Reading mode: Light Dark

5DMK2 vs K20D high ISO comparison

Go to last post
B

Brad Hollister

Well-known member
Messages
127
Reaction score
0
Location
titletown, US
A previous thread in this forum by another person who owns both these cameras inspired me to do the high ISO comparison in as fair a manner as I can manage (in my living room). I've had the Canon a couple weeks now, and it is love hate. Love the high ISO IQ, appreciate the auto focus upgrade, but really, really miss the K20D ergonomics. The K20D is a joy to use, but I haven't been able to get by in my most common photo settings without flash. And, I dislike flash.

5DMK2, 50mm F1.2L vs K20D 31mm Limited.

Test Method: Each camera shot in RAW at ISO 1600, 3200, 6400; identical exposure settings for both cameras. Center focus point focused on the UPC symbol of the Sooby Doo ball for all photos (real lab equipment, eh?)

ISO 1600, 1/50sec, F2.0
ISO 3200, 1/80sec, F2.0
ISO 6400, 1/160sec, F2.0

All RAW files processed in Lightroom2. Same white balance eye dropper point (grey lego brick). Same RAW processing settings including NO exposure compensation, identical sharpening and NO noise reduction, etc, etc.

My opinion of results: Canon ISO 6400 is roughly 1.67 stops better than the K20D for high ISO noise (Canon ISO 6400 = Pentax ISO 2000). No surprise there, what did surprise me is that for the same ISO setting the 5DMK2 sensor is approximately 1/3 stop more sensitive than the K20D. To get the same histogram when using the same exposure settings, same ISO, and shooting the same scene, the K20D file needs to be pushed in Lightroom by 1/3 stop. Because of the ISO sensitivity difference, I'd rate the 5DMK2 a full 2 stops better in high ISO performance.

Don't take my word for it, check out the full size jpg files here:

5D2 ISO1600



5D2 ISO3200



5D2 ISO6400



K20D ISO1600



K20D ISO3200



K20D ISO6400

 
 
The only question I have if whether it is supposed to encourage people to use Pentax or discourage from it. I think k20D holds pretty well against mighty 5D2, and with 31mm Limited on it especially.

-
 
Should be obvious the 5D MKII is clear winner starting from 3200 when talking ISO performance.

And in terms of resolution!

And DOF is shallower on a FF camera too.

Congratulations on your purchase :)

--
Thanks, Ilya.
 
White balance seems off on the k20 Brad?
--
Regards Dean - Capturing Creation
 
Azures said:
The only question I have if whether it is supposed to encourage
people to use Pentax or discourage from it. I think k20D holds pretty
well against mighty 5D2, and with 31mm Limited on it especially.

-
Click to expand...
Thank you. K20D is no slouch, especially considering price difference. It boils down to what you are photographing and how much light you have. For my type of photography, the extra 2 stops in the 5d2 are worth it. I'm shooting people in low light, SR is little help because of motion blur. Answer will be different for someone else, especially if you never go above ISO 800.

I wish Pentax would go full frame, but I now think it won't be happening anytime soon.
 
ozdean said:
White balance seems off on the k20 Brad?
--
Regards Dean - Capturing Creation
Click to expand...
I guess a lego brick is no substitute for a grey card!

I tried to keep all things as equal as I could, so I used the white balance eye drop tool on the same grey lego brick for all 6 photos.

I didn't want someone accusing me of manipulating white balance to somehow sway results (not that I think that is possible, but someone else might)

I agree, white balance needs correction. Although not the worst for compact flourescent lighting.
 
Brad Hollister said:
Thank you. K20D is no slouch, especially considering price
difference. It boils down to what you are photographing and how much
light you have. For my type of photography, the extra 2 stops in the
5d2 are worth it. I'm shooting people in low light, SR is little
help because of motion blur. Answer will be different for someone
else, especially if you never go above ISO 800.

I wish Pentax would go full frame, but I now think it won't be
happening anytime soon.
Click to expand...
--

Thanks for your test images. One thing to also consider is that the Canon overstate their ISO values according to doxmark.com. I can't do a screen capture of the stated ISO vs. measured ISO on doxmark.com. For example, at anything above ISO1600, the 5DMkII is actually 1/2 stop slower than the K20D. As a result, your test images of your 5DMkII is actually 1/2 stop slowers with respect to the K20D and the 2 stops advantage may actually be 1.5 stops. Thanks again for the test.
 
leeak said:
--

Thanks for your test images. One thing to also consider though the
5DMkII overstate their ISO values according to doxmark.com. I can't
do a screen capture of the stated ISO vs. measured ISO. For example,
at anything above ISO1600, the 5DMkII is actually 1/2 stop slower
than the K20D. As a result, your test images of your 5DMkII is
actually 1/2 stop slowers with respect to the K20D and the 2 stops
advantage may actually be 1.5 stops. Thanks again for the test.
Click to expand...
You are welcome, it is something I wanted to do as part of making a decision on which camera to sell.

Interesting what doxmark.com said about ISO sensitivity, my results are the exact opposite. Look how much darker the k20d images are. If anything canon understates ISO.
 
Actually I appreciate this kind of tests, but please consider that you have been testing a camera, the Canon 5DMkII, which is three times more expensive than the other (the Pentax K20).

2stops advantage is quite a lot, but we are tlking about sensors with very different pixel pitch. I'd like to see a last generatio 10 Mpx APS cmos sensor...you'd give away a bit of resolution but probably could get the same high ISO IQ...

--

Mr. President, I'm not saying we wouldn't get our hair mussed. But I do say... no more than ten to twenty million killed, tops. Uh... depended on the breaks.
 
It's a FF vs. a APS-C sensor: twice the surface area, so the 5D had to be better than the K20D.

Also, Pentax has an off-hands approach to noise, unlike Canon, which uses some NR on its RAW files (before in-camera processing) from ISO 800, if I remember correctly.

Add Canon's DIGIC IV processor and very efficient on-sensor noise reduction hardware and there you have it: a DSLR with very little noise up to ISO 6400.

A friend of mine has the 5DMkII and I agree on the high ISO IQ: it's not a D3, but it's not far behind.

The Canon ergonomics and functionalities are realy weak, though. I'd rather buy a D700 just because of that. But since the 5DMkII has an HD video mode and I'm into indie film making, maybe I should... sigh!

Pentax needs a FF... if only for those who really want it.

--
Once you've mastered the technique and the equipment, you can concentrate on
the more important aspects of photography: originality, atmosphere, emotion
and — ultimately — soul.
— Jeff
 
Brad Hollister said:
I tried to keep all things as equal as I could, so I used the white
balance eye drop tool on the same grey lego brick for all 6 photos.
Click to expand...
Then, you could have kept DoF (and diffraction effects) constant, by applying the crop factor to the focal AND f-ratio (so as to get same actual aperture in mm)
APS-C : 31mm @ f/1.8
FF: 46.5mm @ f/2.7 - or 50mm @ f/2.9 with a slightly different FoV

then the images would really be equivalent, yielding probably very similar results in this test.

And then you see the advantage of FF : the FF lens allows shallower DoF and thus more light. The sensor itself doesnt have much to do with actually gathering light.

But you also see that the "high ISO" advantage of F.F. is only valid when using that shallower DoF, the APS-C is just as good with low light there when you need higher DoF (e.g. need several people to be in focus, etc...)

--
Samusan
 
Excellent comparison. The 5dMk2 is certainly a very capable camera.

The only quibble I have is that you are assuming that f2 on the two lenses is the same. They may well be, but I've found that I get brighter or darker pictures doing tests of the same scene using different lenses at the same aperture, FWIW.
 
That doesn't explain why other APS-C DSLRs have better high ISO performance.

And it's debatable if Pentax's high ISO details are actual details or just noise texture.

I like my K20D but ISO 3200 is for poststamp-sized prints only and 6400 is a bad joke.

--
http://www.flickr.com/photos/denis_volk
 
tigrebleu said:
It's a FF vs. a APS-C sensor: twice the surface area, so the 5D had
to be better than the K20D.
Click to expand...
Although the 5DM2 has about 40% more photosites, too.
tigrebleu said:
Also, Pentax has an off-hands approach to noise, unlike Canon, which
uses some NR on its RAW files (before in-camera processing) from ISO
800, if I remember correctly.
Click to expand...
I'm not so sure that Pentax is the one that still has the hands off approach to noise, at least any more. Starting with the K20D there were some pretty definite signs of a little bit of noise reduction to the raw data files for ISO 1600 and above, and this may be true for the K200D for ISO 800 and above - I haven't finished checking yet. If it is true for the K200D, I suspect it may be true for the K2000/K-m as well.

I've never seen any signs of noise suppression for raw data in Canon raw files, to the extent that they don't even offset the data so black is zero.

I think we have to be men enough to admit that Canon truly has excellent high ISO results with their multi generation CMOS sensor designs. It's just a pity that their low ISO dynamic range is let down by an analog to digital converter chain that is much noisier than it should be as compared to the competition (Nikon), in spite of having a 14-bit depth.

Regards, GordonBGood
 
GordonBGood said:
tigrebleu said:
It's a FF vs. a APS-C sensor: twice the surface area, so the 5D had
to be better than the K20D.
Click to expand...
Although the 5DM2 has about 40% more photosites, too.
Click to expand...
Sure but that has no real impact when looking at the results at a normalized size (10x15 print, or screen display with proper downsampling algorithm).

Comparing sensors qualtiy between FF at f/2.0 and APS-C at same f-ratio but smaller focal is like comparing say, K20D at ISO 1600 to D300 at ISO 711. So, yeah, it's not really equivalent, the outcome should be easy to guess, no ?

--
Samusan
 
GordonBGood said:
tigrebleu said:
It's a FF vs. a APS-C sensor: twice the surface area, so the 5D had
to be better than the K20D.
Click to expand...
Although the 5DM2 has about 40% more photosites, too.
GordonBGood said:
Also, Pentax has an off-hands approach to noise, unlike Canon, which
uses some NR on its RAW files (before in-camera processing) from ISO
800, if I remember correctly.
Click to expand...
I'm not so sure that Pentax is the one that still has the hands off
approach to noise, at least any more. Starting with the K20D there
were some pretty definite signs of a little bit of noise reduction to
the raw data files for ISO 1600 and above, and this may be true for
the K200D for ISO 800 and above - I haven't finished checking yet.
If it is true for the K200D, I suspect it may be true for the
K2000/K-m as well.

I've never seen any signs of noise suppression for raw data in Canon
raw files, to the extent that they don't even offset the data so
black is zero.

I think we have to be men enough to admit that Canon truly has
excellent high ISO results with their multi generation CMOS sensor
designs. It's just a pity that their low ISO dynamic range is let
down by an analog to digital converter chain that is much noisier
than it should be as compared to the competition (Nikon), in spite of
having a 14-bit depth.

Regards, GordonBGood
Click to expand...
GG,

Hey! can you answer this question on the capabilities of pulling out details and accurate color from underexposed areas of a shot.

I am interested in managing high dynamic range images, and would like to know the advantages of cameras like the ND700 or C5DMKII, in this regard. Low light performance is self explanatory.

In a high dynamic range image, If I were to expose, say the sky properly and leave the landscape in shadow (underexpose), how many MORE stops of underexposure can be retrieved later during post processing in a FF camera vs K20D. You pick a suitable print size for comparision, and the quality of the shadow details and color be par with say using a high quality ND filter of the same stop.

Hope this question even makes sense :(
--

K10D, Pentax AF-540 Flash, Sigma 10-20, Pentax DA 21, Pentax FA 43, Sigma 100-300 F4, Sigma 1.4X TC, Tamron 18-250.

May modify my images for posting as a reply to my posts and crticisms are truly welcome.

'Life is not a journey to the grave with the intention of arriving safely in a pretty and well-preserved body, but rather to skid in broadside, thoroughly used up, totally worn out, and loudly proclaiming... 'Wow! What a ride!'

 
nigbat said:
Excellent comparison. The 5dMk2 is certainly a very capable camera.

The only quibble I have is that you are assuming that f2 on the two
lenses is the same. They may well be, but I've found that I get
brighter or darker pictures doing tests of the same scene using
different lenses at the same aperture, FWIW.
Click to expand...
Good point. It's daylight here now so I've lost my controlled conditions. Maybe tonight I'll try other lens combinations to see if the ISO sensitivity difference has another explanation.

Brad
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top