Archibald wrote in post #18371445
FF or crop is an interesting question for macro, especially when considering the 5DS with its high megapixels. I decided to do a (theoretical) comparison. To compare these formats, one has to decide on some premises, or things get real confusing. Here are my premises for this exercise.
- Crop vs FF
- Equal pixel pitch
- Same picture (equal framing, perspective, and DOF)
- Base case focal length of 100mm and magnification of 1 for the crop sensor.
In this comparison, the FF would need a focal length of 124mm and a magnification of 1.6x to match the crop picture. To get equal DOF, the FF would need f/13.8 vs f/11 (marked) for the crop body. As you suggest, the effective aperture would be a lot smaller for FF.
The FF would end up with almost double the Airy diameter, meaning that diffraction is worse and its effective resolution would be significantly less than for the crop case. Conclusion: the crop format (for the stated premises) has a significant advantage.
That one had me head-scratching for a while...
CASE 1: With FF if I mount a 100mm lens and get a 1:1 image, I frame a subject area which is 24mm x 36mm ; with APS-C if I mount a 100mm lens and get a 1:1 image, I frame a subject area which is 15mm x 22.5mm ; if both cameras had same pixel count (e.g. 4000 pixels) vertically, both cameras would put the same number of pixels on the same subject area, so resolution is identical! Erroneous analysis removed...
For example, if my subject was 15mm tall, at 1:1 in FF or APS-C the subject is IDENTICAL in size on the sensor at the same pixel pitch, and only the total area (around the subject) is larger on FF.
With FF if I mount a 100mm lens and get a 1:1 (1X) image, I frame a subject area which is 24mm x 36mm ; with APS-C if I mount a 100mm lens and I frame the subject area (which is 24mm x 36mm) onto a 15mm x 22.5mm area sensor, the subject is imaged at 0.6X.
If both cameras had same 4000 pixels vertically, both cameras would put the same number of pixels on the same image area, so subject resolution ought to be identical, when considering the sensor alone, but this is NOT considering 'the whole picture' so to speak...!
But the lens is NOT delivering the same line-pairs per millimeter at subject magnification at 0.6X on APS-C as it does at 1X to the FF sensor. A lens which delivers 100 l-p/mm on the sensor delivers 2400 l-p total resolution to FF sensor, but delivers only 1500 l-p total resolution to the APS-C sensor. With with equal subject area in each frame, the FF sensor is delivering 60% more resolution even though both sensors capture with the same 4000 pixels vertically in the shot.
...of 2400 line-pairs delivered to 4000 pxiels vs. 1500 line-pairs delivered to 4000 pixels; FF wins.
CASE 2: With FF if I mount a 100mm lens and get a 1:1 image, I frame a subject area which is 24mm x 36mm ; with APS-C if I mount a 100mm lens, and I frame a subject area which is 24mm x 36mm, and get a 0.6:1 image; if both cameras had sensors with same pixel pitch, let's assume 240 pixels per millimeter...
With the FF camera I image a 24mm tall subject at full size at 1X, with 240 pixels/millimeter, the subject is 5760 pixels tall; with the APS-C camera I image the 24mm tall subject at 0.6X into a 15mm tall sensor space, so the subject is only 3600 pixels tall. Subject resolution is 60% better with the FF image than with the APS-C image, considering only pixel count.
Bring lens resolution back into the discussion...if I have 100 line-pairs/millimeter at the sensor, it delivers 2400 l-p total resolution to FF sensor, but delivers only 1500 l-p total resolution to the APS-C sensor.
2400 l-p projected to 5760 pixels on FF vs. 1500 l-p projected to 3500 pixels on APS-C...2.4 pixels per line-pair on FF vs. 2.33 pixels per line-pair on APS-C; FF wins (barely)
My results are the complete opposite from Archibald's analysis. Somebody else will need to come up with the tie-breaker analysis.