END OF AN ERA, FRACTALFORUMS.COM IS CONTINUED ON FRACTALFORUMS.ORG

it was a great time but no longer maintainable by c.Kleinhuis contact him for any data retrieval,
thanks and see you perhaps in 10 years again

this forum will stay online for reference
News: Support us via Flattr FLATTR Link
 
*
Welcome, Guest. Please login or register. April 19, 2024, 11:36:39 PM


Login with username, password and session length


The All New FractalForums is now in Public Beta Testing! Visit FractalForums.org and check it out!


Pages: [1]   Go Down
  Print  
Share this topic on DiggShare this topic on FacebookShare this topic on GoogleShare this topic on RedditShare this topic on StumbleUponShare this topic on Twitter
Author Topic: Anti-Aliasing using the distance esitmation formula... good idea?  (Read 3451 times)
0 Members and 5 Guests are viewing this topic.
laser blaster
Iterator
*
Posts: 178


« on: March 29, 2013, 12:25:38 AM »

From what I can gather, supersampling seems to be the common method for anti-aliasing mandelbrot renders. But why not use the distance estimation formula to approximate the M-set coverage area within a pixel? If you wanted to do it right, you'd have to estimate the distance for both interior and exterior pixels. Sure, it wouldn't be exact, but I have a hunch it would give nice smooth results in a much faster time.

Has anyone tried this? Is it a good idea, or a bad one?
Logged
makc
Strange Attractor
***
Posts: 272



« Reply #1 on: March 29, 2013, 01:04:47 AM »

and if you use de-based coloring, you don't need aa at all smiley
Logged
laser blaster
Iterator
*
Posts: 178


« Reply #2 on: March 29, 2013, 04:04:51 AM »

I just implemented it into my humble little fractal explorer, and it works pretty dang well. It well exceeds the quality of 4xSSAA, yet it's nearly as fast as no AA at all. Right now, I color all the interior pixels black, and darken the exterior pixels based on their distance, which results in a slight thickening of the thin tendrils (which may or may not be desirable). I might try and implement more accurate AA using interior DE as well as exterior, but I'm not sure if it will be worth it... it depends on how complex the interior DE algorithm is.

The only problem is, this method doesn't AA the color bands, so I need to find a way to AA those too, without resorting to SSAA. And you're right, DE coloring would obviate the need for any AA at all, but I think the escape time algorithm looks much more interesting.
Logged
laser blaster
Iterator
*
Posts: 178


« Reply #3 on: March 29, 2013, 04:30:01 AM »

Ehh... now after zooming in and taking a closer look at the results, I'm not too crazy about this method.

The problem is that it underestimates the distance around long, thin tendrils, and especially the big needle sticking out of the left of the main cardioid, and it overestimates the distance in the valleys, especially the elephant valley. This leads to jagged valleys, and over-blurring near the needle (right now, since I've implemented exterior DE only, features on the needle have a 'hairy tarantula' look about them). It works great on the other areas of the M-set, though.

Now, this is a long shot, but does anyone know of a more accurate DE formula? I'm using exactly the one found here: http://www.ozone3d.net/tutorials/mandelbrot_set_p6.php

Otherwise, if I could find some sort of metric to differentiate valley-like areas from tendril areas from 'normal' areas, then I could easily come up with an approximate correction term for the DE. Any ideas?
« Last Edit: March 29, 2013, 04:33:58 AM by laser blaster » Logged
makc
Strange Attractor
***
Posts: 272



« Reply #4 on: March 29, 2013, 10:50:47 AM »

I need to find a way to AA those too
Clicky here

I think the escape time algorithm looks much more interesting.
You are not alone wink

it overestimates the distance in the valleys, especially the elephant valley.
Here are some de images, are you getting anything like that?
Logged
laser blaster
Iterator
*
Posts: 178


« Reply #5 on: March 29, 2013, 08:46:51 PM »

Yeah, my results look just like those images, but the accuracy is highly dependent on the area you're looking at. After thinking about it for a while, I don't think it's due to an inaccuracy with my DE formula. The problem is that distance is an imperfect metric for approximating pixel coverage in the first place. In areas with lots of thin tendrils, a pixel can have a very close distance to the M-set, yet have very little area coverage because the features are just so thin. In places with more compact geometry, like elephant valley, the opposite effect occurs.

It's a shame, because when this method works, it works really well. I tried to come up with a kludgy solution to correct for this inaccuracy, but it was slow and not worth a damn. So I guess the only solution is to adjust the distance threshold manually to best suit the area you're viewing. A slight inconvenience, which, in my opinion, is well worth the savings in rendering speed.

Once I finish implementing interior DE, I'll post some comparison pics.
Logged
Pages: [1]   Go Down
  Print  
 
Jump to:  

Related Topics
Subject Started by Replies Views Last post
Anti-aliasing for pro results Images Showcase (Rate My Fractal) twinbee 4 7383 Last post September 26, 2008, 06:59:21 PM
by twinbee
Anti-aliasing renders in Mandelbulb 3D Help & Support PrimitiveMind 2 806 Last post October 17, 2010, 07:03:00 PM
by PrimitiveMind
aliasing vs anti-aliasing Non-Fractal related Chit-Chat teamfresh 4 3357 Last post November 29, 2010, 08:20:27 PM
by teamfresh
Simple anti aliasing / supersampling help! Programming richardrosenman 6 2296 Last post April 15, 2011, 02:41:21 AM
by Softology
Anti-aliasing again Mandelbulber Ike1970 2 3648 Last post May 01, 2011, 06:20:40 PM
by Buddhi

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM
Page created in 0.154 seconds with 23 queries. (Pretty URLs adds 0.009s, 2q)