APS-C lenses tested on an APS-C and a fullframe. Doubt

[Alejandro]

I tested several APS-C lenses mounted on an APS-C camera (Sony A580 with 16.20 Mpix sensor) and with those same lenses on a full-frame camera (Sony A7 ii with 24.24 Mpix) as a result of which the images, when using a sector reduced to APS-C of the sensor with the last one, ended up being 10.32 Mpix. So far, everything as expected.

The RAW files of the images of the same photographed object taken with the A580 weighed 16.1 Mb, while the images taken with the A7 ii (with lower resolution in Mpix) weighed more, 20.7 Mb (and I quote that if the photos had they been taken at full-frame they would have weighed 47.1 Mb). Regarding this result, I have my theories (although I do not rule out that they could be wrong), the main one being that the files obtained with the A7 ii probably contain a lot of information about each shot, such as a greater dynamic range.

In order to better be able to evaluate possible differences in the resolution of one and the other photos, I put together a row and pile of boxes of tea and remedies full of very small legends and bar codes. When opening the photos taken with the same lens to the size of the monitor screen, they all covered the same portion of the photographed object(s). Then I compared the images by zooming in, and also zooming in a bit more to the photos taken with the A7 ii (to also match their screen sizes to those of the A580, due to the difference in Megapixels between each other images). It turned out that photos taken in APS-C on the A7 ii looked higher in resolution/sharpness than those taken with the same lens on the A580 (which is APS-C), particularly those taken at larger apertures and more noticeably towards the edges.
Is this attributable to the fact that there may be a difference in the quality of the sensors? Greater micro-contrast? Can the size of the photodiodes influence this?

I googled to see if I could find any references or articles on the subject, but I couldn't find anything. Maybe someone knows the answers or can give an opinion on the subject.

 

(PS. Sorry for my English, it is not my native language)

[Pieter]

I believe this has to do with the fact that the A580 outputs 12-bits RAW files whereas the A7ii produces 14-bits RAW files. The A7ii images therefore containt much more color depth and may indeed look more detailed.

[Alejandro]

7 hours ago, Pieter said:

I believe this has to do with the fact that the A580 outputs 12-bits RAW files whereas the A7ii produces 14-bits RAW files. The A7ii images therefore containt much more color depth and may indeed look more detailed.

Thank you Pieter, what you say makes sense.

Because of what you mentioned, I did a bit of further information searching. It seems that the Sony a7 ii produces 14 bits Raw files if shooting Single Shot (not in Continuous mode in which case the output would be a 12 bits depth file). Also, that newer Sony A7 models can produce 14 bits Raw files as well when shooting in Continuous mode if the camera is set to Uncompressed Raw. With the Sony A7 ii I did the referred tests, shooting Single Shot in Uncompressed mode. I checked that, as you said, the A580 produces 12 bits Raw files, which I think/suspect are in compressed mode.

These finding encourage me to use two of my APS-C lenses on the a7 ii. The Sony DT 55-300mm f/4.5-5.6 SAM tele (while lacking a better long telephoto zoom lens), and occasionally the Sony DT 35mm f/1.8 SAM lens (wich I think might be useful for some amateurish macro if coupling on to it a Reynox macro lens that I have, and for filming video with it in APS-C mode). 10.32 Mpix are more than enough for media display and small prints if not needing to heavily crop a photo, and the resulting images taken with the a7 ii seem to look better than if taken with the A580 if not unnecessarily zooming into an image (because of the lesser count of megapixels that is) and have more dynamic range. The a7 ii is also more comfortable than the A580, and full of nice and useful features .

Anyway, I am open to suggestions and/or differing opinions.

Edited March 3, 2023 by Alejandro