Thought I would do a test today using some of Canon's goto Portrait lens and try to determine exactly what SS would be required to achieve a very high probability of achieving a 'tack sharp' image using the 5DSR handheld.
For most people the 70-200 IS is one of those lens but today I decided to use just the primes, in the standard portrait range. It is probably a fairly good assumption that the 100L with IS is a good proxy for what the 70-200II IS would require.
The setup is I went out to my backyard and brought an envelope with small type face writing on the back and placed it next to a spray paint bottle, also with writing, and then stopped back 15 feet and snapped off 2 pictures at each of the SS I wanted to test.
The SS tested were: 1/60, 1/80, 1/125, 1/200, 1/250, 1/320, 1/400, 1/500, 1/1000
I used Aperture F4 - each of these lens (85L, 100L, 135L) is tack tack sharp in the center at F4.
The test was done quickly with near identical lighting but the camera was placed on Auto ISO so it determined the exposure. Center point focus was used. Below I have shown one of the pre-cropped pictures to give you and idea of the setup and then I did massive 2:1 LR crops to show just the writing on the labels. Because of framing differences between the focal lengths of 85,100,135 there are different number of pixels in each crop but what I'm really trying to determine was what is the minimum SS so there is no question you can get tack sharp.
Being a math guy, I know that 2 handheld shots is no where near enough to give you a high degree of confidence, especially with non IS lenses (with IS lens so far I have found that if one is good nearly all will be good with a very high certainty) but that is just not true with non IS lenses.
The punch line is that with IS lens I think you can comfortably use 1/ 2/3rds FL - so 60th / 80th is good for the 100L (and very likely the 70-200 even at 200mm). I have lots of other experience with the 100-400L II IS and 1/60th is perfectly good at 400mm as well (with a very high hit rate).
The other thing that I saw looking at the various snaps with the 85 & 135L is that just because you might get one of them almost tack sharp at 1/80th you can then get 2 not nearly tack sharp pictures at 1/200th - so this is the frustration with these lenses handheld and taking a single shot and thinking you have it when you actually don't.
The final thing I will say is that when I looked very closely at the images, there is sharp and there is tack sharp when you are pixel peeping and for most images at reasonable distances there will be no telling that difference so using 1/3X FL you can be pretty certain that it is going to look sharp (90% of the time) but it you start to zoom in that is when you really see that it isn't tack sharp and 1/4x or 1/5x or higher is really need for non IS lenses (handheld) to really nail it with high consistency.
I'm left wondering how a high pixel density stabilized body such as the upcoming A7RII would do with IBIS. Without the lens being stabilized I don't think you will get the typical 3 stops of stabilization but could it get 2 stops?
The setup
100L at 1/60IMAGE LINK: https://flic.kr/p/v9Fdnv
85L @ 1/60IMAGE LINK: https://flic.kr/p/w6XzEX
85L @ 1/200IMAGE LINK: https://flic.kr/p/w6xvXD
85L @ 1/320IMAGE LINK: https://flic.kr/p/vNVTsf
135L @ 1/125 (not sharp)IMAGE LINK: https://flic.kr/p/vP47rB
135L @ 1/250 (not sharp)IMAGE LINK: https://flic.kr/p/v9wCRQ
135L @ 1/500 (not quite - but 1/400 was a little better)IMAGE LINK: https://flic.kr/p/vNVP1C









