odd observation, K20D vs canon 30D

Gooshin · 08-26-2008, 09:24 PM

so me and my buddy went out for coffee this night,

we sat down for a smoke at a pretty dark fountain/park area and just started comparing shots (cuz were leet nerds like that)

i dont know if this has something to do with noise reduction, in camera jpeg setup or what, but given identical ISO settings and aperture values, his camera delivers brighter pictures, particulary in the shadow areas, at FASTER shutter speeds than mine

this is compounded as i increase ISO, shutter speed goes up but the shadow areas get darker and darker.

.....as soon as my parents are back from their trip ( i lent them my K100D) i will run some stress tests between the two cameras, but so far i am becoming less and less pleased with the high ISO performance in the dark of the K20D compared to K100D, let alone the Canon 30D.

perhaps use the K100D as a night time camera...

on the other hand, daylight ISO 100-200 and Flash photos from the K20D are absolutley stellar and bring me much delight.

i will work harder at making my night-time technique better, but so far something seems fishy, and i cant quite put my finger on it.

Gooshin · 08-26-2008, 09:58 PM

here are some photos to illustrate my case

altho i will point out that overall sharpness and detail does not decrease!

ISO 1600 F2.0 1/25

followed briefly by

ISO 6400 F2.0 1/100

second set

ISO 400 F2.0 1/6 (!!!!)

followed briefly by

ISO 6400 F2.0 1/125

Gooshin · 08-26-2008, 10:03 PM

i am drawing a conclusion here that "high iso" does not mean that you should shoot at night or in conditions where light simply does not exist!

in fact as the above ISO 400 picture shows, aside from the ridiculously slow speed the dynamic range is significantly higher than the higher ISO outputs.

rather, reserve high ISO for cases where there IS an abudance of ambiet light, but its simply not as intense, and you need a faster shutter speed to complete a shot.

an example of this are the following two shots, ISO 2000 and ISO 1600 respectivly.

pretty much a non existence of any intrusive noise, and a good (percieved) dynamic range for total composition.

*k100d* · 08-26-2008, 10:23 PM

there have been some arguments on DPR about the accuracy of the ISO ratings on the K20D. perhaps you might want to check out those discussions. there doesn't seem to be a firm conclusion
k20d ISO: overoptimistic ?: Pentax SLR Talk Forum: Digital Photography Review

*Clicker* · 08-26-2008, 10:25 PM

Better not compare it to the higher model Canons (5D and up) and Nikons then LOL

stevebrot · 08-26-2008, 10:54 PM

Out of curiosity, are you shooting JPEG or RAW? Could it be that the response curve is not strictly linear as the ISO increases and that the in-camera JPEG converter is shifting the curve down. (Different response curves for high-ISO media is a "feature" for film with th slower films tending to retain greater shadow detail.)

Steve

cpopham · 08-26-2008, 11:17 PM

I don't own a K20, but I recall reading someone saying that if you have the Extended Dynamic Range setting on, all it does with RAW files is underexposes by a stop. I don't know if this affects manual mode as well.

Also, ISO 1600 like that at night?! Holy crap I think I'd better start saving for a K20

Gooshin · 08-27-2008, 03:24 AM

shooting in RAW

"increased dynamic range" turned off

Lowell Goudge · 08-27-2008, 04:53 AM

Why not do a simple test or two.

Specifically using my favorite subject for tests a uniformly lit block wall.

I would suggest 2 slight variants on the test.

first, using Auto exposure, and the lens set to the same F stop for all shots, let the camera set shutter speed with you changing ISO..

Plot the grey scale vs ISO to see if it is consistent. This will tell you a lot.

then, to test dynamic range at High ISO, (contrast set to enutral) take a meter reading with the camera set to the middle of the apature range. Lock the shutter speed and ISO, and then shoot at every apature. Measure the grey scale again.

I have found that normally the grey scale value changes by about 45 points (between 25 and 235) per f stop. If yoou find that it changes by more then this will demonstrate that High ISO has less dynamic range. To compensate, reduce teh contrast when shooting at High ISO.

Gooshin · 08-27-2008, 04:57 AM

Originally Posted by Lowell Goudge

To compensate, reduce teh contrast when shooting at High ISO.

is this relevant for RAW shooting though?

i will try the wall test, i just need to figure out how i can get a uniformly lit wall... hmmm.

also "plot the gray area", what software do i use to accomplish such a thing?