rec.autos.simulators

Found and fixed a goof up I made that caused the exhaust valve to act as though
it was sort of closing and not closing at the same time.  This seemed to be the
reason for all the static.  Second, the amount of cylinder filling is now a
function of rpm (volumetric efficiency changes as the engine revs through its
range).

Sorry, there's no slider controls or anything yet :-)

I'd like to get your input on this one!

http://www.racesimcentral.net/

By the way, I goofed up when compiling a couple of these so the names are
switched around on maybe two or three of these.  EngineTame.exe is a high
revving racing engine (all 5.7 liter V-8's again), for instance.  The original
EngineTame.exe vanished..

Todd Wasson
Racing Software
http://www.racesimcentral.net/
My car sim
http://www.racesimcentral.net/

Hi Todd,

Cant believe no one has commented on this.  Great piece of work!  Its
definitely developing into something very nice.  There is still something
slightly synthetic sounding about it, not sure what it is, but its
progressing nicely in the right direction.  Keep up the good work!

All the best,
Ash
http://www.siroccoracing.com

Rich

Thanks Ash :-)  There is a lot to do on it yet that should really transform the
sound.  Whether it will transform it into something better or worse remains to
be seen of course :-)

Todd Wasson
Racing Software
http://PerformanceSimulations.com
My car sim
http://performancesimulations.com/scnshot4.htm

Thats great news, cant wait to hear part 3!!! :)))

All the best,
Ash
http://www.siroccoracing.com

Todd,

Excellent work you are doing, any more insight of how you are doing it?  I
think you said you based it on some papers, any public references to them?
The sound is convincing although still a bit synthetic.

Cheers and keep up the good work!
Tom

The only thing that gets me, though, is there just seems to be a general
lack of bass.  Having grown up driving 350 Chevy's (5.7L goodness), I can
tell you that not a one has that little bass to the sound...

I oughta figure this out for myself one of these days, though...so I can
figure out what exactly makes my brother's '94 Z28 sound the way it does, so
I can make my '70 GMC sound similar, hehe. :D

Milhouse

Just ran across this - very nicely done, with some interesting effects, but
through my SB Audigy it's still super-static-y.  Love to hear what it must
sound like clear...

Thanks for sharing,

SB

Thanks, Tom :-)

The system works basically by modelling all the cylinders individually as the
pistons move up and down (connecting rod length effects this too, but I can't
hear a difference).  The exhaust valve opens and the gases are forced out.
What you're actually hearing is the cylinder pressure, not the exhaust tip
pressure as you should.  I think this is probably why it sounds a little
synthetic.  For most of the loop the curve is much like a sine wave, but you
get a wobble in the waveform when the valve opens.  

That part wasn't done with any papers, but the vibration in the formula 1
engine sample was inspired by an SAE paper Stefan (last name escapes me at the
moment) referred me to after Haqsau asked about it.  I don't think that would
have dawned on me any time soon, so thanks guys :-)

Basically it's letting the engine block shift back and forth randomly a couple
degrees as it's spinning (only in the F-1 sample).  This causes the sound to
change by altering the piston position in an irregular fashion, so I think it
would be very cool to tie this into the car's orientation so when you hit a
curb or something, you get that little "jumpy" sound as you accelerate away.
Of course that would be most noticeable with stiff springs :-)

Todd Wasson
Racing Software
http://PerformanceSimulations.com
My car sim
http://performancesimulations.com/scnshot4.htm

Interesting.  I get a good amount of bass on the speakers at work.  Somebody
there has a big subwoofer hooked up to his system, so I'll have to try it on
that and see what happens.

Yes, that would be cool :-)

Todd Wasson
Racing Software
http://PerformanceSimulations.com
My car sim
http://performancesimulations.com/scnshot4.htm

I've tried a few different types of motors and they sound ok, I guess, but the
V8's are my favorite so you'll hear lots of them :-)  I set it up like a Harley
Davidson early on and it sounded pretty convincing at low speed.  (All the
pistons fire at the same time, brap brap brap).

When I get time (I'm leaving for the Indy 500 in a few days, woohoo!), I'll see
if I can post some other engine types.  I'm more anxious to get an exhaust
system in and see how it sounds.  It should really change it a lot, but it
might wind up being pretty lousy :-)

Todd Wasson
Racing Software
http://PerformanceSimulations.com
My car sim
http://performancesimulations.com/scnshot4.htm

Thanks, Steve.  I'll make a point of putting up some that don't have as much
static.  To me it starts sounding a bit more synthetic without the cycle by
cycle variation in cylinder filling (which causes some of the static).  

Todd Wasson
Racing Software
http://PerformanceSimulations.com
My car sim
http://performancesimulations.com/scnshot4.htm

p_c(CAD) = p_ivc (V_ivc / V(CAD))^k_c, where k_c is the polytropic

                                       coefficient for the compression.
p_e(CAD) = p_3 (V_3 / V(CAD))^k_e, where k_e is the polytropic
                                    coefficient for the expansion.
"3" is the point of highest pressure and lowest volume in the Otto cycle.
          ^  3
          |  **
          |  * **
log10(p) |  *   **  4
          |  2**   **
          |     **  *
          |       ***
          |         1
          +------------------>
               log10(V)

The total (approximate) pressure curve is then generated using a
"pressure ratio function" of your own choice. It is common to use
the Wiebe function, but I have been using an ordinary sigmoid:

f(x) = 1/(1+exp(k1*(CAD-k2)),

(for simplicity).

The design variables are p_ivc, p_3, k1 and k2.  By adding noise to
k1 and k2 it is possible to simulate cycle to cycle variations.
k1 will simulate influence on the combustion rate and k2 variations
in ignition delay.

The total pressure becomes (an interpolation between p_c and p_e)):

p(CAD) = f(CAD) * p_e(CAD) + (1 - f(CAD)) * p_c(CAD).

:)

How about filtering the pressure?  Why not apply a digital filter to the
pressure and try to play around with some filter parameters to obtain
nice sounds?  I don't know, but phasing effects may give a more
convincing sound maybe?  Try playing around with some discretised
analog filters, like Butterworth and such and see what you might come
up with. I would recommend you to make prototypes in e.g. Scilab or
Octave, which both are free high-level languages for scientific
computing. It makes testing much easier!

--
/S

/* http://www.s2.chalmers.se/~last */
(remove "nospam" from email to reply)

Steve

--
/S

/* http://www.s2.chalmers.se/~last */
(remove "nospam" from email to reply)