Welcome to Fractal Forums

Can anyone point me to how to anti-alias in Mandelbulb3d? Sources, tutorials, etc.

Thanks,
Jim
Please email responses to jimi2tee@gmail.com

The AA in M3D is simply done using the Viewing arrows (1:1, 1:2, etc...) at the top of the main window. When you save as PNG or JPG, the downsized image is saved (WYSIWYG). It's as simple as that.

As bib wrote
----

my receipt to antialias (maybe somebody has a better routine ? )

1 Render larger then your final wished result (at least 50% bigger)

2 NEVER save as jpg but as png

3 apply a very slight gaussian blur ( from 0.5 to max 2.0 according to image and image size, i use 2 only for gigantic render, something less then 1 is usually fine...note that most of graphic editor allow to use decimals for gaussian blur BUT only by putting the number in the numerical box...sliders and arrows only change by 1 integer

4 scale down as required

I Render Large, save @1:1 in Png and do edge treatments by hand as Postwork in Photoshop before Resample/Resize and save for Web as Jpg. A true rendered  antialias would be good but not sure it's worth the extreme added render time, I like putting the personal touch on an image after the machine is done making it.

a good hint i think. i'd recommed a blur 0.3 - 0.5 for double size renders and 1 - 1.5 for tripple size renders. above that values you run the risk of blured downscaled images. in my observation this method works better, than the sometimes suggested gamma correction before downscaling.

i'm not sure as well, but a triple size render normally takes 9 times longer to finish...

Yes bigger values will blur
but the low value we indicated will not only minimize jagging but also give the sensation of more sharpness (because more clean edges)

And after all that as last touch to give even more sharpness is always possible apply Unsharp mask (that even if the name suggest the contrary is for sharpening )

NB perfectionist will apply unsharp mask only on the L channel in Lab color space to avoid any color shift

But to this there is a alternative that may sound more simple :
1 duplicate the layer and apply unsharp mask on the duplicate
2 change duplicate layer mode in Value (for gimp ) or Luminosity or Lightness (PS )
3 save the visible

Just wanted to add that for a downscale of 2 and 3, m3d does a very good anti-alias filter before saving in an image format.
It is a thresholded lancszos8 filter that prevents ringing and outperforms a simple gaussian blur a lot, in terms of aliasing and sharpness.

Of course, it is all up to your own preferences, i am only a bit proud of this technical aspect because it is something i developed myself.  :)
(If it is to sharp, bluring afterwards ishould be no problem)

I am wrong or render at 1:2 or 1:3 took exactly  the same time that render at full size ?

I do agree that in that case M3D does a pretty good job but ,maybe because i am used to bitmap editing and for me that is really quick to do manually (less then a minute for render) i prefer render big and downscale manually

(may be different for animations i did not try yet

maybe it's a matter of belief, but downscaling is allways the last step in my chain, so i prefer the image editor as well.
but for images straight "out of the box" mb does a really good job.

maybe you could implement an interface for photoshop plugins jesse  :dink: (don't take serious please)