Crop vs FF (Colours & DR)

D

DDirector

Veteran
I honestly never bothered to upgrade my stills camera from crop to FF - I have the Sony a6500.
A big reason is that I shoot the majority of my work on a telephoto lens. I love having the extra reach.

But, I've been comparing my photos to one from an a7rIII and I'm not sure if I was just being biased, but it seemed like the colours and DR from an a7rIII were better.
It is expected to be better since its more money, but is that actually the case?

Will the Dynamic Range and colour science be better?

They were not my photos though, they were someone else's but I felt like I had a slightly easier time grading them and it wasn't being clipped as fast.
But it could be from the particulate photos I had to edit.
 
I like to refer to the testing done by Dxomark. I don't know if they're 100% accurate gospel truth, but at least you can compare how different cameras fare under the same testing regimen.

Portrait score indicates color depth in bits, landscape score indicates dynamic range in stops, sports score indicates usable ISO (by what threshold I do not know).

The A7Riii beats the A6500 by varying degrees in each measure. (Actually, it basically doubles the performance in every category - more than an extra bit of color, more than an extra stop of DR, more than an extra stop of ISO).

https://www.dxomark.com/best-sony-cameras
 
Last edited:
It depends on what you shoot and your shooting scenario, but in my opinion, full frame isn't warranted for most shooting scenarios.
It also depends on the sensor, for example DXOMark rates my APS-C Nikon D7000 at 13.9 Evs and Canon's full frame 5D MK III at 11.7 Evs. Sony sensors seem to have the best dynamic range.

I know cameras have improved a lot, but I'm still content with my nearly decade-old Nikon D7000 for photography. Grad NDs, CPL, and bracket photos allow enough dynamic range for most scenes. It would be nice to have better AF and low light performance, but Nikon hasn't come out with an appealing enough camera to convince me to upgrade, the Z6 comes close, but it's still a little lacking on the video side.
 
It has little to do with the sensor size, tends to have more to do with the time, money, research, experience, aesthetics, and energy (etc etc) put into the camera.

Smaller sensors tend to have smaller pixels or lower resolution or higher demands on optics. In stills world, crop sensors tend to be lower tier. I think you may be experiencing all of the above.
 
James0b57 said:
It has little to do with the sensor size, tends to have more to do with the time, money, research, experience, aesthetics, and energy (etc etc) put into the camera.

Smaller sensors tend to have smaller pixels or lower resolution or higher demands on optics. In stills world, crop sensors tend to be lower tier. I think you may be experiencing all of the above.
Click to expand...

I agree in the sense that the Alexa delivers better color and DR performance than most or all stills cameras. But the Alexa LF has better DR, low-light (and possibly color?) performance than the Alexa. As well as being higher resolution. So, I think that with the same pixel pitch and type and overall imaging pipeline, a larger sensor will have some fundamental benefits. (Of course, the tradeoffs include less DOF and slower rolling shutter.)
 
Doesn't the LF use the same sensor as the Classic, just with a couple parallel sensors, like a digital version of Cinerama?
 
ahalpert said:
I agree in the sense that the Alexa delivers better color and DR performance than most or all stills cameras. But the Alexa LF has better DR, low-light (and possibly color?) performance than the Alexa. As well as being higher resolution. So, I think that with the same pixel pitch and type and overall imaging pipeline, a larger sensor will have some fundamental benefits. (Of course, the tradeoffs include less DOF and slower rolling shutter.)
Click to expand...

Haha! I’m guessing you just skimmed over my post.

The OP is likely long gone and he already had at least three good answers explained in different ways.
 
James0b57 said:
Haha! I’m guessing you just skimmed over my post.

The OP is likely long gone and he already had at least three good answers explained in different ways.
Click to expand...
lol nope. Still here.

Interesting input. I will look into it more.
I think I did notice those test numbers before. They seem very tiny differences, so I didn't think it would really noticable.

Will do more digging into this.

Curious to know how/if the test results change if sitting in crop mode on a FF camera.
 
DDirector said:
lol nope. Still here.

Interesting input. I will look into it more.
I think I did notice those test numbers before. They seem very tiny differences, so I didn't think it would really noticable.

Will do more digging into this.

Curious to know how/if the test results change if sitting in crop mode on a FF camera.
Click to expand...

I'm not sure, but I think you would be right back where you started. The noise will be larger, so you won't be able to push the gain as high. Your dynamic range will be decreased because the image is noisier. And the resolution will be decreased. But that's assuming that the pixels are the same size/type and have the same processing on the larger sensor as on the smaller one (as is the case with the Alexa LF and the regular Alexa).
 
DDirector said:
lol nope. Still here.

Interesting input. I will look into it more.
I think I did notice those test numbers before. They seem very tiny differences, so I didn't think it would really noticable.

Will do more digging into this.

Curious to know how/if the test results change if sitting in crop mode on a FF camera.
Click to expand...

At the rate this thread is going, just stick with FF35.


The only time full frame is better, is when all things are equal.
 
ahalpert said:
Yes I think so, but the net effect is that it has less noise (due to larger light-gathering area for the equivalent shot?) and it also has a half-stop more dynamic range (possibly also due to lower noise? I'm not sure of the physics on that).
Click to expand...

I think it's more to do with more specialised and updated internals. The A65 is also a small run highly specialised camera. I'd guess there's less than 60 of them in the world. In order to make one they have to find 3 regular sensors that match very closely already and then stitch them together and then a lot of them still don't work out.

So you're getting this kind of best of the best binning effect where the individual best sensors are chosen and then the best groups of three of those sensors get to make into a camera that has more couture finished updated boards and electronics.

I have a friend that's behind the scenes at Arri and she told me they go to an awful lot of trouble to join those three sensors together and they reject a lot more than they keep. And it's not like they can then recycle those sensors. Once they're mounted and joined that's it. They're throwing a bunch of sensors away that would have been "above average" on a normal Alexa...

I think any effect of the actual sensor size on DR and certainly on colour gamut is insignificant in real terms. The improvements are more to do with other factors as described...

JB
 
John Brawley said:
I think it's more to do with more specialised and updated internals. The A65 is also a small run highly specialised camera. I'd guess there's less than 60 of them in the world. In order to make one they have to find 3 regular sensors that match very closely already and then stitch them together and then a lot of them still don't work out.

So you're getting this kind of best of the best binning effect where the individual best sensors are chosen and then the best groups of three of those sensors get to make into a camera that has more couture finished updated boards and electronics.

I have a friend that's behind the scenes at Arri and she told me they go to an awful lot of trouble to join those three sensors together and they reject a lot more than they keep. And it's not like they can then recycle those sensors. Once they're mounted and joined that's it. They're throwing a bunch of sensors away that would have been "above average" on a normal Alexa...

I think any effect of the actual sensor size on DR and certainly on colour gamut is insignificant in real terms. The improvements are more to do with other factors as described...

JB
Click to expand...

That's very interesting about the A65 manufacture.

Yes, it seems that I was wrong about inherent advantages of larger sensors, although they do have qualified advantages. At the same resolution and thus larger pixel pitch, the larger sensor should have better low-light performance and thus more DR at least at higher ISOs (although yes, the Alexa LF has the same pixel pitch).

My experience with larger sensors from the same manufacturer has always been better low-light and DR performance, but that may be strictly due to the pixel pitch and, as you mentioned, manufacturing quality.

Here are 2 graphs from dxomark showing that sensors with larger pixel pitch have better low-light performance regardless of design year, but that sensors with more recent design years have more DR regardless of pixel pitch. The pixel pitch (eg 8.4) is notes for some data points:

ISO performance:
high-ISO.jpg

Dynamic range:
Dynamicrange.jpg

So, the moral of the story is that you can't make good shots unless you're shooting on an A65.
 
ahalpert said:
That's very interesting about the A65 manufacture.

Yes, it seems that I was wrong about inherent advantages of larger sensors, although they do have qualified advantages. At the same resolution and thus larger pixel pitch, the larger sensor should have better low-light performance and thus more DR at least at higher ISOs (although yes, the Alexa LF has the same pixel pitch).

My experience with larger sensors from the same manufacturer has always been better low-light and DR performance, but that may be strictly due to the pixel pitch and, as you mentioned, manufacturing quality.

Here are 2 graphs from dxomark showing that sensors with larger pixel pitch have better low-light performance regardless of design year, but that sensors with more recent design years have more DR regardless of pixel pitch. The pixel pitch (eg 8.4) is notes for some data points:

ISO performance:
View attachment 139010

Dynamic range:
View attachment 139011

So, the moral of the story is that you can't make good shots unless you're shooting on an A65.
Click to expand...

Larger sensors tend to have worse rolling shutter.
 
James0b57 said:
Larger sensors tend to have worse rolling shutter.
Click to expand...

True, but I think that as long as the RS is fast enough, no one minds. Alexa mini was 3ms, C300mk2 was 6ms, and C300mk3 in DGO mode is 15.3ms (all according to cinema5D) but I don't hear anyone complaining about it. (Although I do think the fast RS is part of why people like the Alexa's motion cadence.)

Also - I did some research on CMOS dynamic range. It does seem that well capacity and therefore pixel pitch corresponds to higher potential dynamic range. But all the various types of noise reduce dynamic range, and since most types of noise scale up with pixel pitch, the dynamic range of a larger pixel is basically equivalent.

And therefore the gains in dynamic range come from reducing noise in various ways, which is why newer and more refined and expensive engineering techniques can yield better results.
 
A lot of mention and comparisons to other video/cine cameras.
I was merely looking for photo only and not video. So, not sure if that changes things in what you're saying when comparing it to video.
 
Maybe consider a different perspective if the DXO testing on page 1 isn't enough; almost every single professional photographer on the planet uses a full-frame or medium-format camera for stills photography.

(Yes, someone can run a successful business and make millions with a 1" sensor stills camera, but just saying.)
 
I hae a D600 its no different for stills from the 5d2.

One has good reviews the other does not.

It only makes a difference when you do high end sport or top fashion.

If its just billboards and glossy mags they are all good enough if not 10 years old.

But yep FF35 is the norm.
 
DDirector said:
A lot of mention and comparisons to other video/cine cameras.
I was merely looking for photo only and not video. So, not sure if that changes things in what you're saying when comparing it to video.
Click to expand...

I would take a look at the evaluations at dxomark because they are evaluating the cameras' stills abilities. And they are measuring their performance in practice, not in theory, which is all that matters for us.

That being said, the physical properties of the sensor/pixels etc as they relate to DR and color depth are the same for stills and for video. What differs is the image processing because video requires a higher framerate than these cameras can shoot for maximum quality stills. So, typically, video performance is inferior in regard to bit depth, dynamic range, sharpness, codec, etc.

And regarding rolling shutter - if you are using the mechanical shutter to take stills, AFAIK the rolling shutter speed does not matter. If you are using the silent shooting mode electronic shutter, you are relying on the rolling shutter and its speed is a factor. But I've only heard photographers complain about rolling shutter in specific circumstances like fluorescent lighting maybe.
 
You must log in or register to reply here.
Back
Top