Basically, let's say I have a model with multiple top surface areas - one of them is at Z=1mm, one at Z=1.25mm, one at Z=1.7mm, and one at Z=2.5mm. My layer height value is set to .5mm.
Currently, i'm not entirely clear on what S3D does in these instances - does it round down, so now I have my 1.25mm top surface at 1mm? Is it rounded up, to 1.5mm? What about the 1.7mm top surface? Lots of vagueness to how this is handled.
However, I have a dual extruder, and I've done experiments doing dual-material prints with each extruder running at different layer height values (usually easily divisible, like .25mm and .5mm, or .2mm and .1mm). S3D handles this like a champ, giving me an extra layer on the smaller layer height compared to the larger one (IE, E1 gets a layer, then E2 gets a layer. Z axis moves up, then E2 gets another layer, then it goes up again and both get another layer).
So, why not make a slicer mode where, just for the sections that aren't easily divisible by the layer height, it 'adds' a higher resolution section?
To use the earlier example, with a .5mm layer height we'd get two layers until 1mm, then we'd get a sub-layer - just for the section that needs it - to give the .25mm top layer. We'd get another .5mm layer everywhere else, then we'd get a .2mm sub layer to cover the .7mm top surface. Then, we go on up to 2mm, then 2.5mm.
Now I've got a model that is extremely accurate in the Z-axis, in terms of measurements, while still keeping the speed that the higher layer size (.5mm) gives me. Otherwise, to truly hit all the points necessary on my imaginary model, I'd have to either use a .05mm layer height (eww), make multiple processes to catch everything (in this case, a .5mm process for two, a .25mm process for two layers, a .1mm process for 5 layers (or a .2mm process for two layers and a .3mm process for one layer), and then another .5mm process just for the last layer. That's 4-5 process just for a single model, and it's a pain in the ass to keep track of exactly where I need to split up my layer heights! On a highly complex model it would be nearly impossible to set up all the processes necessary for it.
Granted, this new process could get very complicated (and CPU intensive) very fast on organic models with a lot of detail - you'd want to have some sort of constraining functions ('go no smaller than a .1mm layer height' 'only applies to areas larger than X millimeters' etc) just so that it doesn't go totally off the rails trying to compute everything.
Anyway, this got long. I would love to see something like this in ANY slicer - but S3D is my slicer of choice so I'd much rather see it in this one.
