Why can't someone post a side-by-side comparison of a 33mp to a 67mp image of the same subject with the subject the same size when posted? One could then judge for themselves whether or not the difference matters to them when posting or viewing images on this site.
mudlake wrote:
Question that may have already been answered:
Why did Sony leave out medium raw? I use it all the time on my A7R5. I love how I can shoot 26mp images and STILL get 26mp when I use crop mode in mRaw. Anyone know why they left this useful tool out on the new camera?
Idea of smaller resolutions raws is appealing but the implementation is always terrible in my view.
People have expectation that smaller raw will be as good as if image from larger raw would be downscaled.
All the common sensors use Bayer mask filter array. Except many Fujifilm APS-C sensors where different mask is used but same concept and Sigma's Foveon which record color information in every pixel, but that is true exception and rare.
So the raw file contains single color component for every pixel. In this sense the actual resolution is lower than state number of megapixels because resolution of greens is only a half and resolution of reds and blues is only a quarter of stated resolution.
Bayer array is hard to downscale. It can be done if you would want 1/4 of MP. But that is too low resolution on most cameras. With 66MP it would be 16.5MP which might be ok, but consider that it would contain just 4MP of blue or red color. You would not get the same thing as if normal raw is demosaiced and downscaled to 16.5MP, it would be worse. It would be small though.
So to provide something between 1/4 and full MP of resolution the cameras that support smaller raws need to demosaic Bayer data into RGB information for each pixel, then downscale it.
So at the start you have 16.5M red, 16.5M blues and 33M greens. You demoisaic it into 66MP of RGB pixels, with 66M red, 66M green, 66M blue. Then you downscale it to those 26MP. So you have 26M red, 26M green, 26M blue. So instead of original 66M 14 bit values you now have 26M*3=78M 14 bit values. This is worse, but it enables further processing. Now you can reduce the bit depth to make it smaller. And to make it better you can precompute white balance (if you can guess right one) before reduction so that it will be done at higher precision. You can also do some other processing that would be worse to do after bit reduction. And then you store only 11 or 12 bits (Nikon cameras do this with small raw formats and I guess Sony too). Sometimes this is done in different colorspace, not RGB, but YCbCr where bit reduction is not perceived by our eyes as bad.
So now you have 78M values at 11 bit depth. That is still not much better than what you have started with. Then you probably skip some. Otherwise you would not get to smaller sizes. The end result is terrible.
So on a7rv if you use Lossless L, the image quality is perfect. If you switch to Lossless M or Lossless S with same exposure parameters you will get less highlight headroom (clipping) by 1.5EV. So they are not so lossless as you may think.
So why not use lossy compression of Bayer data at its original resolution?
So … Compressed HQ is now an option. I expect the algorithm is not publicly known but perhaps someone can make some good guesses — or share actual experience vs Lossless?
Imagemaster wrote:
Why can't someone post a side-by-side comparison of a 33mp to a 67mp image of the same subject with the subject the same size when posted? One could then judge for themselves whether or not the difference matters to them when posting or viewing images on this site.
Just pretend this is an example.
I think there are probably several GM and probably even some G lenses that will out-resolve the 33mp sensor and those would likely render somewhat better in a comparison as you show, at least using 100% crops from each. I’m not sure how the 1.4x linear scaling would impact that, though probably enough to make the difference pretty marginal if downsizing the larger to match the smaller. If both are full images downsized to say 4056px width and converted to sRGB, I doubt there’s going to be any visible difference.
I think the better comparison would be a really good lens over the 50, 60 and 67 mp sensors and 100% crops from them compared, all resized to match the 60. I doubt anyone could see any *resolution* differences.
Again, I agree there remain other reasons besides resolution to buy one of these cameras over the other.
Jack Flesher wrote:
I think there are probably several GM and probably even some G lenses that will out-resolve the 33mp sensor and those would likely render somewhat better in a comparison as you show, at least using 100% crops from each. I’m not sure how the 1.4x linear scaling would impact that, though probably enough to make the difference pretty marginal.
I think the better comparison would be a really good lens over the 50, 60 and 67 mp sensors and 100% crops from them compared, all resized to match the 60. I doubt anyone could see any *resolution* differences.
I really don't care what lens is used for the comparison, as long as it the same lens on each camera. And I don't care about seeing 100% crops.
Jul 04, 2026 at 09:24 AM
Steve Spencer Offline • • • • • • • Upload & Sell: On
Jack Flesher wrote:
I think there are probably several GM and probably even some G lenses that will out-resolve the 33mp sensor and those would likely render somewhat better in a comparison as you show, at least using 100% crops from each. I’m not sure how the 1.4x linear scaling would impact that, though probably enough to make the difference pretty marginal if downsizing the larger to match the smaller. If both are full images downsized to say 4056px width and converted to sRGB, I doubt there’s going to be any visible difference.
I think the better comparison would be a really good lens over the 50, 60 and 67 mp sensors and 100% crops from them compared, all resized to match the 60. I doubt anyone could see any *resolution* differences.
Again, I agree there remain other reasons besides resolution to buy one of these cameras over the other. ...Show more →
I think you should read this piece by Roger Cicala and Brandon Dube from lens rentals blog:
Two things are clear from their testing. First, there are a number of lenses that will resolve some detail at even very high resolutions. Second, it is hard to predict which lenses will be able to resolve at high resolutions from how they resolve at lower resolutions.
From their tests it looks like there are a lot of lenses, even though it is hard to predict which ones, that will have some improvement even with 100 MP sensors and of course new lenses could be designed that would be better at ultra high resolutions. There will be diminishing returns. We are already seeing that as we increase from 33MP to 67MP, but it is not as though there won't be any increase in performance. It might not be enough to matter to hardly anyone, but whether it matters will be a subjective decisions.
Finally let me quote Roger from the addendum to that article on system MTF and whether a camera out resolves a lens:
"System MTF = Camera MTF x Lens MTF. MTF maxes at 1.0 because 1.0 is perfect. So let’s say my camera MTF is 0.7, and my lens MTF is 0.7, then my system MTF is 0.49 (Lens MTF x Camera MTF). This is actually a pretty reasonable system.
Now, let’s say I get a much better camera with much higher resolution; the camera MTF is 0.9. The system MTF with the same lens also increases: 0.7 X 0.9 = 0.63. On the other hand, I could do the same thing if I bought a much better lens and kept it on the same camera. The camera basically never ‘out resolves the lens.’"
It is because system MTF works this way that I don't like talking about a lens out resolving a sensor. No sensor is perfect. No lens is perfect. A high resolution sensor helps both good and bad lenses. They just don't help bad lenses much. Conversely a good lens will look better than a bad lens even on a lower resolution camera, but the differences between a good lens and a bad lens will be exaggerated on a high resolution camera.
Two things are clear from their testing. First, there are a number of lenses that will resolve some detail at even very high resolutions. Second, it is hard to predict which lenses will be able to resolve at high resolutions from how they resolve at lower resolutions.
From their tests it looks like there are a lot of lenses, even though it is hard to predict which ones, that will have some improvement even with 100 MP sensors and of course new lenses could be designed that would be better at ultra high resolutions. There will be diminishing returns. We are already seeing that as we increase from 33MP to 67MP, but it is not as though there won't be any increase in performance. It might not be enough to matter to hardly anyone, but whether it matters will be a subjective decisions.
Finally let me quote Roger from the addendum to that article on system MTF and whether a camera out resolves a lens:
"System MTF = Camera MTF x Lens MTF. MTF maxes at 1.0 because 1.0 is perfect. So let’s say my camera MTF is 0.7, and my lens MTF is 0.7, then my system MTF is 0.49 (Lens MTF x Camera MTF). This is actually a pretty reasonable system.
Now, let’s say I get a much better camera with much higher resolution; the camera MTF is 0.9. The system MTF with the same lens also increases: 0.7 X 0.9 = 0.63. On the other hand, I could do the same thing if I bought a much better lens and kept it on the same camera. The camera basically never ‘out resolves the lens.’"
It is because system MTF works this way that I don't like talking about a lens out resolving a sensor. No sensor is perfect. No lens is perfect. A high resolution sensor helps both good and bad lenses. They just don't help bad lenses much. Conversely a good lens will look better than a bad lens even on a lower resolution camera, but the differences between a good lens and a bad lens will be exaggerated on a high resolution camera. ...Show more →
And as I’ve said as nausium, “show me”. All of that is nothing but theoretical conjecture that nobody has been able to show with an actual example, including Roger.
ruthenium wrote:
while I tend to agree with you on the "diminishing returns" of adding more pixels to a FF sensor.
In this subject, to put it simply: the returns diminish but don’t disappear.
Whether or not those decreasing returns make a significant difference depends on a bunch of variables. For many photographers they won’t make a difference; for some they may.
Jul 04, 2026 at 10:17 AM
Steve Spencer Offline • • • • • • • Upload & Sell: On
Jack Flesher wrote:
And as I’ve said as nausium, “show me”. All of that is nothing but theoretical conjecture that nobody has been able to show with an actual example, including Roger.
I am not sure what you want to see. Do you want to see a Sony 42 MP sensor compared to a Sony 60MP sensor? If so, see Fred's thread comparing the Sony A7r IIIa with the SonyA7r IV linked earlier in this thread. To me I can see the effect of the increased resolution. I wouldn't call the effect big, but it is there when viewed at the magnification level Fred provided. My own subjective evaluation is that the difference although perceptible isn't that important to me, but others can easily feel differently.
If you want to compare a FF 60 or 67 MP sensor to a FF 100 MP sensor, obviously no one can do that. So, I'm not sure what exactly it is you want to see that hasn't been shown and could be shown.
Douglas L wrote:
From yesterday. Fast FPS won't hurt for certain situation when you need it.
Great capture! Question per your exif — the image, at least to my eyes, doesn’t look anywhere near 1.7 stops under-exposed. Did you under-expose on purpose to then lift in post; and if this is a normal workflow practice for you, I’d be interested in knowing your reasons for doing it. Again, great shot!
Douglas L wrote:
From yesterday. Fast FPS won't hurt for certain situation when you need it.
Beautiful image, Douglas! Happy Fourth Of July to those who are celebrating it. Just got back from an enthusiastic community-oriented parade across the border in Kentucky, now getting ready to go celebrate with family, BBQ, fresh baked apple pie, and ice cold adult beverages 😎
