Path: chuka.playstation.co.uk!scea!peter_alau@playstation.sony.com
From: shade@dragonshadow.com (Scott Cartier)
Newsgroups: scee.yaroze.programming.2d_graphics
Subject: Re: What do'you mean I have to brush up on me maths!
Date: Mon, 12 Jul 1999 16:53:50 GMT
Organization: SCEA News Server
Lines: 26
Message-ID: <378a1a59.1200158692@news.scea.sony.com>
References: <7ko1jf$33u25@chuka.playstation.co.uk> <376fbd21.251404588@news.scea.sony.com> <7lkucs$ko217@chuka.playstation.co.uk> <7lnteg$i9a1@chuka.playstation.co.uk> <37823145.682258342@news.scea.sony.com> <7mah71$t1v6@chuka.playstation.co.uk>
NNTP-Posting-Host: vmlabs66.vmlabs.com
X-Newsreader: Forte Free Agent 1.11/32.235

>p.s. Scott - did you find out which method (sqrt or inv_tan) was faster?

Nope, I haven't tried that experiment.


>Oh and why can't you just use a simple look-up table for inv_tan?

The problem is that the result of (dy / dx) can be any real number.

SIN and COS are nice because you know the angle will be from 0-359
degrees and can make a look-up table of the appropriate size.  Have an
angle of 248 degrees?  Simply look up array element 248 to get the SIN
& COS.

INV_TAN is harder because the value you would use as the index is the
result of (dy / dx).  This can be any real number from negative
infinity to positive infinity.  How do you make a finite-sized look-up
table for such a thing?  Compounding the problem is that the result
isn't linearly proportional to the input.  Or, inv_tan(1) is not
merely 2 times inv_tan(0.5).

That said, I'm sure there's a way (there always is), but I leave that
for the smarter folks.  :)

Scott