Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

3D M sets in Fractint: the *real* solution

2 views
Skip to first unread message

David M. Read

unread,
Jun 25, 1991, 6:09:40 PM6/25/91
to
I've gotten dozens of replies to my question concerning how to produce
Beauty-of-Fractals-Cover-type mandelbrot Mesas, and they all include
basically the same idea: set inside color = maxiter in the (x) menu
before generating the image. I've tried it, and it works beautifully.

Someone also pointed out to me today that the best way to do this is
to try setting the ``continuous potential'' parameters in the (y)
menu...so I went & read the manual section on that, and gave it a try.

It works much better! The range of slowly-varying colors gives the 3D
pictures a great depth that they don't have without it.

The drawback is that it's kinda slow; a high bailout value (required
to get the range) forces one to use the flotating point algorithm. Of
course, if you were using the FP algorithm already (i.e. you've zoomed
way in), you'll never notice it!

Tim (if you're listening): can the integer math section be written so
that the 127 bailout limit is raised? Even if it were something like
255 or 511, it would be much better than 127!

Thanks to all the people who gave me those tips! (and extra thanks to
the fractint authors for providing the continuous potential option...
I love it!)

--
Dave Read (rea...@ccwf.cc.utexas.edu) | In large, friendly
UT-Austin Nuclear Physics Graduate Student (Slave) | letters were the words
Sometimes I wish life were a game of Asteroids... | Don't Panic.
So I could hit the Hyperspace button. |

Timothy Wegner

unread,
Jun 27, 1991, 9:10:28 AM6/27/91
to
rea...@ccwf.cc.utexas.edu (David M. Read) writes:

>Tim (if you're listening): can the integer math section be written so
>that the 127 bailout limit is raised? Even if it were something like
>255 or 511, it would be much better than 127!

I haven't looked at the continuous potential code for a long time. I believe
that continuous potential uses 32 bit integers shifted 16 bits, so that
should leave plenty of room. I'll check it out - that may be a left-over
limitation. However I am warning you that there is an exponential in the
continuous potential formula that may be the reason we set it where we did.

Note that the price of coprocessors just fell precipitously. A 33 mhz 80387
is now $250 with the clones around $200. That is probably the way to go!
Frankly I think continuous potential needs bailouts like 1000 or 2000 so
I would recommend floating point rather than integer.

0 new messages