This is really sad news, and my thoughts are with his family and friends.

The Superfly render engine has changed a lot under the hood. In P12 one thread would render one bucket, so you'd set the number of buckets to match the threads available for CPU rendering. However in P13 that has changed - all threads render one single bucket at a time.

Two options - either increase the size of the bucket (more threads = bigger bucket, though RAM will play a role too) or switch Progressive Rendering on which will then ignore the bucket setting.

The Log will give you an exact time for each render to compare different settings.

In addition, I've found that P13 needs less samples than P12, I'd suggest try halving the setting first. With both P12 and P13 installed you should be able to do some comparisons. Also the Denoise post-process option is imo better than P12.

Basically I've found that P13 is significantly faster for me.

Bug report submitted - pix/in setting not being saved with image file. 'Tis a very minor bug - if you set the physical size and resolution it will render the exact pixels required; it's just that pix/in is always set to 72dpi in the exported file. (I set a render size of 6x4" at 300pix/in which is 1800x1200 pixels, but in Photoshop the file is 72pix/in so the dimensions are 25x16.667" - still 1800x1200 pixels and easy enough to change, but that setting should carry over as it used to).

So to agree with and back up JoEtzoid and ChromeStar, that resolution setting has nothing to do with rendering. A pixel is just an RGB number, so pix/in is meaningless UNLESS you specify a physical size.

In Poser you EITHER set a physical size and resolution in the render dimensions dialog OR you set the number of pixels.

If you set the number of pixels e.g. 640x480 and then change the pix/in value it will be ignored UNLESS you change dimensions from pixels to inches or cm.

@igohigh - you need to set the render size in either inches or cm for changing the dpi to have any effect. 

Resolution when rendering is controlled by the total number of pixels in the image.

In the example screens you've posted, the image has the same resolution in both 11 & 12 - 640 x 480 pixels.

That number of pixels can be printed at either 8.888 x 6.666 inches at 72 dpi, or 2.133 x 1.6 inches at 300 dpi.

If you want to print at 8.888 x 6.666 inches at 300 dpi then you need to set that size in the render size dialog and let Poser figure out how many pixels are needed (it's 2666 x 2000, according to Photoshop).

So if you tell Poser what print size and resolution you need it will calculate the number of pixels needed to achieve that. 

If you tell it how many pixels you want you'll have to sort the print size and resolution yourself.

DPI is not rendered, only pixels are rendered.

The Poser bundled version with leotard - everything bar the shader is default.

If you use Photoshop's Image Size dialog, uncheck Resample, and change the dpi to whatever you want - the dimensions will change but the pixel count will remain the same.

Looks very good to me, though I'd suggest trying a render using a single light source behind the figure too - I have found it useful using the ears to judge subsurface amount.

@hborre that would make sense. AFAIK radius is depth split into RGB values and the subsurface value is an overall multiplier, so higher RGB depth + lower multiplier should look similar to lower RGB depth + higher multiplier. I'll try your values when I fire Poser up next.

@primorge SF vs Blender Cycles is an easy answer for me - I don't use Blender, therefore Superfly always wins :D

Animation is an interesting question - personally I'd look at the number of frames I would need vs computing power available, then figure out a time budget per frame. That would determine the complexity of the materials, and that would influence my choice of engine.

Rendered with 40 samples, single area light at 100% plus jpg textures on the Poser Ground for bounced light. Raw renders have had minor sharpen, contrast and colour adjust in Photoshop - after all, Blender and Iray and other 3d apps use 2d post-processing to optimise their results, so it seems only fair. This is about an hour of tweaks and test renders.

This is a quickie setup for Principled BSDF with La Femme. Highlighted nodes are new additions to the default LF shader. The Specular map is not used as I'm not good with math nodes - I took the original image into Photoshop, inverted it and used Levels to crush the blacks to a more reasonable value. Roughness = the nearer to black the sharper, brighter and more mirror-like the highlights will be. There is of course no such thing as Specular in Superfly, it's all raytraced reflections.

The SSS Method at the base of the PBSDF is set to Burley by default. Random Walk is better as it calculates the actual volume of the mesh so it's the most accurate SSS calculation method available in Superfly. Only the SkinHead and SkinTorso mat zones were altered, but I did change the Scatter Group ID for the eye materials to 3 to get rid of the blue cast from the intersecting meshes.

There is a bug report in for this issue, has been for some time. It may be related to the issue with higher-res morphs, it's something to do with accuracy of vertex positions. Basically it's not a trivial problem, hence no fix yet. Please drop a bug report in via posersoftware.com if you can as the more feedback the devs get the better, thanks.

What you need is a little X-ray vision :D

Shoulders are very hard, there are 4 joints to emulate. The improvement to that collar bend is really good.

Found my thread in case it's useful - examples of Principled BSDF here - it's ridiculously useful, and simple to set up.

The clear stones in the render use the shader below, which I attempted to duplicate by following this example.

Gemstones have a base reflectance of 16%, so Specular is set to 2 and I set Transmission to 1. I've played around with varying the colour, roughness and IOR (there are many examples online, here is one). I also tried adding a little Transmission Roughness to a couple of stones - to enable that the Distribution must be changed to GGX instead of the default Multiscatter GGX.

The mesh is a primitive from Lightwave, there is a single infinite light and I've used a jpg panoramic sky plus simple reflective surface for the Ground.