The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
DX0 ratings . . . Didn't the A55 do well
- Thread starter jonoslack
- Start date
C
curious80
Guest
Well I have had my views change a couple of times about how useful are the DxoMark results. I have gone from being a firm believer to not believing at all to believing them somewhat again. However I definitely owe this to DxoMark that it made me question the supposedly great improvements in APS-C sensors in terms of high-iso performance.
DxoMark suggests similar high-ISO SNR performance for most good APS-C sensors from last 4-5 years (before D7000 and K5). Whereas typical belief is that APS-C sensors have improved by one or two stops in last few years. To test that out I downloaded raw files from Imaging-Resource for bunch of APS-C sensors from past few years and passed them through Lightroom 3.2 with noise reduction and sharpening turned to 0. And as far as I could see I didn't notice any real improvement over the years. For example here is a comparison @ iso 1600 between 30D and the supposed high-iso champion NEX5. Here I have upscaled 30D to 14MP to match NEX5 and then took 100% crops. Despite being upscaled, the old 30D seems to be atleast as good if not better than the NEX5:
NEX5:
30D:
DxoMark suggests similar high-ISO SNR performance for most good APS-C sensors from last 4-5 years (before D7000 and K5). Whereas typical belief is that APS-C sensors have improved by one or two stops in last few years. To test that out I downloaded raw files from Imaging-Resource for bunch of APS-C sensors from past few years and passed them through Lightroom 3.2 with noise reduction and sharpening turned to 0. And as far as I could see I didn't notice any real improvement over the years. For example here is a comparison @ iso 1600 between 30D and the supposed high-iso champion NEX5. Here I have upscaled 30D to 14MP to match NEX5 and then took 100% crops. Despite being upscaled, the old 30D seems to be atleast as good if not better than the NEX5:
NEX5:
30D:
C
curious80
Guest
If you replied to my post then I don't really understand your comment
ptomsu
Workshop Member
Was not meant as reply to your post, but as my general statement what I think about DXO testsIf you replied to my post then I don't really understand your comment
barjohn
New member
The 30D test above is a bit misleading for several reasons. Merely setting LR to 0 for NR etc, does not mean that you are getting no NR. Canon is known to do some in camera NR even in RAW processing. Look closely at the 30D image and you can see the soft smearing of fine detail to reduce noise. It is especially visible in the label on the thread and the lower pink fabric with the splatter pattern.
C
curious80
Guest
I would argue that 30D actually has more detail in at least some parts of the image. For example with the fabric with read leaves, the details in NEX image have been smeared due to noise whereas 30D manages to preserve more detail. You also have to keep in mind that 30D has significantly less resolution - it has only 8MP whereas NEX has almost twice the pixels at 14MP. 30D image in this example had been significantly upscaled which is one cause of the softness. And it is amazing that despite that extreme upsizing it can deliver better detail than NEX for some parts of the image.
Anyways my point was not to prove that 30D sensor was superior to NEX. It would be a very broad generalization to make based on just a single crop from a single image. Plus despite these being studio images with consistent exposure and lighting etc they are not identical images - for example I think the plane of focus is slightly different in both images which obviously impacts details and so on. The main point was just to argue that these two sensors seem to be roughly in the same league instead of the general belief that the newer aps-c sensors have one or two stops high-iso advantage over old sensors. DxoMark correctly predicts this which is why I find DxoMark as a useful resource even though I don't agree to everything which they say.
Anyways my point was not to prove that 30D sensor was superior to NEX. It would be a very broad generalization to make based on just a single crop from a single image. Plus despite these being studio images with consistent exposure and lighting etc they are not identical images - for example I think the plane of focus is slightly different in both images which obviously impacts details and so on. The main point was just to argue that these two sensors seem to be roughly in the same league instead of the general belief that the newer aps-c sensors have one or two stops high-iso advantage over old sensors. DxoMark correctly predicts this which is why I find DxoMark as a useful resource even though I don't agree to everything which they say.
A
Andrea Buso
Guest
It may also have something to do with the lenses used on the respective cameras.