.
Those of you that have seen my deep zoom videos may have noticed that at a few points along the way the sequence "jerks" or stutters with distortion. This has nagged at me for years and I have pulled my hair out trying to figure it out. Turns out to be rather simple, and I have been kicking myself constantly for not having realized it sooner.
I have been in communication with Michael Condron, who runs the new
http://www.hpdz.net/ web site (you all should check out), and he has been comparing his home-grown software (which is apparently quite good) with fractint; we may collaborate on animations in the future if we can get our codes to match. He pointed out that fractint may not be jumping to a high enough level of precision soon enough when in arbitrary precision math mode.
I re-rendered a couple of the jerky sequences using the "BFDIGITS" command to force a higher precision, and yes - it works! They render perfectly smooth. So now I am re-doing the famous Universe-3 and the current Universe-6 with this correction.
Since fractint doesn't "know" I am rendering thousands of images for an animation, it has to be assumed that this problem can occur unpredictably at any point or time when a deep image is being rendered. Thus people rendering deep images may see a distorted oval where a circle should be etc., and would never know that it is a distortion - it would just look like another cool fractal!
I wouldn't call this a "bug" or "software issue", it seems to me that a small change in the fractint SC should fix it, maybe the number of a single variable. It looks like a whole lot of tight "C" and "ASM" code, mostly way over my humble head, but a fix might be:
I think correcting this in fractint may be as simple as changing a single variable in the code:
IF (magnification = xxx)
THEN (precision = xxx * c)
END IF
something like that.
Of course, higher precision means longer render time, so a balance must be struck. My observation so far suggests that a single order increment should fix it.
Stay Amazed!
JoTz
.
