Subject: Re: Re[2]: [linux-audio-dev] Plugin GUIs
From: Paul Sladen (paul_AT_sladen.org)
Date: Sun Mar 25 2001 - 23:57:33 EEST
On Mon, 26 Mar 2001, oRick Burnett wrote:
> This may just be me but I cannot believe that doing real-time
> rendering would be faster than a bitmap lookup? Am I just missing
> something? The other problem is that the rendered images that you are
> doing on the fly will not be able to have the same 'quality' in
> certain instances. I am just imagining that here we have some midi
> control changing parameters while we are playing a stream or
> something. Lets say maybe three of four different knobs and sliders
> are changing. To me that seems like a TON of extra overhead for
> rendering each of those frames.
Indeed; (Sorry I should have emphasied the `=-)' smilies more ;-), my
point is that is not really pratical; in the same way, that storing such
high-resolution images of widget frame, I believe to be neither nessesary,
nor partical.
Never minding the alising effects that you are going to get when using a
source image outide of the range 0.5-1.5x of the size of the destination
size. The other extreme is to use vectors.
I'm switching a piece of (er oh, -- "commerical") DTP from seperate GDI
and Xlib drivers to a cross-platform OpenGL one.
The problem is fonts though; and as far as I can tell, at the higher
resolutions (think of a headline zoomed to full screen, with only three
letters visible) that it is going to be more efficent to draw the
characters out of primitives, rather than rendering to a pixmap, and
bliting to the screen. At the lower resolutions I do just that instead.
So, how does this relate; well what about avoiding such questions, and
allowing the (a?) library to do the drawing. It has three choices (just
like how my font sub-system does:
a) Scale an existing pixmap and blit
(scaling artifacts)
b) Build a pixmap [from vector data] at the apporiate size
then blit to screen
(first time is slower, since `rendering' into cache)
c) Render directly to the screen using vector data
(inefficent in large quanities at small sizes)
Anyway, I'll let you know over the next couple of weeks as I try some
things out. and finish it.
Paul
-- e: paul_AT_sladen.org t: 0115 922 7162
This archive was generated by hypermail 2b28 : Sat Apr 07 2001 - 15:39:36 EEST