in terms of input and stored knowledge, and give it more computational power to work with, and presto, solved problem.

[gnaztee]

Maybe, because it makes no sense to speculate about yet another chip worth 50 cents?

How stupid of me. I mean, I should have thought before I posted such silly speculation on this forum of high-minded, concrete ideas. Maybe I should have asked about a 50 cent dedicated ISO button, or does price have something to do with it? Do my speculations need to be more expensive? Please, oh please, somebody give me guidance on this!

[Yohan Pamudji]

It can already. The difficulty is telling the difference between a coal mine and a gravel pit, or between snow scene and an empty parking lot. And as I said, the answer is: smarter matrix metering. Make more information available to the system both in terms of input and stored knowledge, and give it more computational power to work with, and presto, solved problem.

What you're suggesting is levels of magnitude more complex than any form of matrix metering. The camera would need to have knowledge of the scene and what is actually being photographed instead of just how much light is being reflected by the scene. The camera would have to be able to identify and label scenes as "snow", "backlit portrait", "cave", etc. Good luck cramming that into a camera anytime soon. This level of artificial intelligence is not even easily obtainable with the most modern AI techniques, let alone some technique that a DSLR CPU can handle.

[falconeye]

Maybe I should have asked about a 50 cent dedicated ISO button, or does price have something to do with it?

Maybe, my answer was a little bref. What I meant is that you get what you pay for.

E.g., with the iPhone which already is a camera with one of those little bargain A-GPS chips (small as the tip of a match) one could learn how useful it is.

Basically, it switches on with the camera wake-up and gets a fix minutes after you took the photo. And the K-7 wouldn't be able to triangulate cell phone

So yes, it may be useful for a large scale positioning. But to actually geotag images one would still need a logger solution which works like a champ btw.


So, what I wanted to say ... maybe, it isn't speculated more because it wouldn't be such a hot or useful feature anyway.

[mattdm]

What you're suggesting is levels of magnitude more complex than any form of matrix metering. The camera would need to have knowledge of the scene and what is actually being photographed instead of just how much light is being reflected by the scene. The camera would have to be able to identify and label scenes as "snow", "backlit portrait", "cave", etc. Good luck cramming that into a camera anytime soon. This level of artificial intelligence is not even easily obtainable with the most modern AI techniques, let alone some technique that a DSLR CPU can handle.

It's orders of magnitude more data, but it's still relatively simple pattern matching — the level of AI isn't really increased. Matrix metering already tries to identify and label scenes as "snow", "backlit portrait", etc. — it just doesn't have much to work with. (Pentax entry-level dSLRs even try to do fancier things than just adjust the exposure, like determine the appropriate scene mode to use.)

The key point is that it doesn't actually have to "know" anything. It just has to have a relatively good match for the scene.

I agree we're unlikely to see it any time soon because it'd probably add $50 in components to the camera. More, if you want to do it outside of live view. And collecting — and selecting! — the exposure database would be a serious undertaking.

[Ishpuini]

Sadly it means to use all the new function of the K-7's "soul" we will have to cut our hair to look like the fab four.

(oh well, goes to get scissors).

Oh D*mn... My hair will never again grow to be long enough, does that mean I'll have to skip the K-7?



Wim

[Mik]

I think he's probably referring to the strap. Maybe the K-7 strap is made of leather.

[sezme]

I think he's probably referring to the strap. Maybe the K-7 strap is made of leather.

Wow - that's quite a leap considering that the camera in the photo doesn't even have a strap.

I think what Ned is talking about is simply this: the add-on exposure meter at the top is unique and gives the SV a unique look and feel; it has a certain something, a certain cool factor that no other camera had. I think Ned's thinking (hoping) that there's something special about the K-7 that will trigger that same visceral response, something that will be obvious once all the specs are revealed.

[Mik]

Wow - that's quite a leap considering that the camera in the photo doesn't even have a strap.

It looks like there's a leather strap hanging below his left hand, even though it's not attached. That would certainly be something to be felt.

[falconeye]

Instead, the entire live-view sensor could be read and a small full-color image thumbnail constructed. This thumbnail could be compared against a much, much larger database and the best match used.

Is this necessary?
- With only 1 zone metering: 18% gray (and you would like to be able to differentiate between snow and asphalt).
- With only 16 zone metering: Some guessing for more than 18% gray
- With, say 100,000 zone metering (LV): Make the brightest zone 90% gray.
- With, say 15 million zone metering (actual image): Make the brightest zone (aka pixel) 100% gray, aka "expose to the right". Or make 99.9% of pixels to expose to the right.

So, the more zones you have, the less intelligence you require.