[A83] Re: arctan
[Prev][Next][Index][Thread]
[A83] Re: arctan
use the series expansion for accuracy (maybe a lookup t?)!
God bless Mathematica 4!
regards
Philip
>From: Patai Gergely <patai.ti@freemail.hu>
>Reply-To: assembly-83@lists.ticalc.org
>To: assembly-83@lists.ticalc.org
>Subject: [A83] arctan
>Date: Fri, 2 Aug 2002 10:56:57 +0200 (CEST)
>
>
>For my turn, I have a really annoying question to all of
>you: fast arctan, or rather arctan2 for two 8-bit
>coordinates with a 8-bit output as well. The algorithm
>doesn't need to have a deadly precision, and error of
>around +/- 5 degrees is acceptable. A maximum running time
>of about 1500 cycles should be possible.
>
>I was thinking about creating a 16x16 table of angles and
>divide the coordinates into two nibbles each (so I get the
>final vector as the sum of two vectors: one made up of the
>higher nibbles and the other from the lower nibbles). The
>higher nibbles could be used to point directly into the
>table, and the lower nibbles could provide some additional
>precision, but I don't know how.
>
>If anyone can think of something useful, I'd be glad to
>hear it.
>
>PG
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx
-- Binary/unsupported file stripped by Listar --
-- Type: application/mathematica
-- File: arctan.nb
Follow-Ups: