[blende8]

At least the images get brighter with higher aperture values.
This is an agreement, probably an universal one.
With my camera I normally get overexposure from about f4 on, with max. values of about +2 from f8 on. It depends a bit on the lens and the environment, but the general tendency is that the images get brighter with increasing aperture.
I think these variances of about 2 stops are not tolerable.

[borno]

Mine is about like jonas', 1-1.5 stops over-exposed below f4 with the stock screen. I've been compensating by taking a reading, then increasing shutter speed 2 or 3 clicks and checking the histogram (PIA). Pentax emailed me and said "the ds screen won't fix this problem", but it looks like it does to me.

[Jonas B]

Here it is again, but now with the kit lens added (the only DA lens I have):

[carpents]

Thank you for that Sean. I understand what you mean with the difference between a grey card and your cardboard: as the whole scene is bigger than the cardboard it doesn't give a true "grey card value" in return. In my series the grey card filled the viewfinder.

The notebook did in fact take up the entire screen with the 200mm.

I wonder how the differences between mine and Seans curves can be explained? Has Pentax done anything about this during the production (I know my camera is a late one compared to yours)?

Mine was indeed from the first batch.

I did a very similar test for a fellow forum member without a gray card, and it turned out the same. What I noticed, however, was that the Konica-Minolta 7D that I also used in the test had a steeper curve than the K10. The DS was indeed pretty flat.

I'm going to add the DS into the mix and post back the results later.

[jeffkrol]

Thanks Jonas. Looks like the DS screen isn't perfect either. Though from f2.8, and throwing out the anomoly at f4 it's pretty good if you rise the baseline to 110. Canon's "baseline" is usually at 127 (they fudge the iso for tha ANSI standard, See Doug Kerrs wonderful papers. Articles by Doug Kerr
Carpents, I take it back, Jonas's are uglier
Jonas re: the 2 added lines. Seems counter -intuitive. You would think that if the orig K10 screen was brighter than the DS screen, the curves should be reveersed.. anyways, flat even underexposure is much easier to deal with......

[Jonas B]

Thanks Jonas. Looks like the DS screen isn't perfect either. Though from f2.8, and throwing out the anomoly at f4 it's pretty good if you rise the baseline to 110. Canon's "baseline" is usually at 127 (they fudge the iso for tha ANSI standard, See Doug Kerrs wonderful papers. Articles by Doug Kerr
Carpents, I take it back, Jonas's are uglier
Jonas re: the 2 added lines. Seems counter -intuitive. You would think that if the orig K10 screen was brighter than the DS screen, the curves should be reveersed.. anyways, flat even underexposure is much easier to deal with......

No the DS screen doesn't make the K10D exposure perfect... But I don't think anyone expect a modern Pentax to exposure perfect, not even when switching viewfinder screens. And yo are right; the DS screen works much better. The little dip at f/4 is puzzling: I have it with another lens as well (not included in the graph and even the Zenitar jumps up from a value of 95 at f/4 to a straight line around 114 between f/5.6 and f/11.

When I use the kit lens I have a positive compensation dialed in, no problem. And the curves are counter-intuitve for the DA lens, yes. But, it shouldn't be a big surprise to you: It is the same with the Takumar lens (and the Zenitar and... and I have double checked my numbers...).

I was for a long time a "97" man. Now I am more leaning towards "110". It is of little importance though as I always try to exposure to the right. The Canon vs ANSI thing is interesting in theory but quite unimportant from a practical point of view.

Jeff, when I bought the K10D I was blinded by the feature set and VPN and gave all this a lot of thoughts. I have a hard time believe it myself but I actually forgot about the exposure problems when I drove away to the camera store. Now I don't know. Spoiled by the 5D (in memory now) I just want a camera that works giving me no surprises. I don't know what to do really about this.

I guess I can give up and just buy DA lenses. Or I don't buy anything more and contemplate the situation. As I sold most of my gear I can afford jumping ship if needed. Any ideas? I would like to have "Live view" and a fast, good normal. Olympus?

[carpents]

I tested some more with 8 lenses this time, throwing in some different coatings as well. I used the Jupiter-9 on both cameras and the Nikon Nikkor-S 50/1.4 on the DS only (it doesn't quite fit on the K10). I also used the D-FA 100/2.8 Macro using the aperture ring instead of dialing in the aperture.

I also tried the Super-Multi-Coated Takumar 135/2.5 both in M mode and Av mode on the K10. It made a big difference.


Note - in this test I used a white-ish board and filled the frame. Most everything else is the same as the first set except I used a tripod.

The data:


The graphs tell quite a story.

The K10 (note the blue line is the Aperture Priority line):


The DS:

[SWEngineer]

JonasB wrote:

I just want a camera that works giving me no surprises. I don't know what to do really about this.

Hi Jonas. I probably don't belong in this discussion (still just have the lowly DS ).
But from Sean's latest test, it looks like all the lenses are consistent at a given aperture. So, maybe a good 'compensator' strategy for K / M lenses in M mode would be:

• Always meter (green button) at F8 (or whatever F you like that works for all your lenses).
• Add whatever fixed exposure offset works for you based on Sean's curve or your own.
• Compensate again if needed for the scene & how you metered it.
• Adjust F#, shutter speed, iso as desired to keep the exposure matched. (e.g. F8 to F5.6 ---> cut ISO or increase shutter speed by factor of 2 etc)

Actually, I use the DS this way a lot... manually resetting the TV & AV in proportion as desired based on the initial meter reading (which I admit isn't consistently at F*).

To further expose my ignorance, is this an issue with both spot & center weight metering or just one?

-Mark

[jeffkrol]

No the DS screen doesn't make the K10D exposure perfect... But I don't think anyone expect a modern Pentax to exposure perfect, not even when switching viewfinder screens. And yo are right; the DS screen works much better. The little dip at f/4 is puzzling: I have it with another lens as well (not included in the graph and even the Zenitar jumps up from a value of 95 at f/4 to a straight line around 114 between f/5.6 and f/11.

When I use the kit lens I have a positive compensation dialed in, no problem. And the curves are counter-intuitve for the DA lens, yes. But, it shouldn't be a big surprise to you: It is the same with the Takumar lens (and the Zenitar and... and I have double checked my numbers...).

I was for a long time a "97" man. Now I am more leaning towards "110". It is of little importance though as I always try to exposure to the right. The Canon vs ANSI thing is interesting in theory but quite unimportant from a practical point of view.

Jeff, when I bought the K10D I was blinded by the feature set and VPN and gave all this a lot of thoughts. I have a hard time believe it myself but I actually forgot about the exposure problems when I drove away to the camera store. Now I don't know. Spoiled by the 5D (in memory now) I just want a camera that works giving me no surprises. I don't know what to do really about this.

I guess I can give up and just buy DA lenses. Or I don't buy anything more and contemplate the situation. As I sold most of my gear I can afford jumping ship if needed. Any ideas? I would like to have "Live view" and a fast, good normal. Olympus?

How about writing to Sigma and getting them top produce a K mount Foveon based DSLR. I'm sure Hoya will sell them the rights pretty cheap... Sorry been reading up on the "merger". Sigma was my second choice, and to be honest, I'm pretty sure the SD9 was out before the D. ALMOST pulled the trigger. Buying all new lenses (had a few Pentax, including the Tamron90 I loved so much) really turned me off. In hindsight, I did just that anyway.... but it was helped by a few used purchases...
As to Olympus, good brand, but too small a sensor for me.

[Jonas B]

Hi Sean - good to see another test round, and such a exhaustive one. This is puzzling. Your curves are pretty much inverted to the curves I get. I have no idea about how to explain that.

Secondly; all our curves are totally unacceptable. In practise we have to accept them of course. The only possible thing to do, as I see it, is to do some experiments with other viewfinder screens. It's also possible to use the DS screen and adjust the exposure a bit. But, I would like to have it more spot on.

Well, I guess this is another chapter in the "Learn your equipment"-book.

best,

[Jonas B]

How about writing to Sigma and getting them top produce a K mount Foveon based DSLR. -snip-
As to Olympus, good brand, but too small a sensor for me.

He, writing to Sigma and ask them to switch mount...

Oly sensors are small, indeed. But now that I have given up on noise they are tempting. You can also mount nearly any lens to those little things. The coming Olympus 510 looks very cool.

[Jonas B]

Hi Jonas. I probably don't belong in this discussion (still just have the lowly DS ).

Code:

Lol. Hi Mark, of course you do.

But from Sean's latest test, it looks like all the lenses are consistent at a given aperture. So, maybe a good 'compensator' strategy for K / M lenses in M mode would be:

• Always meter (green button) at F8 (or whatever F you like that works for all your lenses).
• Add whatever fixed exposure offset works for you based on Sean's curve or your own.
• Compensate again if needed for the scene & how you metered it.
• Adjust F#, shutter speed, iso as desired to keep the exposure matched. (e.g. F8 to F5.6 ---> cut ISO or increase shutter speed by factor of 2 etc)

Actually, I use the DS this way a lot... manually resetting the TV & AV in proportion as desired based on the initial meter reading (which I admit isn't consistently at F*).

Code:

Really? Don't you fnd this...well, tedious?
I could just as well by an incident light meter and use that,
just as I used to do. Then I could use the inbuilt meter for
hasty snaps only, together with bracketing.

I want to be able concentrate on the light and the composition.
The light meters worked fine back in the 70's and I don't understand
why they can't make it work now.

To further expose my ignorance, is this an issue with both spot & center weight metering or just one?

Code:

Both.

kindest regards,

[jeffkrol]

I want to be able concentrate on the light and the composition.
The light meters worked fine back in the 70's and I don't understand
why they can't make it work now. Kindest regards,

That one baffels me as well. Some think that the meters didn't work that good and most errors were covered up by processing. Shooting mostly slides at that time, I do remember a lot of meter failures w/ my Superprogram. Well actually my failures
Could be that the digital train is just more sensitive than film and slides, causing minor variations in the meter to be exaggerated. I could probably buy this theory.
Still want Pentax to build a "meter baseline adjust" into the firmware and menu..... won't help w/ 1 lens variation, but may help w/ lens to lens variation.

[Jonas B]

That one baffels me as well. Some think that the meters didn't work that good and most errors were covered up by processing. Shooting mostly slides at that time, I do remember a lot of meter failures w/ my Superprogram. Well actually my failures
Could be that the digital train is just more sensitive than film and slides, causing minor variations in the meter to be exaggerated. I could probably buy this theory.
Still want Pentax to build a "meter baseline adjust" into the firmware and menu..... won't help w/ 1 lens variation, but may help w/ lens to lens variation.

My last fSLR camera is the Olympus OM-4Ti. When I started to use it I had to learn the buttons and the odd placement of the shutter dial and... Anyway; I took a few test films (cheap ones) trying different things and among them a series of exposure controls with the three different lenses I have. All are spot on (slides). It was the same with the beaten up Leica M5 (two lenses only with that one) I had for a while until I realized I didn't like rangefinders. I also don't remember any strange jumps in the range of 1.5 EV with the older Canon F-1. Now there are computers inside the cameras but we can't get the light metered?

Yes, a metering baseline, free for the user to dial in, remembered by the camera for several lense - that would be a dream. If they also can add a function in M mode where the shutter speed changes with the momentary light (meaning you only need to read the light once until you change the aperture setting). Free simple stuff, just firmware changes/addons.

regards,