Print Page - Mandelbrot 3D slope tests

Fractal Art => Movies Showcase (Rate My Movie) => Topic started by: Kalles Fraktaler on December 05, 2014, 08:59:54 PM

Title: Mandelbrot 3D slope tests
Post by: Kalles Fraktaler on December 05, 2014, 08:59:54 PM

Maybe all of you have seen this many times before

https://www.youtube.com/watch?v=ztm5YtY3thE

Zooming is not easy because since the distance between two pixels decreases per frame, the slope tilt also decreases.

https://www.youtube.com/watch?v=xbzUWGTxIek

Title: Re: Mandelbrot 3D slope tests
Post by: TheRedshiftRider on December 05, 2014, 09:20:30 PM

It looks great.

Title: Re: Mandelbrot 3D slope tests
Post by: kubinator4321 on December 05, 2014, 09:42:34 PM

:verysurprised:

Seriously, this zoom looked great. I guess I underestimated the iteration-based sloping. Great job in making that.

Title: Re: Mandelbrot 3D slope tests
Post by: superheal on December 06, 2014, 02:54:35 PM

That's really nice!

Can you share the algorithm you used to create that effect?!

Title: Re: Mandelbrot 3D slope tests
Post by: Chillheimer on December 06, 2014, 06:03:02 PM

very, very cool!
I really look forward to the next version!

Title: Re: Mandelbrot 3D slope tests
Post by: Kalles Fraktaler on December 06, 2014, 08:15:09 PM

Quote from: superheal on December 06, 2014, 02:54:35 PM

That's really nice!

Can you share the algorithm you used to create that effect?!

Thanks

Here is the algorithm, it normalizes the quote of the nearby pixels with atan.

Code:

		// put the ratio of the current pixel p1 with the next p2 in the variable diff
		diff = p1/p2;
		// raise this value with nSlopePower
		diff = diff^nSlopePower
		if(diff>1){
			// make the diff value between 0 and 1 
			diff = (atan(diff)-pi/4)/(pi/4);
			color = (1-diff)*color + diff*black
		}
		else{
			diff=1/diff;
			diff = (atan(diff)-pi/4)/(pi/4);
			color = (1-diff)*color + diff*white;
		}

However it is not suitable for zoom Movies, because if the joints should not be visible, the nSlopePower needs to grow exponetially as a zoom goes deeper, actually the same value as the magnification. This is since the slope is calculated from nearby pixels, the distance between nearby pixels is decreasing exponentially. It get ridiculus already at 1e5 eliminating all gradients on the shadows.

Quote from: Chillheimer on December 06, 2014, 06:03:02 PM

very, very cool!
I really look forward to the next version!

If it is OK for still images only, I can add it right away.
The rotating zoom movie I created by render frames for each Movie frames, and used my first movie assembler program, if you remember that one, from jpeg images...

Title: Re: Mandelbrot 3D slope tests
Post by: mclarekin on December 07, 2014, 03:52:50 AM

Impressive. That slope test2 maybe the best I've ever seen :) :) :).

Welcome to Fractal Forums

Fractal Art => Movies Showcase (Rate My Movie) => Topic started by: Kalles Fraktaler on December 05, 2014, 08:59:54 PM