7d mkii vs 5d mkiii

All I can say is the 7DII has motion blur/out of focus or something going on here. Should not be this much difference, IMHO - at least there isn't between my two (5D3 and 7D2) but I'll go try some tests.

TeamSpeed wrote in post #18643813
I see a bit more detail with the 5D4 (and with the 5D3), but never have any real complaints about the 7D2. This is from the SL2 at 6400 from an inside aquarium in Texas, but the 7D2 is nearly identical in high ISO performance and low light. I don't find it muddy at all, and noise management is pretty good.

This is from my 7D2, in situations where I use it more often.

The 5D3 is a great camera, I used it for a couple of years. Now having the 5D4, I do find I don't often pick up the 7D2 these days, but will always recommend it to others.

I was the same and decided to pull out the 7D2 this year. After getting the 400 DO II I'm leaning back to towards the 5D4 but
I would not have any issues shooting with my 7D2 if I had to.

I was experimenting with both the 7D2 and 5D3 when the 7D2 came out. I took pictures on an overcast day exposures where the same for both. You can see the noise which I call the muddy with the 7D2 around the edges of the black molding.

The 5D3 was far cleaner. While I have no issues with high ISO and low light shots with my 7D2 as I can ETTR and correct in post but I won't even bother with it unless I have a good sunny day. FF's are more forgiving.

Personally, I would just call that noise in general (but muddy works ), and whatever processes that raw may produce different results than another processor.

The 7D2 will have about a stop worse noise than the 5D3, so if you shoot the 5D3 at a higher stop ISO, it should produce very similar results.

Here is an exact comparison between the two, using the same exposure and raw processing.

digital paradise wrote in post #18643641
Cleaner is also a good reference. In heavy overcast I don't even bother with the 7D2. I call it the muddy look.

Well, there is no such thing as a muddy sensor, or a muddy Canon RAW file. All Canon DSLRs have RAW data that is a linear measurement of light levels across the scene (plus noise), and RAW contrast depends almost completely on the lens and scene. Default conversions hit smaller pixels harder with NR because people run right to 100% pixel views to falsely gauge "IQ", when they are looking at "PQ".

If you need to magnify it more, that is not a sensor surface hunger; that's a magnification issue.

If one finds that they use the 5D3 when the subject is large and close, but tend to use the 7D2 when it is far or small, and then the 7D2 images look noisier or more noise-reduced or less sharp at 100% or after cropping, resized to the screen, that is not a weakness of the 7D2 sensor. That is a weakness of the situation - the subject or desired composition was a smaller (millimeter by millimeter) area of the sensor - it is not the sensor's fault that this area was smaller; it was the situation's fault, and you chose the smaller sensor, one reason being that you don't have to record pixels that serve no purpose, and in this 7D2/5D3 comparison, the 7D2 gives significantly more resolution (over double the pixels-on-subject).

The 7D2 is a better sensor than a 1.6x crop from the 5D3, hands down, at any ISO; the 7D2 has more resolution and less noise with less banding than a 1.6x crop from the 5D3 at any ISO or tonal level. It is unanimous across the board - nothing better about the 5D3 unless you fall prey to the illusion of 100% pixel views that magnify the 5D3 sensor area less. To see the same magnification of sensor area, you need to use 1.53x times the pixel magnification for the 5D3. You don't even need to actually do it; no 20MP APS-C with a sharp lens is as non-detailed-looking at 100% as the 5D3 or any camera is at 153%. The only way 153% can look sharp on a ~100PPI monitor is by using nearest neighbor, and that causes gross distortion, as 47% of the rows and columns of pixels are used as is, and 53% of them are doubled, giving a distorted, pixelated look.

To use your "light hungry" metaphor, a 1.6x crop from the 5D3 is more "light hungry" than the 7D2.

I don't have a 5D3, but I have a 6D, which has less noise and less banded noise than the 5D3 across the board, and I find the 6D (cropped), to be inferior to the 7D2 despite the fact that the 6D (cropped) has 1/3 stop less read noise at high ISO. It's at least partly the higher resolution, which reduces the ability of single-pixel noise impulses to obscure actual detail, but the character of the 7D2 noise seems to be more random, spatially, with finer chromatic noise.

The 3.1MP D30 might embarrass them both, by that standard, as long as you didn't need deep shadows at base ISO (the D30 is terrible there).

100% pixel view is an illusion, not a quality. 100% pixel views on historic, coarse monitors around 100 PPI that are sharp are aliased; period. You need to get up around 400 PPI or greater before a display can be both sharp and reasonably sampled, from a standard viewing distance. Acuity should occur over a gradient of a few fine pixels.

We could say that the 5D3 at 100% pixel view on q coarse monitor protects the sensor are quality by defaulting to lower magnification of each square mm of the sensor. Of course, we can leapfrog the 5D3 by binning the 7D2 conversions 2x2; then the 5D3 puts more final pixels-on-subject, and has noisier, softer 100% pixel views. We can play this leapfrog game ad infinitum, if we don't come to our senses and realize that imaging qualities occur at the final (possibly cropped) image level, not the arbitrary magnification resulting from 100% pixel views. Not images downsized with nearest neighbor, either.

Think about the future - higher resolution displays are going to favor higher-resolution captures of the subject or final composition. They don't have the room to hide the limitations. Use the sensor/camera that is best for what you are doing, and judge by the final image quality; not the pixel quality or full image quality (if you aren't using the entire frame).