[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Pierre Phaneuf wrote:
> Steve Baker wrote:
> > > ...Voodoo3 hardware seems to be the biggest bang for the buck, mostly
> > > lacking in the visual departments (only 16 bit and picture always
> > > look too dark with the 3Dfx chipset, dunno why).
> > It looks dark because you don't have the gamma set up right.
> Yes, but I didn't properly word that... The colors seems "dull". It's
> cool when you're just coming from software rendering, but if you compare
> nVidia output and 3Dfx output, you'll see the difference immediately.
Dull? Hmmm - that could still be incorrect gamma correction. Have you
actually tweaked the Voodoo's gamma settings?
> > > Yes, knowing assembler is important, and even with todays compilers,
> > > sometimes you have to use assembler. For example, we have a sound mixer
> > > for Quadra (no threads, no sound server process, doesn't skip or lag!)
> > > that adds the samples together and then uses if's and whatever to
> > > properly clip it.
> > That's exactly what I have in PLIB - and use on my Tux game.
> You mean synchronous sound?
I'm not quite sure what you mean by that. I mean that I don't have
a separate thread and that there is an absolute minimum of delay
between the game stimulus and the commencement of the audio...and it
doesn't skip or lag unless the graphic frame rate gets suddenly
> > > This is a good case that could be replaced with
> > > assembler.
> > But is it? My code is in C++ and consumes less than 1% or the CPU on my
> > 266MHz PC - and only 7% on my ancient 66MHz 486. (That's playing three
> > 22KHz samples - each with pitch and volume modified on-the-fly and then
> > mixed together)
> > There is no way I'd rewrite that in machine code to save just a half
> > percent of the CPU...especially since I'd then have to make it work
> > for Alphas, Mac's, MIPS, PPC's, etc.
> No, at the quality you're mixing, no way that would be useful! But
> Quadra mixes up to 8 samples at an output rate of 44.1 KHz, at 16 bit,
> in stereo. The input samples are of various rate and it also supports
> pitch, volume and panning effects.
I can do that too - although it seems somewhat overkill.
I havn't tried winding the numbers up that high - but there are only
a couple of factors of two there - it's still not going to be over
one or two percent.
Steve Baker http://web2.airmail.net/sjbaker1
firstname.lastname@example.org (home) http://www.woodsoup.org/~sbaker