bclaff
Forum Pro
Are you certain that the same sharpening algorithms and settings are being used?
(I doubt it.)
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
The image qualities of Sony vs Canon sensors
- Thread starter chironNYC
- Start date
(I doubt it.)
chironNYC
Senior Member
Ron Poelman
Veteran Member
Doesn't make it any less right.
chironNYC
Senior Member
he never said that the sensors were the same, you said that... this is a very complex subject, influenced by many things, including yes the sensor, depending on how it's used.
if you wanted to look at sensor performance in the real world, maybe this a7r vs. 5dmk3 shadow recovery comparison: http://www.fredmiranda.com/A7R-review/
another approach would be to use bill claffs sensor page:
http://www.photonstophotos.net/Charts/PDR.htm#Canon EOS 5DS R,Nikon D5,Sony ILCE-7R
--
dan
Last edited:
SilvanBromide
Veteran Member
No and in fact it couldn't be any less right.
As a former Canon shooter I know well that with the same photographer and the same techniques, the RAW coming off the Sony sensors looks different from the RAW I routinely got from Canons. Post processing mostly serves to amplify (but in any case not eradicate) those differences.
bclaff
Forum Pro
I don't think my dynamic range measurements speak to the original posters concerns.he never said that the sensors were the same, you said that... this is a very complex subject, influenced by many things, including yes the sensor, depending on how it's used.
if you wanted to look at sensor performance in the real world, maybe this a7r vs. 5dmk3 shadow recovery comparison: http://www.fredmiranda.com/A7R-review/
another approach would be to use bill claffs sensor page:
http://www.photonstophotos.net/Charts/PDR.htm#Canon EOS 5DS R,Nikon D5,Sony ILCE-7R
But my comments on the OLPF, microlens, and color sensitivity do.
As well as the differences in firmware regarding color processing, sharpening, etc.
Regards,
--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )
B
BBQue
Guest
That was mostly hype and initial over-excitement due to the high ISO possibilities of this camera. And yes, 12MP can look really smooth at 100%. But if you downsample A7rii to 12MP it looks just as good as the A7s. (I have both cameras).
chironNYC
Senior Member
B
BBQue
Guest
How do you know for sure that this was due to differences with the sensor? Did you use the same raw processor? Does the raw processor treat Sony files exactly same as Canon files?
I agree that there are differences, I just am not sure how much of that to attribute to the sensor hardware - which is the topic of this thread.
(((I also used Canon (the venerable 5D) before I switched to Sony.)))
Last edited:
chironNYC
Senior Member
According to someone with as discerning an eye as Michael Reichmann, these cameras with their very different sensors have quite a different look, and he wrote multiple times that the A7s had the look of medium format to him, something that he attributed to the unusually large size of the pixels in that camera's sensor. They both might look good, or have as little noise when the A7r files were resized, but they still looked quite different to him.
I don't quite understand why anyone would even want to argue that the sensor doesn't make a difference to the qualities of the image. It seems like a settled topic. Differences in dynamic range and color sensitivity alone would alter how images look.
Last edited:
B
BBQue
Guest
I do respect Michael Reichmann very much - and at the time he wrote that review, the A7s surely was a revolutionary camera with aa unmatched high ISO look.
I never claimed that sensors make no difference - that would be silly indeed. But I simply do not believe that you can look at 100 random Flickr images from Sony versus Canon cameras and then attribute the differences you are seeing to the sensor hardware (as opposed to the post processing skills of the photographers). Just my two cents.
Of course there are differencies. But then there are also differencies in lenses (even from the same manufacturer), in firmware (even different versions for the same camera), in PP software, in peoples settings of the camera, in peoples PP, in screens, in printing etc. When you have "filtered the light" through all those variables, I think it will be very hard to determine exactly which comes from sensor differencies. An not too relevant either because a picture can and will be tweaked in so many ways. When I look at the finished pictures taken with my old Nikon D700 and my Sony A7S (both 12MP sensors), I can't tell the difference after they have been through PP, despite there is an age difference of more than 6 years between the two sensors.
I think that's why many give answers to the effect of "There are differencies, but who cares?". You may determine differencies by technical measurements with specialist instruments, but not by looking on finished pictures. I'm talking about "rendering" here, not resolution (where you can of course see a difference between a 12MP and a 42MP sensor by pixel peeping, provided you keep the original file size of the latter).
--
"Good photos is not about fancy gear. It's about how you see as a photographer and about being ressourceful with whatever gear you have." (Alfred Eisenstadt, Life photographer)
Last edited:
chironNYC
Senior Member
chironNYC
Senior Member
Astrophotographer 10
Forum Pro
- Messages
- 13,956
- Solutions
- 11
- Reaction score
- 6,160
Pixel size would play a part aside from sharpening routines etc.
The advantages of a large pixel are usually the large well depth and greater sensitivity as it catches more light being larger. Well depth is how many electrons each pixel can hold before its full. In astrophotography I have used many different CCDs but usually the pixel sizes are larger than what is used in CMOS cameras. Part of that may be CCDs would tend to have larger pixels due to no surrounding circuitry on each pixel.
Another factor would be antiblooming. Antiblooming is a feature of most CCDs and I assume CMOS where once the wells are fully saturated additional electrons are bled off by an antiblooming mechanism otherwise you get blooms or weird horizontal lines coming out of stars for instance. If that is not set right you can get artifacts like a bump on the side of a bright star. What that would look like on a daylight camera I am not sure. Probably weird artifacts on bright light pointy sources.
One camera I use has large pixels - 9 microns, exKodak sensor. It has massive full well depth of around 1000,000 electrons (like the original Canon 5D). What I find there is that camera can give smoother images because things like stars overexpose easily and it can take a longer exposure before it starts to overexpose. So processing the image later the stars can take much more processing before it breaks down. On the other hand I have used a highly sensitive but small pixelled very low noise but very low full well depth Sony CCD with 4.54 micron pixels and it was the opposite. The stars tended to break down easily when processing and form fluffy/fuzzy edges easily.
I am not sure how it works in daylight cameras but in telescopes there is definitely a formula for matching pixel size to the focal length of the scope and to the usual atmospheric turbulence of the imaging location.
But I imagine that matching lens sharpnesses to the pixel size has a big impact as we are finding as higher megapixel cameras come out. Old lenses tend now to show up as not being as sharp as we thought they were. A7r2 and A7r were showing that up quite clearly and is the case for GM lenses being capable of handling high mp cameras.
Greg.
The advantages of a large pixel are usually the large well depth and greater sensitivity as it catches more light being larger. Well depth is how many electrons each pixel can hold before its full. In astrophotography I have used many different CCDs but usually the pixel sizes are larger than what is used in CMOS cameras. Part of that may be CCDs would tend to have larger pixels due to no surrounding circuitry on each pixel.
Another factor would be antiblooming. Antiblooming is a feature of most CCDs and I assume CMOS where once the wells are fully saturated additional electrons are bled off by an antiblooming mechanism otherwise you get blooms or weird horizontal lines coming out of stars for instance. If that is not set right you can get artifacts like a bump on the side of a bright star. What that would look like on a daylight camera I am not sure. Probably weird artifacts on bright light pointy sources.
One camera I use has large pixels - 9 microns, exKodak sensor. It has massive full well depth of around 1000,000 electrons (like the original Canon 5D). What I find there is that camera can give smoother images because things like stars overexpose easily and it can take a longer exposure before it starts to overexpose. So processing the image later the stars can take much more processing before it breaks down. On the other hand I have used a highly sensitive but small pixelled very low noise but very low full well depth Sony CCD with 4.54 micron pixels and it was the opposite. The stars tended to break down easily when processing and form fluffy/fuzzy edges easily.
I am not sure how it works in daylight cameras but in telescopes there is definitely a formula for matching pixel size to the focal length of the scope and to the usual atmospheric turbulence of the imaging location.
But I imagine that matching lens sharpnesses to the pixel size has a big impact as we are finding as higher megapixel cameras come out. Old lenses tend now to show up as not being as sharp as we thought they were. A7r2 and A7r were showing that up quite clearly and is the case for GM lenses being capable of handling high mp cameras.
Greg.
Astrophotographer 10
Forum Pro
- Messages
- 13,956
- Solutions
- 11
- Reaction score
- 6,160
I think you can see differences from different sensors. Certainly in the astrophotography world I could pretty often tell what camera was used if you showed me the image only with no details and be right more than 50% of the time.
Greg.
Greg.
chironNYC
Senior Member
One of the more experienced professional photographers over at fredmiranda.com answered my question this way:
"Sony sensors have a ton of DR so right out of the gate it will have smoother transitions and look a little more flat than Canon because it's seeing into the shadows better. That's one basic the other is for some reason the 42mpx just has a really nice tonal palette between colors and tone. Call that tonal range which is part of the DR but it goes beyond that and I saw it the first day I shot the A7rII over my A7r. This 42mpx sensor is in my mind the closest sensor I have seen to a MF back. It just has a very nice tonal range and transmission fall off off between tone and color. Than you get into the lenses and than you get the character , look of images that is very pleasing with this sensor. No joking here this is the best sensor I have seen in 35mm format and I'm digital only for the last 25 years. I owned 5 MF backs and this sensor matches up really damn good. So out of the gate your starting with a really nice sensor that has great qualities to it. I was thrilled the first day I shot this sensor."
So, he is suggesting that the differences in some of the image characteristics of Sony sensors come from:
1. "smoother transitions and look a little more flat than Canon because it's seeing into the shadows better"
2. "the 42mpx just has a really nice tonal palette between colors and tone"--or color sensitivity in relation to tonal or exposure sensitivity
3. "It just has a very nice tonal range and transmission fall off off between tone and color"--so, again, the idea of the way the colors and the exposure intensities roll-off and transition throughout the image, which is likely a function of color sensitivity, dynamic range, and possibly other sensor factors (and other non-sensor but brand-specific factors, like raw processing).
This is very much what I see when I look at images from Sony sensors in comparison to images from the Canon sensors..
"Sony sensors have a ton of DR so right out of the gate it will have smoother transitions and look a little more flat than Canon because it's seeing into the shadows better. That's one basic the other is for some reason the 42mpx just has a really nice tonal palette between colors and tone. Call that tonal range which is part of the DR but it goes beyond that and I saw it the first day I shot the A7rII over my A7r. This 42mpx sensor is in my mind the closest sensor I have seen to a MF back. It just has a very nice tonal range and transmission fall off off between tone and color. Than you get into the lenses and than you get the character , look of images that is very pleasing with this sensor. No joking here this is the best sensor I have seen in 35mm format and I'm digital only for the last 25 years. I owned 5 MF backs and this sensor matches up really damn good. So out of the gate your starting with a really nice sensor that has great qualities to it. I was thrilled the first day I shot this sensor."
So, he is suggesting that the differences in some of the image characteristics of Sony sensors come from:
1. "smoother transitions and look a little more flat than Canon because it's seeing into the shadows better"
2. "the 42mpx just has a really nice tonal palette between colors and tone"--or color sensitivity in relation to tonal or exposure sensitivity
3. "It just has a very nice tonal range and transmission fall off off between tone and color"--so, again, the idea of the way the colors and the exposure intensities roll-off and transition throughout the image, which is likely a function of color sensitivity, dynamic range, and possibly other sensor factors (and other non-sensor but brand-specific factors, like raw processing).
This is very much what I see when I look at images from Sony sensors in comparison to images from the Canon sensors..
chironNYC
Senior Member
Greg,
This is an extremely interesting answer, especially your point about well depth in relation to how much exposure a pixel can stand and still not break down, thus in effect producing a smoother range of exposure transitions within an image. And it is interesting that this was one of the characteristics of the original Canon 5D, which did have a very lovely smooth image quality that seems in many ways more like the current Sonys than it is like the current Canons. I am sure that there are many other factors that can contribute and i don't actually know what the well depth is on Sony sensors, but this seems like it is likely to be an important one. For whatever reasons, the smoothness of the transitions is to me one of the defining characteristics of the Sony sensors.
Astrophotography is such a demanding arena for sensors and lenses that it seems like it would be very informative about many technical aspects of image-making.
Last edited:
bclaff
Forum Pro
I was referring to algorithms and settings in the camera. These can never be made identical.
Keyboard shortcuts
- f
- Forum