From owner-freebsd-multimedia  Mon Feb 26 07:29:18 1996
Return-Path: owner-multimedia
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id HAA24970
          for multimedia-outgoing; Mon, 26 Feb 1996 07:29:18 -0800 (PST)
Received: from plains.nodak.edu (tinguely@plains.NoDak.edu [134.129.111.64])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id HAA24963
          for <multimedia@freebsd.org>; Mon, 26 Feb 1996 07:29:14 -0800 (PST)
Received: (from tinguely@localhost) by plains.nodak.edu (8.7.1/8.7.1) id JAA13002; Mon, 26 Feb 1996 09:28:44 -0600 (CST)
Date: Mon, 26 Feb 1996 09:28:44 -0600 (CST)
From: Mark Tinguely <tinguely@plains.nodak.edu>
Message-Id: <199602261528.JAA13002@plains.nodak.edu>
To: leonard@dstc.edu.au, multimedia@freebsd.org
Subject: Re: ComputerEyes PCI/1024
Sender: owner-multimedia@freebsd.org
Precedence: bulk

Digitial Vision wanted me to write a driver for the Computer Eyes/1024
since I wrote one for their much older 8 bit ISA Computer Eyes/RT.
I decided to pass on the driver because the card is too expensive (a Meteor
clone is much cheaper) the CE/1024 uses a ISA bus rather than the cheaper
PCI bus meteor clone, but most convincing was the fact the Computer Eyes/1024
like the older RT provides only RGB 24 capture mode. this means you will have
to convert the RGB 24 to YUV in software. I found that the software
RGB 24 -> YUV significantly reduces the frame rate.

In the CE/RT driver, I made the big mistake of doing the RGB -> YUV
using floating point, a much better way would be converting RGB 24 to 
RGB 8 with shift and then doing a table lookup to convert RGB 8 to YUV
like nv does it for X capture mode. Even using this method the frame
rate will be lower than the cheaper metoer clone.

-mark.