Pentax 18-55 Kit lens underexposure?

[PentaxScott]

I recently jumped into DSLR photography (upgrading from a MZ-7) in September of 2007. I purchased a K10D kit, opting for the 18-55 kit lens as a basic backup to my Tamron 24-135 3.5-5.6 lens. I have been using both the kit lens and the Tamron equally depending on whether my situation called for the use of wide angle (18mm on the kit).

During some recent tests however, Ive noticed that my 18-55 kit lens seems to have a habit of underexposure. I took some test shots with my K10D in Program, spot metered, and shifting the aperture on the Tamron to match the samples on the kit lens, but allowed the camera to set the shutter. The Pentax kit lens was 1/3 stop darker on the 24 and 35mm focal length tests and a FULL stop darker on the 50 mm test.

Am I missing something? Even though the Tamron lens was originally designed for 35mm film use, I dont see why there should be this much discrepancy in the exposure. Why is the meter getting such vastly different readings? I have left the links for the 24mm and 35mm tests and displayed the 50mm tests. All shots were taken at 1600, no flash, and NO modifications (including WB corrections) were made other than to crop the image to upload on photobucket.

Kit lens @ 24mm f4.0 1/25

Tamron lens @ 24mm f4.0 1/20

Kit lens @ 35mm f4.5 1/25

Tamron lens @ 35mm f4.5 1/20


Kit lens @ 50mm f5.6 1/20


Tamron lens @ 50mm 1/10

[jeffkrol]

Originally Posted by PentaxScott

I recently jumped into DSLR photography (upgrading from a MZ-7) in September of 2007. I purchased a K10D kit, opting for the 18-55 kit lens as a basic backup to my Tamron 24-135 3.5-5.6 lens. I have been using both the kit lens and the Tamron equally depending on whether my situation called for the use of wide angle (18mm on the kit).

During some recent tests however, Ive noticed that my 18-55 kit lens seems to have a habit of underexposure. I took some test shots with my K10D in Program, spot metered, and shifting the aperture on the Tamron to match the samples on the kit lens, but allowed the camera to set the shutter. The Pentax kit lens was 1/3 stop darker on the 24 and 35mm focal length tests and a FULL stop darker on the 50 mm test.

Am I missing something? Even though the Tamron lens was originally designed for 35mm film use, I dont see why there should be this much discrepancy in the exposure. Why is the meter getting such vastly different readings? I have left the links for the 24mm and 35mm tests and displayed the 50mm tests. All shots were taken at 1600, no flash, and NO modifications (including WB corrections) were made other than to crop the image to upload on photobucket.

All sorts of possibilities but my guess is that the "kit" lens aperature mechanics are just off.
The lens meters "wide open" and the extrapolates the f stop shutter speed from there (you can see at full open the 2 are close. 5/100's of a second different which is no big deal in my book and closer to a 1/4 stop I guess from the image histogram averages).
The key in my mind is how linear the error. If ALL exposures at ALL f stops are 1 stop darker it is at least a linear error and easy to deal with. I would ignore the error at the very smallest aperature as most lenses seem to exhibit anomolies at that end.
Second possibilty is f4 is NOT f4 in the kit and therefore when the computer extrapolates up it is doing so from an incorrect starting point.
Either way the error is managable if linear.
I'm assuming the Tamron is at least an "a" lens w/ metering wide open.
Just do a series of white wall tests to calibrate your brain to the lens
Oh, and shooting w/ a monitor (or any fluctuating electronics such as flourescents) is not a good idea as that could create another error.

[fastoys2nv]

I find myself generally shooting 0.5 comp with a K100D and kit lens. Seem to get the same general results that you have post in your test shots.

[PentaxScott]

Thanks for the advice.

I originally did this test to look at IQ, and resolving power of some of the older film lenses I had when I noticed this anomaly. I should have thought at that point to arrange a composition away from the computer monitor.

I retested the shot focusing on a coca cola sitting on a news paper with a plain wall as a background. The test was done under incandescant light with no flash, again on a tripod. I noticed the 24mm (f/4.0)and 35mm (f/4.5) focal length tests still to be about 1/3 underexposed compared to the tamron and the 50mm (f/5.6) shot was still a full stop underexposed. I changed the ISO to 100 and recieved the same results, so I have to guess its not the camera meter at this point. Adjusting the ISO back to 1600 and shooting at f/16 on both lenses set to 50mm again showed the same 1 stop underexposure on the kit lens.

jeffkrol said he has noticed that the kit lens seems to be underexposing a bit as well for him. Anyone else have this issue? Do I have a bum lens, or is this common?

(Also, the Tamron in question is the SP AF 24-135mm f/3.5-5.6 AD Aspherical (IF) lens that was released about the same time as the Pentax 24-90mm lens. Its always performed perfectly in the past)

[Canada_Rockies]

Originally Posted by PentaxScott

Thanks for the advice.

I originally did this test to look at IQ, and resolving power of some of the older film lenses I had when I noticed this anomaly. I should have thought at that point to arrange a composition away from the computer monitor.

I retested the shot focusing on a coca cola sitting on a news paper with a plain wall as a background. The test was done under incandescant light with no flash, again on a tripod. I noticed the 24mm (f/4.0)and 35mm (f/4.5) focal length tests still to be about 1/3 underexposed compared to the tamron and the 50mm (f/5.6) shot was still a full stop underexposed. I changed the ISO to 100 and recieved the same results, so I have to guess its not the camera meter at this point. Adjusting the ISO back to 1600 and shooting at f/16 on both lenses set to 50mm again showed the same 1 stop underexposure on the kit lens.

jeffkrol said he has noticed that the kit lens seems to be underexposing a bit as well for him. Anyone else have this issue? Do I have a bum lens, or is this common?

(Also, the Tamron in question is the SP AF 24-135mm f/3.5-5.6 AD Aspherical (IF) lens that was released about the same time as the Pentax 24-90mm lens. Its always performed perfectly in the past)

Check the EXIF data for your images - your Tamron may be switching the meter to center weighted from matrix. If you cannot do that, repeat the tests with the camera meter set to center weighted and see if the exposures are closer to each other.

[RiceHigh]

All Pentax DSLRs I have used have erratic metering which tends to underexpose quite often but exposure will usually be brighter with film lenses, no matter the lens is Pentax original or 3rd party. In fact, different lenses could produce quite different exposure results, as you've shown us. Even worse, spot metering and centre-weighted average produce different levels of underexposure whereas the matrix metering is rather unpredictable but inaccurate.

For more information, see my Geocities homepage or my particular article on the underexposure tendency of Pentax DSLRs with digital lenses at my Blog:-

RiceHigh's Pentax Blog: Underexposure Tendency of K10D and K100D

Originally Posted by PentaxScott

I recently jumped into DSLR photography (upgrading from a MZ-7) in September of 2007. I purchased a K10D kit, opting for the 18-55 kit lens as a basic backup to my Tamron 24-135 3.5-5.6 lens. I have been using both the kit lens and the Tamron equally depending on whether my situation called for the use of wide angle (18mm on the kit).

During some recent tests however, Ive noticed that my 18-55 kit lens seems to have a habit of underexposure. I took some test shots with my K10D in Program, spot metered, and shifting the aperture on the Tamron to match the samples on the kit lens, but allowed the camera to set the shutter. The Pentax kit lens was 1/3 stop darker on the 24 and 35mm focal length tests and a FULL stop darker on the 50 mm test.

Am I missing something? Even though the Tamron lens was originally designed for 35mm film use, I dont see why there should be this much discrepancy in the exposure. Why is the meter getting such vastly different readings? I have left the links for the 24mm and 35mm tests and displayed the 50mm tests. All shots were taken at 1600, no flash, and NO modifications (including WB corrections) were made other than to crop the image to upload on photobucket.

Kit lens @ 24mm f4.0 1/25

Tamron lens @ 24mm f4.0 1/20

Kit lens @ 35mm f4.5 1/25

Tamron lens @ 35mm f4.5 1/20


Kit lens @ 50mm f5.6 1/20


Tamron lens @ 50mm 1/10

[blende8]

Originally Posted by RiceHigh

In fact, different lenses could produce quite different exposure results,

That is absolutely normal., because it depends on the transmittance.

[RiceHigh]

Originally Posted by blende8

That is absolutely normal., because it depends on the transmittance.

Not really, the f-number should be catered for this already, in the optical design. A lens should be calibrated properly for its values as rated.

Also, for metering part, it's TTL (thro the lens), isn't it?

Sorry that I must mention again that Canon EOS system is very consistent and accurate in fixed pattern meterings namely CWA, spot and partial ones. So, I'm afraid the transmittance reason is not substantiated here as this uncounterexample proves this is not valid.

A short note is on the other hand, Nikon system is not as consistent as Canon's but Nikon's the Matrix metering is the most intelligent in compensating all those discrepenacies, IMHO again.

[jeffkrol]

Originally Posted by RiceHigh

Not really, the f-number should be catered for this already, in the optical design. A lens should be calibrated properly for its values as rated.

Also, for metering part, it's TTL (thro the lens), isn't it?

TTL at base aperature, extrapolated up from there. Pentax has to deal w/ a total mechanical system from there. Canon, I believe has electronic controlled aperatures in all the NEW lenses. I suspect that they are more accurate, to a degree. I'm sure Pentax can do much better once they throw out all legacy support and go w/ a new electronic mount.

Originally Posted by RiceHigh

Sorry that I must mention again that Canon EOS system is very consistent and accurate in fixed pattern meterings namely CWA, spot and partial ones. So, I'm afraid the transmittance reason is not substantiated here as this uncounterexample proves this is not valid.

A short note is on the other hand, Nikon system is not as consistent as Canon's but Nikon's the Matrix metering is the most intelligent in compensating all those discrepenacies, IMHO again.

Time to do us a favor:
Take your 5D and do a white wall series w/ 1 lens going from largest to smallest aperature and post the MEAN of each histogram. Just do it in spot metering. You don't even need to post the images. A table of numbers will do.Then we can see how consistent it is.. Simple test, takes only minutes. Then we can go on from here......

[Jodokast96]

Spot metering? You mean that cheesy 3.5% they try to pass off as spot metering?

[blende8]

Originally Posted by RiceHigh

Not really, the f-number should be catered for this already, in the optical design. A lens should be calibrated properly for its values as rated.

Please look up how the f-number is defined.
It has nothing to do with transmittance.
An f1.2 lens does not necessarily bring more light to the sensor than an f1.4 lens.

[jeffkrol]

Originally Posted by Jodokast96

Spot metering? You mean that cheesy 3.5% they try to pass off as spot metering?

Yea, that one