I don't recall this in earlier versions, but I've upgraded to SR4 with Poser Pro 2014.  Now when I do some renders, and use DoF, the focus has been off.

Typically I'd select the head, then use "Calc DoF Focal Distance."  Used to work like a charm.  Then I noticed blurry eyes.  So I tried selecting the eyes.  Still blurry.

Does it calculate off of the closest point to the camera, or center of mass?  Or am I just copy and pasting wrong?

It is also F-Stop dependent, the higher the number, the greater the Depth of Field.  

Right.  And that's where I noticed it, with a portrait shot with a low F-stop.  I think it may have been 2.8.  The eyes were a little blurry, but the neck and ear clear, which in real units would be a few inches behind.

Personally. f-2.8 is too shallow.  I would prefer f-5.6 or f-8 to increase the depth of sharpness, perhaps focusing on the bridge of the nose, if not the eyes themselves.

Yeah, I typically don't use such a shallow F-Stop myself.  That was an extreme that I noticed.  F-5.6 or F8 or so is where I usually roam as well.

But for the focus, the center point is often not where I want to focus, so I use the calculation.  I don't want to move the camera and mess up with the manual slider.

Where is this "Cal DoF Focal Distance" you "use"? Is it a button - a menu item? Where is it?

Obviously it is something I don't use, since I don't have any idea what you're referring to.

I simply grab the focal_Distance parameter dial (left mouse down and hold) and start moving my mouse. A cross-hair appears (the Focus Distance Guide) in the preview window which moves as I move the mouse. When the cross-hair intersects what I want in focus, I let go of the mouse button. 

Then it renders in focus exactly where I left the cross-hair.

( You can also permanently show the Focus Distance Guide by the menu item Display / Guides / Focus Distance Guide. )

@BB: It's in the Python script panel under Render / IO. Also, I don't know if it's just me, but it gives the results in feet, regardless of what your preferences are set to use.

Those default crosshairs can be a pain. If my camera isn't focussed directly onto a characters head, but rather aimed at the body, I can never tell what portion of the face will be in focus. I found myself using a script by Semidieu to put visible planes on the camera that I can work with.

Quote - @BB: It's in the Python script panel under Render / IO. Also, I don't know if it's just me, but it gives the results in feet, regardless of what your preferences are set to use.

Those default crosshairs can be a pain. If my camera isn't focussed directly onto a characters head, but rather aimed at the body, I can never tell what portion of the face will be in focus. I found myself using a script by Semidieu to put visible planes on the camera that I can work with.

Yup, that's exactly why I use it.  If the face isn't in the crosshairs, it's difficult.

imo it's more advantagous to use the z-depth render pass in addition to the focus cross hair, that way you can dial in the amount of blur in post.

I wish that were true.  I have been playing around with it, and will use it when I can.  It fails when there is transparency, like a portrait of hair. 

I'll use it when I can.  It does shorten render times, and lets you set the focus distance and amount of blur on the fly which is very handy.