Re: Sources and Buffers

DAve Shillito <[email protected]> Wed, 6 Feb 2013 13:00:12 +0000
Newsgroups gmane.comp.lib.openal
Message-ID <CAFoMBTvHp+-wykURkKOWHKTjA3BgCbYN7+0ObzYeoDrAjf=1Ug@mail.gmail.com>
--===============1758331061==
Content-Type: multipart/alternative; boundary=20cf3074b45820fc5d04d50de89c

--20cf3074b45820fc5d04d50de89c
Content-Type: text/plain; charset=ISO-8859-1

Hi,

Sorry to come to this party late, I only check my mailing lists
occasionally.

We hit the problem of rapidly changing the notch setting on US Locos for
Microsoft Train Simulator.

What we eventually realised is that if you take a loop from one of the
middle notch settings and pitch it down, it actually approximated the low
notch loop pretty well, and likewise if you pitched it up it approximated
the high notch loop.

As such we just used one of the middle loops with the pitch and volume
driven by the notch setting (actually the engine rpm, which is driven from
the notch, and also took into account the load on the loco).
We had intended to use this as our bottom 'level of detail' for low end
PC's, but decided it gave better, and more importantly, instantaneous
results for the sound in general.

The fixed notches mean provide the 'stepped' nature to the sound.
The "surge" as you change notches can be accomplished by allowing a slight
ramping of the values, overshooting and returning to the desired rpm, or by
adding a separate oneshot "surge element" when the notch changes.

I would also think that adding a distortion effect to the buffer could
improve the overall effect (although we never tried that at the time - it
was about 10 years ago and we didn't have the capability then :) ).

We also used the same concept on Rail Simulator.

YMMV, but I thought I'd make the suggestion, I know it feels wrong to not
use original source material, but it's worth using the source as reference
and seeing what you can synthesize (even if it is only using a simple
pitch/volume change).


DAve

--20cf3074b45820fc5d04d50de89c
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hi, <br><br>Sorry to come to this party late, I only check my mailing lists=
 occasionally.<br><br>We hit the problem of rapidly changing the notch sett=
ing on US Locos for Microsoft Train Simulator.<br><br>What
 we eventually realised is that if you take a loop from one of the=20
middle notch settings and pitch it down, it actually approximated the=20
low notch loop pretty well, and likewise if you pitched it up it=20
approximated the high notch loop.<br><br>As such we just used one of the
 middle loops with the pitch and volume driven by the notch setting=20
(actually the engine rpm, which is driven from the notch, and also took=20
into account the load on the loco).<br>We had intended to use this as our b=
ottom &#39;level of detail&#39; for low end=20
PC&#39;s, but decided it gave better, and more importantly, instantaneous r=
esults for the=20
sound in general.<br><br>The fixed notches mean provide the &#39;stepped&#3=
9; nature to the sound.<br>The
 &quot;surge&quot; as you change notches can be accomplished by allowing a =
slight=20
ramping of the values, overshooting and returning to the desired rpm, or
 by adding a separate oneshot &quot;surge element&quot; when the notch chan=
ges. <br><br>I
 would also think that adding a distortion effect to the buffer could=20
improve the overall effect (although we never tried that at the time -=20
it was about 10 years ago and we didn&#39;t have the capability then :) ).<=
br><br>We also used the same concept on Rail Simulator.<br><br>YMMV, but I =
thought I&#39;d make the suggestion, I know it feels wrong to not use origi=
nal source material, but it&#39;s worth using the source as reference and s=
eeing what you can synthesize (even if it is only using a simple pitch/volu=
me change).<br>
<br><br>DAve<br><br><br>

--20cf3074b45820fc5d04d50de89c--

--===============1758331061==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Openal mailing list
[email protected]
http://opensource.creative.com/mailman/listinfo/openal

--===============1758331061==--