Firstly, I know I don't have the ultimate light set up here, but it works well enough for my purposes and for my tests here....

Now to summarize some things I've learned from this:

Min shading rate is the free standing render setting variable that is a serious factor in determining render times. This one is unquestionable – going from 1.0 down to 0.50 doubled render time for the wild hair, and caused similar time increases in all my other portrait tests. A smaller size definitely improved image quality – but by how much? You'll have to decide for yourselves what you can live with, but right now I'm satisfied that 1.0 is good enough for most work I'll be doing at this range.

Now I'm light on reflection/refraction in these tests – only the corneas and eyeballs have any – so number of ray traces could become a bigger factor if there's lots of that kind of thing going on. I usually don't have much of it, so I guess I'll deal with it later if I need to.

Pixel samples didn't become really important until I moved out of close ups and into the main camera shots. Again you'll have to decide for yourselves what you'll be happy with, and it will also depend on what distance you're shooting from, but going with 4 for non-portraits gives decent results – it seemed to take very little longer to render the scene going from 2 pixel samples to 4 pixel samples. For portraits with the face camera 2 pixel samples worked fine for me – when you're in tight antialiasing isn't as big of an issue.

Bucket size is nearly as important in regards to render settings, however it is dependent on other things – anything which impacts the amount of memory available to be exact. Shrinking texture sizes may help here if you are using many different large size image maps in the render. The more memory you can free up, the larger bucket size you can use and the quicker your render will be. One thing I've verified is that for me Poser no longer chokes if I set too high a bucket size. It seems to automatically adjust to a smaller bucket size which will allow it to complete the render. For instance when I set the max bucket size at 1024 it never actually used that buffer size – I think it reverted down to 256. Also when I had the bucket size at 256, it stayed at 256 when I used smaller texture maps/a lower maximum texture size setting. But when I upped the texture sizes and the maximum texture size setting I could see the bucket size ended up being decreased – I think down to 128 – even though I had left the max setting at 256. What this means is you can probably leave your max bucket size pretty high and Poser will adjust down from there to complete the render if it needs to. As I mentioned at the top I'm on SR2 for Poser 6 – so this probably doesn't work in previous releases as I recall getting the magic out of memory popup before when I'd set the bucket size too high.

Maximum Texture size is a tricky thing. You can definitely notice a difference in the quality of the render in terms of reduced detail and sharpness when the textures get reduced – regardless of whether you reduce them by hand yourself prior to rendering or you set your Maximum Texture Size lower than the textures actually are. The render time improvement obtained by reducing texture size was negligible in my testing. This may be due to me not using many different figures/props with lots of textures, but the wild hair itself is a pretty heavy object. I'm leaning towards thinking texture reduction isn't a real player in reducing render time EXCEPT in the case that your textures are causing your bucket size to shrink to unacceptable levels. As mentioned above I've got about 1.5 GB RAM, so I didn't run into this problem. My bucket size never dipped below 128. Shrinking textures prior to rendering, rather than having poser do it, probably saves you anywhere between 30 seconds to a minute per texture – this savings is seen in a faster “loading textures” time and that's about it. For my money the loss of detail, clarity and sharpness is not worth reducing the textures – at least I haven't run into a situation that it would be worth it yet.

In terms of hair, the biggest time killer of all are hairs that use complex material set ups combined with elaborate transparency/alpha maps. The Wild hair has 4 materials used in rendering, the Long Hair Evolution has 3 materials used. However for each material on the Wild hair there are three maps used and one specular node.The Long Hair Evolution materials have 2 maps each and no nodes. Eyeballing the transparency/alpha maps for the two hairs in question, it seems to me that there's quite a bit more and quite a bit more detailled mapping going on with the Wild Hair. I haven't delved deeply into the material settings for the Wild hair, but it seems to me that the extra work required to avoid painted highlights, replacing them with procedural highlights along with the more complex transparency maps cause the user to pay a very heavy price in terms of render time. The Wild hair definitely renders more realistically when everything is on – but is the extra realism worth the extra hour of render time? I know I'll probably only use the Wild hair with it's full material setup only on very special occasions. The lesson here is to pay close attention to the complexity of the shaders you use when there's a busy transparency map happening – and hair is the prime culprit in this area, so use your hair figures and props carefully.

I hope some of this proves useful/educational to some folks around here, and of course I'd enjoy hearing what others have to contribute as well.

Unzipped