Actual usable data from different sensors

AndyWest

AndyWest

Messages
8,400
Name
Andy
Edit My Images
No
I wonder if anyone can help. I like to heavily process evening landscape/sunset images a lot and currently have M4/3 gear only. I have had APS-C and full frame before but find M4/3 so much fun to use and easier as it's less weight etc. When I had a 6D I found that the RAW files were a joy to process. Loads of data and sharp files. I could get a lot from them. With M4/3 it 'seems' that there is less information from the RAW files to play with. I'm not sure if this is the case or if I'm just being daft. SO, my question is:

3 cameras, 1 Full frame, 1 APS-C and 1 M4/3

All 3 are 16 Megapixel, Would I get more RAW data to play with in editing with the larger sensor?

I'm guessing yes because a 16 megapixel compact will not give the same IQ as a 16 megapixel FF camera. I'm guessing as each pixel is larger you get more light and better IQ but actual usable editing data is the bit I don't really understand. Is there more of it in larger sensors?

M4/3 seems to be great for normal daylight photography but when push comes to shove they are not quite there in low light situations or am I wrong.

Thanks in advance.
 
I guess as long as it's the same settings used (exposure/iso etc) then the only difference would be the dynamic range of the sensor
 
Phal said:
I guess as long as it's the same settings used (exposure/iso etc) then the only difference would be the dynamic range of the sensor
Click to expand...

Yep, you would think but compare at 100% a RAW from a compact and a RAW from FF, The differences are huge especially in detail/IQ. Where is all this extra data sitting?
 
The sensor is only as good as the lens in front of it too so if you're looking for more detail then I'd look at that first. Compacts won't have very good lenses to shoot with so their sensor will be rather handicapped.
 
Phal said:
The sensor is only as good as the lens in front of it too so if you're looking for more detail then I'd look at that first. Compacts won't have very good lenses to shoot with so their sensor will be rather handicapped.
Click to expand...
My scenario is if all were equal i.e same quality lens at same equiv aperture etc. There must be more to it than just lenses.
 
From personal experience, and this is a"gut" feeling, of shooting (sometimes at high ISO values), with a 12mp full frame sensor, a 10mp crop (APS-C) sensor and a small sensored P&S, and almost always in RAW, I know which image files I would rather work with.
I shoot indoor classical music (no flash) and the P&S at ISO 1600(+) is almost a totally waste of time so I never use it for those shoots. The full frame sensor is better than the crop sensor for these shoots.

I use the same lenses for both full frame and crop bodies when shooting hand held in low light.
 
Last edited:
Obviously FF sensors will have the advantage on high ISO shots :)

If you notice a difference in detail and sharpness then I would say the factor there is the lens and available sensor resolution. If you notice less detail that you can recover from shadows/highlights in your processing then I would say the sensor has less DR than you're used to.
 
I'd be surprised to see any difference in dynamic range and ISO performance comparing 24 MP FF sensor to 12 MP APS-C to 6 MP m43. This is because the photodiodes are all the same size, just different crop of the same image.

But you are comparing at same resolution, meaning the smaller the sensor, the smaller photodiode. Each photodiode has to be able to produce the same image with less light. Think of photodiodes as water buckets: larger FF photodiodes has wider opening, can gather 0 to 100 water at 1/30s; quarter sized m43 photodiodes has has much smaller sized opening, can only gather 0 to 25 water with same amount of time. After gathering the light and converting into digital data, the m43 processor has to produce the same image with less range of samples.

High ISO shooting is just boosting ADC sensitivity. Using same numbering system, say we shoot at ISO 800, that is 3 stops from ISO 100, 2^3=8 times sensitivity. So the FF water bucket has range of 0 to 12.5 to be counted as 0 to 100. A slight difference between each diodes will be amplified as noise. The M43 water bucket will need 0 to 3.13 to be counted as 0 to 25, which has more chance to amplify photodiode variations.

So smaller photodiodes has less dynamic range? yes. But the other question is: is it noticeable enough to worth the effort of carrying FF glass around? no. ADC technology and image processing have come a long way to able to bend laws of physics, to our eyes.
 
I have moved from a Nikon D7000 (APS-C) to Sony A7 (FF) and the difference is there when working with RAW files.
So much more data seems to be held within the FF RAW files compared to the Nikon.
I guess it just gives me more room to adjust the photo's which we taken poorly too :)
 
Could also be that the more 'pro' cameras have greater bit depth in their Raw files (14 or more bits versus 12)?
 
ChrisR said:
Could also be that the more 'pro' cameras have greater bit depth in their Raw files (14 or more bits versus 12)?
Click to expand...
Yes I think that has a lot to do with it. I hadn't thought of bit depth.

Serge Ramelli (LR & PS Tutorial guy) has recently been shooting with a Sony A7R over the 5D Mk2 and he keeps going on about how much more detail there is in the A7 files. Resolution has much to do with this too i guess.

So compare a 6D @ 20Megapixels and a Sony RX100 at 20 Megapixels and you will find I'm sure more scope for editing in the FF sensor images. So this must be down to either Bit depth or something else.
 
larger pixel - higher signal to noise ratio. You can't cheat physics, and the following will always be true MF > 35mm > APS-C > 4/3 > compacts. Better SNR results in better colour range, dynamic range and less noise. Top of the range cameras also do far less aggressive on the chip processing and produce more neutral files that can handle a lot of post work. This is not to say m43 will be bad, but recent FF will always be at least slightly to a lot better given the circumstances and the needs.
 
Larger sensors always win this kind of contest. It's simply a matter of physical size that drives every benefit (all things being equal, or at least similar generation tech).

FF is 4x larger than M4/3, it collects 4x as many photons/light. That's two stops more dynamic range for starters.

The other factor is larger sensors don't work the lens so hard, because they require less enlargement for same size output. So resolution demands are greatly reduced, and when resolution goes down, then image contrast goes up - better sharpness, more punch. Basic MTF lens theory.

It's just phsyics. All the other stuff like pixel density and spacing and well depth or whatever, is just tinkering around the edges. Bigger is always better.
 
Very interesting stuff I know that the larger sensor on full frame cameras will always be better but am I right in saying that as technology improves smaller sensors will get better so the difference will be less important?
Will say a Canon crop camera in a few years time be as good as say the 5D mk3 is now?
 
LCPete said:
Very interesting stuff I know that the larger sensor on full frame cameras will always be better but am I right in saying that as technology improves smaller sensors will get better so the difference will be less important?
Will say a Canon crop camera in a few years time be as good as say the 5D mk3 is now?
Click to expand...

Maybe...? but the limitations of optics will still remain, and with that thinking the then current 5D would be able to cleanly see in the dark.
 
HoppyUK said:
FF is 4x larger than M4/3, it collects 4x as many photons/light. That's two stops more dynamic range for starters.
Click to expand...

This is the precise bit that I always struggle with and remember I was an electronic guy not an optical one.

A FF sensor is bigger than a MFT sensor but most of the additional light falling on a FF sensor falls outside of what would be a MFT sized area in the middle of the FF sensor. The actual number of photons captured by an individual pixel within a MFT sized area of the FF chip I'd have thought should be the same as the number captured by an identical pixel on a MFT sensor assuming the same exposure and identical sized pixel. Unless I've got something very wrong. My reasoning is that if I nail my light meter to my garage wall and take a reading on a particular brick the reading wont change if I knock down most of the wall. The same amount of light, the same number of photons, will I assume fall on the same brick even if it's now part of a smaller wall.

We should be able to test this quite easily these days as there are FF cameras that can work in an APS-C crop mode and I'm pretty sure that if we shoot with one of those or even simpler just crop a FF image to MFT size there will not be a 2 stop difference in DR or indeed any other major difference between the two pictures unless we do something else like enlarge the smaller image more and thus make any issues more visible or when shooting change some settings. Given two identically exposed images, one FF and one smaller, and if the technology is the same the only differences between the two images should I'd have thought be those created by different levels of enlargement otherwise if we cropped a FF image we'd lose two stops of DR and do all the other bad things that happen to smaller format pictures and I don't think that's what generally happens.

Any differences we see between pictures taken with different cameras other than those due to different enlargement should I'd have thought be due not to the physical size of the sensor as such but rather the pixel size, A/D converter, microlenses, glass in front of the sensor, other gubbins, how the camera processes the images, the camera settings and the lens.

Oh well, like I said I'm an electronics guy not an optical one.

And back to one of the OP's questions...

AndyWest said:
M4/3 seems to be great for normal daylight photography but when push comes to shove they are not quite there in low light situations or am I wrong.
Click to expand...

I used to have a 5D and I now have a A7, a G1 and a GX7. The G1 is nowhere near as good as the 5D was at higher ISO's but the GX7 is much closer. I wont say it's better as I haven't compared the two that closely. Anyway, at higher ISO's as the GX7 is much closer / pretty close or possibly better than the old 5D (and better than my APS-C 20D) and much better than the G1 despite having the same sized MFT sensor so there's obviously a lot more going on than just sensor size as such.
 
LCPete said:
Very interesting stuff I know that the larger sensor on full frame cameras will always be better but am I right in saying that as technology improves smaller sensors will get better so the difference will be less important?
Will say a Canon crop camera in a few years time be as good as say the 5D mk3 is now?
Click to expand...

Yes and no. Technological development will make smaller sensors more efficient, and make lenses sharper. It is also easier to make sharper lenses for smaller formats. So yes, it may be that some years in the future a M4/3 camera will match a 5D3 of today, but since the same advances will also be applied to larger formats, the differential will remain.

Perhaps more to the point, smaller formats will likely be 'good enough' for most people (they pretty much are already) so the benefits of FF will be marginalised. I'm not sure though, that FF will lose its appeal. We always want more, even if we don't need it: better sharpness, greater dynamic range, higher ISO, shallower DoF, etc. FF will always be better at these things.
 
woof woof said:
This is the precise bit that I always struggle with and remember I was an electronic guy not an optical one.

A FF sensor is bigger than a MFT sensor but most of the additional light falling on a FF sensor falls outside of what would be a MFT sized area in the middle of the FF sensor. The actual number of photons captured by an individual pixel within a MFT sized area of the FF chip I'd have thought should be the same as the number captured by an identical pixel on a MFT sensor assuming the same exposure and identical sized pixel. Unless I've got something very wrong. My reasoning is that if I nail my light meter to my garage wall and take a reading on a particular brick the reading wont change if I knock down most of the wall. The same amount of light, the same number of photons, will I assume fall on the same brick even if it's now part of a smaller wall.

We should be able to test this quite easily these days as there are FF cameras that can work in an APS-C crop mode and I'm pretty sure that if we shoot with one of those or even simpler just crop a FF image to MFT size there will not be a 2 stop difference in DR or indeed any other major difference between the two pictures unless we do something else like enlarge the smaller image more and thus make any issues more visible or when shooting change some settings. Given two identically exposed images, one FF and one smaller, and if the technology is the same the only differences between the two images should I'd have thought be those created by different levels of enlargement otherwise if we cropped a FF image we'd lose two stops of DR and do all the other bad things that happen to smaller format pictures and I don't think that's what generally happens.

Any differences we see between pictures taken with different cameras other than those due to different enlargement should I'd have thought be due not to the physical size of the sensor as such but rather the pixel size, A/D converter, microlenses, glass in front of the sensor, other gubbins, how the camera processes the images, the camera settings and the lens.

Oh well, like I said I'm an electronics guy not an optical one.

And back to one of the OP's questions...



I used to have a 5D and I now have a A7, a G1 and a GX7. The G1 is nowhere near as good as the 5D was at higher ISO's but the GX7 is much closer. I wont say it's better as I haven't compared the two that closely. Anyway, at higher ISO's as the GX7 is much closer / pretty close or possibly better than the old 5D (and better than my APS-C 20D) and much better than the G1 despite having the same sized MFT sensor so there's obviously a lot more going on than just sensor size as such.
Click to expand...

If you take a FF image and crop it down to M4/3 size, you effectively have a M4/3 format camera. You've just thrown away three-quarters of the image, and all the benefits of the larger sensor area with it.
 
HoppyUK said:
Yes and no. Technological development will make smaller sensors more efficient, and make lenses sharper. It is also easier to make sharper lenses for smaller formats. So yes, it may be that some years in the future a M4/3 camera will match a 5D3 of today, but since the same advances will also be applied to larger formats, the differential will remain.

Perhaps more to the point, smaller formats will likely be 'good enough' for most people (they pretty much are already) so the benefits of FF will be marginalised. I'm not sure though, that FF will lose its appeal. We always want more, even if we don't need it: better sharpness, greater dynamic range, higher ISO, shallower DoF, etc. FF will always be better at these things.
Click to expand...


You have a good point there about current cameras being good enough for most people my 7D is plenty good enough for me already
Im normally at ISO 400 or 800 and the image quality is excellent
its only occasionally in the winter in low light that I struggle and have to raise the ISO and then its not so good
 
Last edited:
if you want tons of real detail in your pictures buy a foveon sensor camera, they kill everything else for detail
 
FF will win due to it's larger photosites. Most of the issues with image quality (noise/DR/color data/etc) is due to a lack of "sensor saturation," or said another way; a lack of "energy" or "information" collected by the photosites.
Think of it as the difference between catching rain with a bunch of teacups as opposed to catching rain with a bunch of buckets. The larger buckets are going to catch more rain (energy) per bucket, and that's what matters.

"Detail" is a matter of "matching dots;" imagine taking a picture of a chessboard of 64 dots. I could use a pretty crappy lens that could only project 64 dots onto a sensor with only 64 pixels and get a fine image (assuming all of the dots lined up).
But that's not what we usually do... we normally take an image existing of billions of points of light, project it with a lens capable of projecting only some fewer MP's of light points (dots), at an aperture that limits it to even fewer MP's of light points, onto a sensor that may have more/fewer MP's of points to collect the light... and then we print it on a printer capable of producing more/fewer points of color/light. The limiting factor to detail will be whichever component is the least capable of reproducing the original billions of points of light. In general, the limitation here is the lens. Most lenses are not capable of producing the very small points of light required by the smaller photosites.

And then there is "sharpness." Our perception of sharpness is largely based upon contrast and not actual detail... it's similar to newspaper print made up of many discernible individual dots or an image made from a bunch of smaller/different images. If you look closely at the fine detail the original text/large image becomes "blurry." And if you back off the fine detail is lost but the larger contrast increases and the text/larger image comes into focus.
The acceptable requirement for sharpness requires very few MP's of detail (<2MP). Here the limitation to contrast is again pixel size.
Consider a black and white image made up of only two pixels from the FF sensor... one will be black and the other white. There will be a "hard line" of definition and greater contrast.
That same area would be covered by 8 pixels on the 4/3 camera. The transition between black and white will be more gradual with a mix of black and white pixels intermixed along the edges...the result is less "definition/contrast" and thus a lower perception of sharpness.
 
You must log in or register to reply here.
Back
Top