[Rener]

The post I replied to was the usual Pentaxiam post of 'Nikon's are blurry/soft/no detail', which I decided to have a go at. There isn't much difference in picture quality in the D90 v K20D. Heck, all DSLRs these days are awesome. It's just gets up my goat when people decide that because they've got X brand, all other brands are inferior. I'm pretty sure nearly anyone would struggle to tell the difference on a photo taken with a D90/K20D/50D once it's been processed.

From Imaging Resource;

Comparison between a $3399.99 Nikon D700 body and $999.99 Pentax K20, in Canadian dollars, Henry's Camera.

ISO 3200



ISO 6400

[grishazzz]

Lower fuel *if* you've got decent stick skills. I've also heard computer-aided automatic transmissions have reduced this gap to close to zero given most people's stick skills.

Not true if you are using the meter as I described. When pointing directly at any given area under a given light, it always tells me very accurately what shutter speed is needed to get an average exposure of a little darker than 18% gray. The part it *doesn't* necessarily do well is figure out where to point in order to get useful meter readings, or how to use the information gleaned from doing this to figure out what exposure I might actually *want*. So that's where I come in. I use the camera as a foolproof light meter - to tell me how what shutter speed I need to get to 18% gray in different areas of the scene - but make I the exposure settings myself based on my own notion of how I want the exposure to come out. Works every time.

I know how to drive stick (14 years).... :-) automatic tranny will never get close to the manual in terms of fuel consumption! never! thats why in Europe where gasoline is expensive, people still drive stick!

Ok, let me ask you this: If camera is capable to meter the scene correctly, why do I have to adjust the EV? :-) Nothing is perfect on earth..... :-)

[Marc Sabatella]

Ok, let me ask you this: If camera is capable to meter the scene correctly, why do I have to adjust the EV? :-)

Because not all scenes average out to a little darker than 18% gray, which is what the meter is calibrated to yield. The meter is essentially *perfect* in telling you how to achieve an exposure that is a little darker than 18% gray (when using center-weighted or spot metering, anyhow, which are "dumb" but completely effective). As I keep saying, it's up to the photographer to decide how to *use* that information - if you decide you want the scene to look lighter than 18% gray, you need to adjust the exposure. It's the decision making process that I don't trust the camera with, which is why I don't use multi-segment metering.

So again, the meter - in center-weighted or spot mode and pointed at an appropriate target - is essentially *perfect* in telling me what exposure will yield a little darker than 18% gray. It is completely trustworthy in that respect. It just isn't very good at all in telling me what exposure will give me the results I *want* - that's something I have to figure out for myself, using the very accurate information given me by the meter as a starting point.

[ManuH]

It's the decision making process that I don't trust the camera with, which is why I don't use multi-segment metering.

That's interesting, myself I almost always use the multi-segment (or spot for some tricky situations) but I may try center-weighted instead. There is a nice explanation of how the Nikon Matrix metering works by Ken Rockwell. You can see how the system tries to guess the "correct" exposure. Very instructive and also shows why you cannot rely on it to get the exposure *you* want all the time (funnily on a lot of examples I prefer the non-AMP version).

But when all is said and done, with digital it's easy to correct any error. You have the histogram and the blinking highlights and shadows to help you.

[Marc Sabatella]

What I like about center-weighted is that it combines the best features (to my way of thinking) of mulit-segment and spot. Like multi-segment, it can give a semi-reasonable guess at a good exposure just by pointing and shooting - or in my case, using M mode, pointing and hitting Green button. And unlike multi-segment metering, it won't underexpose to protect a small highlight, leading to brighter expoures overall (I more often need to dial in negative compensation than positive). But like spot metering, you can use it to get a read on specific areas of a scene. It's not as precise as spot, of course. So you can't use it to meter specific objects. But it's very good at giving me a general "average" read of the shadow area of a scene versus the light area of scene, giving me a quicker idea of overall dynamic range I am dealing with than having to spot meter a bunch of objects.

Using the meter in this way, I spend a few seconds familiaring myself with the scene(s) around me and what shutter speeds will be needed to expose for the lights or for the shadows, and then I can pretty much stop thinking about metering and just start shooting, adjusting shutter speed accordiing to whether I am interested in the light or the shadow, based on what I learned in my original scan.

And this is what I mean in saying the data itself is *extremely* reliable. It basically never reports anything other than the proper exposure value to render a given area as a little darker than 18% on average.

And BTW, this is why the DOF preview is so important to me with manual lenses - as you can, seeing the meter during this scan is essential to the whole deal.