Canon sensors in 7D and 60D behind the curve?

DuncanM1 said:

The K-5 is capable of producing quite usable images straight from the camera at an effective ISO of 25600 or 51200, which is beyond the ability of the Canon crop sensor cameras.

Click to expand...

I'm certainly not arguing that is not the case though!

I would possibly be drawn into an argument as to the quality of the output at these ISO's though ;-)

Keith Reeder said:

There is nothing "inferior" about the Canon sensor

Click to expand...

Sorry, but the Sony/Pentax 16MP sensor has lower noise and greater dynamic range, and these are the two most crucial measures of sensor performance and will result in noticably superior IQ, especially at high ISO. Here's an example of high ISO output straight from the camera that illustrates the superiority of the Pentax output:



(effective ISO=51200)

I have looked at K-5's over the camera counter and this is the kind of output that one can expect from this camera.

no text

you know that a lot of Nikonians says Canon does it since the first Rebel (300D) ?

DuncanM1 said:

You are trying to argue that sensor performance is irrelevant to IQ...and to some extent, (which is why I still use a D50 and K10D) especially at low ISO, this is true and you can still do good work with a 60D/7D despite its inferior sensor, but that doesn't mean that there is no difference in sensor IQ. The K-5 is capable of producing quite usable images straight from the camera at an effective ISO of 25600 or 51200, which is beyond the ability of the Canon crop sensor cameras.

Click to expand...

As MANY other people have said, the DXO ratings are really not indicative of the final results that can be obtained by any camera. Here's a (terrible) shot taken at with a Canon 7D, cleaned up a bit in CS5, shot at ISO 51200 by underexposing 2 stops at ISO 12800 and pushing 2 stops when processing.





There is some banding which could be cleaned up fairly easily with something like Topaz DeNoise or Nik Dfine (which I do not own). The banding probably causes the lower DXO scores overall for Canon. Having said that there is a fair amount of detaill at 100%.

In fact there really is little difference between all recent APS-C sensors, the only thing being Canon has had the sensor in some of their cameras for over a year.

I really wish that when people spewed out nonsence about some camera, they would at least post an IMAGE to prove their point, not some semi-meaningless numbers or other peoples opinions.

Tim

DuncanM1 said:

I have looked at K-5's over the camera counter and this is the kind of output that one can expect from this camera.

Click to expand...

Sorry buddy, but I don't find that image particularly impressive. Perhaps you could show us something that we could actually compare to the Canon sensor if you wish to make a statement about how it compares ;-)

Keith Reeder said:

We're arguing - if you were to talk the time to read and understand the thread properly - that DxO's nonsense proves nothing about fanboy claims that the 7D/60D's sensor is in any way "inferior" to that of the A55/D7000/K-5

Click to expand...

You may not like the word "inferior" but the Canon sensor is behind the Sony sensor on all four DxO metrics.

And that's on the 'print' tab (that is, after resolution normalization).
On the screen tab, the results are even better.

The differences are minor for SNR and Tonal Range, where the Sony sensor has just an edge.

However, at ISOs 100-400, the Sony sensor is quite a bit better on the Dynamic Range and Color Sensitivity metrics.

In fact, the Sony sensor has a better Dynamic Range at ISOs 100-400 than the 5DII.
That's quite an achievement for a 1.5x sensor.

Many on this forum had doubts about the quality of the 7D sensor when the camera was announced last year. It now seems that there was a reason for that after all.

(Btw, so glad that the whole 'you-must-downsize-to-compare' rhetoric does not apply anymore).

Spazmedia said:

There is some banding which could be cleaned up fairly easily with something like Topaz DeNoise or Nik Dfine (which I do not own). The banding probably causes the lower DXO scores overall for Canon. Having said that there is a fair amount of detaill at 100%.

Click to expand...

Just for the hell of it I finally downloaded Topaz Denoise. Here is what I was able to obtain with a 7D at 51200 and a bit of post processing (and my skills are fairly limited!):





In both cases not great. But then again the K5 results are not great at ISO 51200 as well.

Definitely behind the curve these days - quite literally.

rhlpetrus said:

And did you really check the D3100's images comapred to the 550D's in that comparison? For example, the base ISO? Or the 6400 ISo comparison?

Click to expand...

I'm going to guess not. The folks who cast doubt on the D3100 output are mainly DxO obsessives who haven't really looked at samples/comparisons.

After doing some reading on the whole DxO process, how they weight the data, that IC RAW tweaking skews their figures, and they ignore pattern noise like banding, I'm not as interested in their findings as I used to be.