DxO

Bill, I think my sensor is okay (an M8), but my brain sure does hurt after trying swallow all that technical stuff. I think I'll ignore whatever the score might be for my camera and just take some pictures. If fact, I may leave the M8 home alone and shoot a couple of rolls on an IIIf and M2 to relax my overwhelmed synapses. I little silver might calm down my over-extended gray matter.
 
I've never been a pixel-peeper or camera number cruncher. I just don't see the point. If you like to look at and play with the numbers that's fine but I think in the end most of it really has little to do with what kind of pictures one takes. If you think about it the sensor is equivalent to the film you'd put in an old fashioned (sic) camera so you still have to factor in subject, light, mood, exposure and the vision of the photographer. It takes a thoughtful combination of all these elements to produce those stunning keeper images we all yearn for. When it comes to comparing cameras I'd much rather look at real world images and listen to what other users actually say about a particular unit. Much more revealing in my opinion...
 
I think I'll go shoot some Plus-X and worry about this some other day.
 
A lot of the DXOmark rankings for sensor performance disagree with my personal experience in real world use. I've said it quite a few times on this forum now that I personally don't think you can quantify how good a sensor is with numbers.

It is interesting though, and possibly somewhat useful.
 
Last edited:
It is good to know that there is someone so much 'across' this stuff, and I've book-marked this link. But like the others, and as you predicted of us, I too am off with my 30,000(?) MP M5 to not contribute to the sensor size / MP wars.
 
Last edited:
All information that can flush out the differences between cameras is good (cost, size, weight, battery capacity and so on). These tests just adds to that and it is up to each of us to use the information.

Of course, after you buy a camera, and your new expensive little darling does not 'win' these tests you will tend to disagree with the results of the test. However remember that the price dump that came 34 min after you bought it hurts more.
 
Bill, thanks for that link. interesting article. I've found the dxo #'s and database correspond with the handful of digitals I've owned that are in the database.
 
If you enjoyed this article, you may want to read on here:
http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/

It goes into more detail and among other things explains why Nikon's compression of RAW files is effectively lossless. Should be the same for Leica's compression.

Well, it's a fascinating read but I doubt that it will make me a better photographer. Especially since I shot film only. ;)
 
Well ... that made my head spin! :eek:

All I know is ... my D700 takes nice pitchers!
 
fdigital said:
A lot of the DXOmark rankings for sensor performance disagree with my personal experience in real world use. I've said it quite a few times on this forum now that I personally don't think you can quantify how good a sensor is with numbers.
Click to expand...

Sorry, but that's nonsensical*. Of course you can. A sensor is simply an array of measuring devices with specific parameters, constructed to some set of tolerances.

And the output of a CCD or CMOS sensor is an array of numbers!

I'd agree if you said that you can't describe a sensor's properties with one number. I'd agree if you said that DxO assigns different weights to various parameters than I do. I'd agree if you said that knowing about the sensor per se doesn't necessarily tell you what the whole imaging chain will do.

But I don't think there's a scientist or an engineer in the world who deals with imaging sensors who would agree with the above statement.

In fact, without quality metrics, the engineers who design sensors would have no rational basis on which to improve sensors, and we would not have seen the rapid improvements in these devices that we have in fact seen over the last three decades.

*So... do you shoot Epson or Leica DRFs, Olympus DSLRs, or both? (I shoot Olympus DSLR, and the 4/3 community, like the Leica DRF community, is heavily enriched in people who think we can't quantify sensor performance. Of course, the obsession should be with photography, not with incremental — but eminently quantifiable — differences in sensor performance. Exactly as Bill says in the first post in this thread.)
 
Last edited:
PKR said:
If you go back to the "old days" of silver based photography, where photographers had to know a little chemistry and math
Click to expand...

Of all the many things to learn about digital, I find sensor technology the least practically useful topic for a photographer. (as opposed to color management, backup strategy etc) You just expose to the right and that's it. As far as I can see, that is all you can do.
The few times you buy a new camera, it may be good to know something about sensors, but then you need to take more factors into account, like lenses, image stabilization, weight, etc. E.g. some combinations of lenses and sensors work well together, others don't. The sensor score tells you nothing about that.
The best you can do is to try for yourself before you buy. So in the end, you don't need any numbers.
 
PKR said:
When I first encountered the problem, I asked a friend who designs camera sensors (I'm near the Si Valley and I'm lucky to know two engineer-scientists who design camera sensors). He explained the very common problem and how different designers deal with it. It's an issue of extrapolation and how many adjacent sites are polled, and where they are located in relation to the non photon receiving site - and how the poll information is used. Some do it better than others.
Click to expand...

That happens in the demosaicing (conversion) AFAIK, all sensors being Bayer types except for the Foveon. Could be something I know nothing about too. FWIW.
 
You must log in or register to reply here.
Back
Top Bottom