p.5 #1 · 5D3, 6D and D600 true RAW (dcraw) comparison
mttran wrote:
You can extract milk from a glass of chocolate milk but surely it 's not a whole milk, what we want is no chocolate mixed milk to begin with.
I'm sorry, but I'm not very good at solving riddles. I'd like to know what you mean though.
A wild guess: RAW performance (milk) can not be improved by a firmware update?
p.5 #2 · 5D3, 6D and D600 true RAW (dcraw) comparison
gdanmitchell wrote:
Just a quick comment, probably to multiple posters, so don't take it personally.
The notion that comparisons must somehow neutralize the cameras to some default state makes some logical sense but little photographic sense. No one simply leaves images as they come from the camera - we tweak, adjust, compensate, and so forth to pull our ideal best image out of what the camera captures.
I've always felt that the best real world test would be one in which a photographer takes images from all of the test cameras and then uses a workflow optimized for each camera to produce the best possible prints for comparison. This reflects that way that photographers actually use cameras to make real photographs. To make it more interesting, you might want to have, say, ten skillful photographers try to produce the very best images that they know how to produce using the same approach and then compare.
To shoot all cameras in exactly the same way and the leave all files in the their default state tells us as much about the actual potential performance of the cameras as putting drivers in three cars and making sure that they always push down the accelerator peddle to exactly 50% of its range of travel tells us about automobile performance. ;-)
For an expressive form that has beautiful and compelling photographs as its ultimate goal and which measure success by the effectiveness of those photographs and the degree to which they move us and communicate a personal vision of the photographer, we sometimes seem to focus excessively and almost obsessively on tiny, insignificant technical differences in hardware that matter very, very little. They are not completely insignificant, just way less significant that our attention to them seems to suggest.
I know those scientific data never lie to us but in real life the final result is composed by many many aspects of a camera and the way how a user to direct it.
p.5 #3 · 5D3, 6D and D600 true RAW (dcraw) comparison
Rickuz wrote:
I'm sorry, but I'm not very good at solving riddles. I'd like to know what you mean though.
A wild guess: RAW performance (milk) can not be improved by a firmware update?
Sorry for the necro- this thread popped up in a Flikr discussion.
I've been pretty certain that the sensor in the 6D is almost exactly the same as what Canon put in the 5DIII and 1DX; they differ primarily in number of pixels and corresponding pixel size, which will make some difference when the RAWs are scrutinized from various perspectives.
Still, it's nice to see them compare well to Sony's newest sensors. If I hadn't originally invested in Canon, I'd probably be shooting a D600 instead of a 6D right now, but I'm glad that my 6D isn't that far off.
RE: Firmware- Firmware can absolutely make a difference in the quality of RAW output. It's the only thing that can! After the firmware (that being the camera's operating system and processing software) gets done with the sensor data and writes it to storage we're limited to 'destructive' editing in post.
p.5 #4 · 5D3, 6D and D600 true RAW (dcraw) comparison
D. Diggler wrote:
I like to see both but non-normalized is more important to me.
normalized is more important to me and basically all that is important when I am comparing between bodies
when it comes to imagining how a body might do at full reach though it might be nice to see non-normalized data to get an idea how it would work when take advantage of full extra MP count, but when it comes to comparing bodies I don't see how it makes sense to go by anything other than normalized values
p.5 #5 · 5D3, 6D and D600 true RAW (dcraw) comparison
And then when they update ACR to ideally support the camera in two months all of that hard fought 'real' world comparing you did is tossed in the bucket. Plus same settings for NR and so on don't do the same things for each camera, or even the same camera using different software versions, and it can hard to tell whether it a camera difference or just how they tweaked the software settings, you can try to balance for apparent visible detail, but it can be tricky. WHo want to take so much time and effort trying to sort that all out when a controlled DR test can be done in like 60 seconds? SNR tests are trickier, but those with the right setup and can do them quickly and easily too. WHy waste time on that stuff? Save the real shooting for real shooting.
It's so much easier under controlled conditions and with simple tests like black frame read noise (although a few real world tests to try to see how bothersome banding looks in detailed scenes might not hurt, as tricky as those are to carry out), so may variables, a slight upper haze dims the sun just a trace, etc. hard to keep things consistent
gdanmitchell wrote:
Just a quick comment, probably to multiple posters, so don't take it personally.
The notion that comparisons must somehow neutralize the cameras to some default state makes some logical sense but little photographic sense. No one simply leaves images as they come from the camera - we tweak, adjust, compensate, and so forth to pull our ideal best image out of what the camera captures.
I've always felt that the best real world test would be one in which a photographer takes images from all of the test cameras and then uses a workflow optimized for each camera to produce the best possible prints for comparison. This reflects that way that photographers actually use cameras to make real photographs. To make it more interesting, you might want to have, say, ten skillful photographers try to produce the very best images that they know how to produce using the same approach and then compare.
To shoot all cameras in exactly the same way and the leave all files in the their default state tells us as much about the actual potential performance of the cameras as putting drivers in three cars and making sure that they always push down the accelerator peddle to exactly 50% of its range of travel tells us about automobile performance. ;-)
For an expressive form that has beautiful and compelling photographs as its ultimate goal and which measure success by the effectiveness of those photographs and the degree to which they move us and communicate a personal vision of the photographer, we sometimes seem to focus excessively and almost obsessively on tiny, insignificant technical differences in hardware that matter very, very little. They are not completely insignificant, just way less significant that our attention to them seems to suggest.
p.5 #6 · 5D3, 6D and D600 true RAW (dcraw) comparison
My blacks on the 5D3 never look that noisey at iso6400. Have a hard time believing this comparison just based on that... All are great cameras though and you should be happy with whatever you have.
p.5 #7 · 5D3, 6D and D600 true RAW (dcraw) comparison
The funny part about all these testing methodology debates is that-- everyone is right.
There's value in controlled, non-edited (or limited editing) comparisons.
There's value in real-world, heavily edited, final output style comparisons.
There's value in non-normalized-for-resolution comparisons.
There's value in normalized-for-resolution comparisons.
And there are weaknesses in all of the above too. They each help us figure something out for different end-use scenarios. At the same time, those doing the testing, often for free (or without much web-traffic-income) have finite time and resources. Otherwise we'd also want tests specifically for portrait lighting/FLs/subjects, landscapes, green-vs-red channel dominance, mixed lighting, and on and on.
Thanks to the OP-- for a useful organization of information. One more piece in the puzzle.
