Now what am I doing wrong?

Guys, not sure if the K10/K20 operate differently to the K-x, but I have absolutely no problems getting accurate metering with any manual lens on my K-x, based on the aperture chosen. In manual mode on the K-x, set the chosen aperture, then press the Exposure Compensation button (not the green button) and the camera meters for the chosen aperture. It is accurate every time. I use this with Pentax-M lenses and also M42 lenses with adaptor.

Does this not work on earlier Pentax DSLR's?

Regards
David

Definitely on the K10D, it does not, with K and M lenses, or A lenses not set to A.

I have no direct experimental experience of the K20D, though I've read that it wasn't fixed, nor K-x K-r, or K-7. I haven't tested it yet on the K-5.

If you meter with the lens wide open, you'll get the right result just by calculating the exposure yourself, applying the reciprocity rule, for the aperture you choose. It's all the camera does in A mode, after all.

It's perfectly possible to use K and M lenses successfully, just not all that convenient for auto-exposure.

There's even an upside - K and M lenses are *very* much cheaper than their A series counterparts

There's a lot of nonsense being posted in this thread about people not having ANY exposure problems with lenses on Pentax cameras when it's well documented that there is

This thread link is one of many showing graphs etc. explaining the problems.

No two lenses meter the same either you have to carry out tests and decide on what aperture to meter on and what compensation to give.

TIP: Use an aperture that doesn't require compensation often f/5.6

Don't meter wide open on any lens below f/2.4 your pics will be far too dark. Again it's well documented..... try Giggle!

Is it also so documented on the K-x?

If not, there isn't a lot of nonsense at all, there's one assertion, followed by a retraction, so fair play to Frogherder on that.

Even the film cameras didn't always meter correctly with all lenses

The link I gave above was from 2009.... This is more recent with a graph of a K-7 as well. link

I use an Expodisk and set the WB off it 1st using the sky then a shot at each aperture using Av if possible, Manual (Green Button) if not.

I then read the RGB value for each shot and compare it to my expected value of 117R 117G 117B.

I then estimate the correction needed say +0.7 and do another test to make sure that I get around 117R 117G 117B

This is the rough result from such a test:

==================================================
K-5
Pentax_A 50mm f/1.7

WB Custom via Expodisc

Av Aperture set from Dial

f/1.7 114 112 126
f/2.0 108 106 120
f/2.8 099 097 110
f/4.0 105 101 115
f/5.6 102 100 113
f/8.0 102 100 113
f/11 105 103 106

WB Set again from Expodisc

f/11 106 106 108
f/4.0 105 105 107

Seems OK from f/2.0
All slightly too dark
===========================================

It would look at 1st glance that f/1.7 is correct, but
all other apertures are too dark, so knowing that f/1.7
should be out anyway I would apply a compensation factor
and use f/4 to f/11 for metering.

This is a Pentax-A lens which is quite well behaved.
Pentax-M lenses can be quite erratic.

This was a Pentax-K lens.
Note: WB out and results are not really usable
=========================================================
K-5
Pentax_K 28mm f/3.5

WB Auto via Expodisc
Seems way out time 6:30pm
possibly too late in the day

Aperture set from lens

f/3.5 116 128 152
f/4.0 116 128 150
f/5.6 129 141 163 Consistent from f/5.6
f/8.0 133 144 166
f/11 132 143 165
f/16 134 145 165
f/22 132 143 161


Seems OK from f/5.6
All slightly too light

Redone setting WB via Custom
Note: shots were done backwards f/5.6 to f/3.5

f/3.5 130 130 132
f/4.0 122 122 124
f/5.6 145 145 145
=====================================================

I eventually settled on using f/8 and -0.7 compensation:

GREY BRACKETED TEST
====================
AWB? showed right on camera

-->>> CORRECT F/8 -0.7 ManExp 116R 116G 116B
==============================================


A Sigma 24mm f/2.8 Test gave:

==============================================
GREY EACH STOP TEST
====================
AWB? showed WRONG
Custom WB hard to set. Non of the shots were right for bright sunlight.

f/2.8 120 118 139 Heavy vignette
f/4 165 163 184 Still vignettes to 126 in corner
f/5.6 169 169 193 Corner OK 154 ::::: Try to meter at f/5.6
f/8 180 178 199
f/11 178 176 197
f/16 186 183 202
f/22 186 182 199
=======================================================

So you can see there is a massive variation Lens to Lens
and it takes quite a bit of time to test each lens

These were just pastes of early tests. I'm now more
careful setting the WB correctly, because I need the
RGB values to be very close to each other.

But that thread (near top of P3) asserts that the problem is resolved in the K-m. There are clearly a lot of similarities between the K-m and the K-x (bar the sensor) so perhaps the fix does apply to the K-x ...

Well fair enough.

But for most people's purposes, the requirement is that, for any given lens, stop-down metering should give the same shutter speeds as wide-open metering, followed by applying standard reciprocity calcs to get the required exposure at the selected aperture.

With the K10D it doesn't, with a bigger error as the smaller apertures are approached.

I don't know about the post-K20D DSLRs, but I'll be testing it out with my K-5.

Sorry Algi, but I totally disagree with you where the K-x is concerned. The M lenses (and any other manual lens for that matter) meter just as accurately as any automatic lens on this camera. I've taken thousands of shots with both manual and automatic lenses and there is no difference in the accuracy of exposure metering with either. The fact that there are variations between different lenses is irrelevant - this will always be the case whether the lens is manual or not.

BTW, most of the images I have taken of late have been taken with old manual lenses. i don't recall having a single instance of exposure miscalculation by the K-x, unless it is me that has made the error.

If anyone out there has a K-x, you can use Manual lenses with confidence and grab some extreme bargains for what are usually fabulous lenses optically for very little money.

Regards
David

PS to my previous comment: The K-x does not require the 'green button' to meter with a manual lens. It use the exposure compensation button when in manual mode to accurately meter the correct exposure. Works with all my old manual lenses that don't have the 'A' setting.

Regards
David

The above is basically how it is described in the K-m manual.

I can only assume that although a solution was developed by the K-m/K-x development team at Pentax, the subsequent K7, K-r, K5 cameras all came from the K20D line with respect to firmware, and thus never benefited from this fix. A real shame.

They use the green button, but otherwise work in exactly the same way. Only reason the kx uses the exposure compensation button is there's no room for a green button! A far as I know the other post-k20d cameras also got improved stop down metering.

Hi David, you might well get a lot of acceptable pictures, just as I would get a lot of acceptable pictures if I used a constant f/8 at 1/200th but light transmission of lenses does vary... it's a fact you can't get away from A multi-coated lens will always transmit more light than a single coated lens. A lens that vignettes very heavily also transmits less light and no camera metering will know about that.

Even lenses of the same focal length from the same maker and from the same era/series don't transmit the same amount of light.

This is the difference between the Pentax-A Series 50mm's f/1.7 and f/1.4 at f/4 using Av and Expodisc they should be the same, but there's almost half a stop difference in the final result and compensation is needed to balance them out. I cut and pasted the crop from one onto the other. Sorry I can't remember which was which (done sometime ago on a K20D), but my final conclusion was that the f/1.4 was better for poor light. One lens needed a + correction the other -


-




I've just tested a SMC Takumar 35mm f/3.5 and found that it meters correctly with a K-5 at f/8 which is great because that's the aperture I would use it with. It underexposes by half a stop wide open. OK f/8 to f/16.


-

Sorry but that's slightly wide of the mark! Either I have a very unusual K-x with a 'Green Button' that others don't or all K-x's have a 'Green Button'! The K-x has both an Exposure Compensation button and a programmable Green Button. If you have one that doesn't I would return it to the shop and ask for a refund!!

Regards
David

I'm afraid it sounds like a mistaken assumption on my part David

But the button works in exactly the same way I take it? If it's doing something else, it would be interesting to know.