Originally posted by Aristophanes Show. Don't tell. Post photos with EXIF to demonstrate. Enough talk.
My dear friend. First comes the experiments (photos), then comes the theory (understanding). We already reached this point, we understand.
->
Falk Lumo: Camera equivalence
So, to debate your issue (and I don't totally disagree with your point of view), you'll have to discuss it in equivalence terms:
A 35mm/1.4 lens for APSC doesn't exist (it would be a 23mm/0.9). The closest thing for APS-C would be 24mm/1.4 (as exists for Nikon) which we can use on APSC too of course. Translating back to FF this means the following statement is true:
APSC forces you to shoot with a 36mm/2.1 equivalent lens rather than a 35/1.4.
Now, we have translated the discussion into known territory, no need to speculate.
Some (as yourself) would say that the difference between a 36/2.1 and a 35/1.4 isn't visible or of no interest. Fair enough and many may share your point of view.
Others though will say that it matters absolutely. It's the only explaination for f/1.2 lenses which have been made and costed a fortune. YMMV
Put please, discuss this in an educated style w/o listing tables of DoF values. It is ridiculous. We do
understand this stuff, don't we?
P.S.
One may argue that DoF is no topic with wide angle. However, this is not true. The defocus blur (characterized by the corresponding circle of confusion diameter coc) simply is:
coc_df = M^2/N depth
where M is the magnification, N the lens aperture FStop and depth (if small compared to the focus distance) is the distance in the real space from the plane in focus (or half the DoF).
And as you can see, the focal length is missing. Defocus blur does only depend on magnification and FStop. And magnification (e.g., for a 36cm face onto 36mm portrait or 1:10 which means 35cm distance with a 35mm lens!) is constant for a given subject, only the perpective is changing. As you can see too, depth decreases with M^2! So, even if the relative depth is kept constant (the subject becomes smaller overall), you still have to keep N propoprtional to 1/M. I.e., larger subjects can be shot with larger apertures w/o the "razor-thin DoF" issue. So, a general argument saying that a large aperture equals a razor-thin DoF is typically from people who simply don't understand.
BTW, this does not hold true for background blur (large depth). The background blur is
coc_bg = M f/N = M d
where f is the focal length (f/N is the physical lens aperture diameter d). So, while defocus blur does not depend on the focal length, subject isolation from the background does. This is another common source of confusion.
coc_bg / coc_df = 1/M f/depth
So, f >> M*depth is required to isolate a subject with a given depth against a background without risking to blur the subject itself. In the face example above (M=0.1 and depth = 35cm/2) we get f >> 17mm. Of course, the entire discussion is based on 35mm equivalent figures. Translate as required. Note that magnification M depends on the crop factor c, M = M' / c.