(Second part of my reply; previous page response is to unfocused.)
TBH ISO 400 really shouldn't matter, most modern cameras exceed in quality up to about ISO 600 and are extremely useable.
I realize that doesn't pass the usual sanity check, but for those conditions, that was my impression. Today (in fluctuating conditions, from very sunny to quite overcast within seconds, and back) I had the impression that moving from ISO 100 to 400 (partly to help achieve a faster shutter speed, which was critical) was of little apparent impact. I suppose I'll look at the 100% versions here later.
Also if you are shooting with a TS-E 90mm, at 1:1 or larger magnification you really need to be shooting at a smaller F stop, 2.8 is far too open at these distances you will be lucky if you got 3% of the image in focus, I would say the minimum F number I would use is F5.6 which will give a nicer effect and still keep a pleasing bokeh.
I think this is something that I can judge best in the field. Some insects are actually quite flat, or there is some particular part of them that is most interesting (and sometimes both are true). Others that have more protruding or deep elements, such as spiders, of course do benefit more from such a technique, but you should not forget the artist's tool of selective focus. That f/2.8 aperture allows me, if I manage it properly (difficult handheld, but that's what the continuous shooting mode is for), to draw the eye to a most interesting part of the image, and blur the parts that are less interesting (especially if, as in the hibiscus situation earlier, some parts of it might wish to be less defined for modesty, as on a personal portrait).
One of my recent challenges was to see if a picture of a butterfly could be taken at a wide f-stop while keeping the background as blurred as possible while not losing too much of the detail on the wing due to a misalignment between the plane of that wing and the zone of focus. Unfortunately that's easier said than done, but it was an interesting experiment. Too bad the tilt function is less effective at close distances. On the plus side, nobody has to stare at all the bits of sand or cracks in the rocks in my driveway where the butterfly perched - instead the attention goes (in the better image of the bunch I had) to the proboscis which was curled slightly and laid on a particular rock.
The TS-E 90mm's native maximum magnification is .29X, or a ratio of about 1:3.45. The 2X does improve this specification, but at some cost. Defocus effects are a primary concern; I don't see a reason to cripple my low-light gathering performance and the boke unless I need depth of field, and even then, carefully manually focusing using Live View is the best bet. As I read somebody read recently: "I don't believe in good bokeh, only bad bokeh." A bit pessimistic, but you start with good performance and this characteristic generally gets worse (even on the TS-E which I find has some slight but definite color fringing in unusual circumstances, but not normally).
you can get some nice effects with 2.8 if you are going to use the images to paint or other uses, but from a visual standpoint there isnt enough of the image in focus to keep the viewer interested in the image.
I find it amusing that you are critiquing an image you haven't ever seen. I needn't say much more than that, I hope!
But just because a lens is a 2.8 doesnt mean it is technically right to use it wide open.Au contraire, mon ami
, there is no other reason to create a manual focus lens like the TS-E with a relatively wide maximum aperture other than to allow its use wide open. As a result there is no problem using it wide open, either. Certainly there are tradeoffs either way, but I find that selecting the right aperture is not merely "FIGHT FOR BEST SHARPNESS!" Even adding in chromatic abberations, vignetting, and other lens characteristics that get helped out by smaller apertures, it is more complicated than that.
And of course Ansel Adams has weighed in as well: "Any good modern lens is corrected for maximum definition at the larger stops. Using a small stop only increases depth..."
Please do be advised that I am working on an APS-C sensor, so the depth of field appears to be greater (as a share of the cropped frame) and so DOF appears to be where it might on the full format sensor at around f/4 or so. It is still very small - no larger or smaller than it would have been, and so still challenging to use at these close distances and a wide aperture.
Among my random points, one stands out to me: For some reason I can't fathom, my Live View often gives a misleadingly sharp image compared to the output, even at magnification levels where I would have assumed that fake sharpness due to pixel skipping would have been the culprit. I haven't been able to figure this out entirely, other than to rule out motion blur. Then again, maybe it is simply pixel skipping (I'd expect moire to be a factor in Live View if that were the case, however).
With macro photography you shouldn't use an overall sharpening technique. There is no point in sharpening areas which are out of focus (or bokeh) because it will degrade the image and add unnecessary noise. Your much better off using a selective sharpening technique using a mask to paint in your sharpening. This will not only excentuate the part of the image you want the viewer to concentrate on it will also add more depth and clarity and increase the overall visual effect. [more clipped]
I haven't been thinking about adding artificial sharpening (I think I mentioned that, but maybe not), but this is a helpful reminder about what that lasso tool / High Pass is for, thanks!