The situation with the Tomahawk X is best explained by this picture:
No real world V10 uses that configuration (for several good reasons).
The other two variants, whilst distinct from each other in terms of the exhaust tract, are actually very similar in terms of their base sound, whereas the X variant is unique in both the base sound and the colouration from the tract (and all that rasp). The 14 000 rpm playground has a lot to do with it as well.
The rasp, by the way, is so overblown because PD don't have intake sounds, yet.
...
@Griffith500 , wouldn't it be possible to simulate the sound of induction noise by using a sort of vacuum/air compressor through an actual induction pipe and filter so it eliminates the sound of engine noise? Possibly even for individual throttle bodies? I mentioned something like this before for Blow Off Valves through a Mighty Car Mods video, an air compressor, piping, BOV of choice and boom, blow off valve separate from the engine to record.
The sound of induction into an ICE is a pressure phenomenon resulting from the intake valves opening (IVO); when the valve opens, a "negative" pressure wave communicates through the ducting and air is pushed in to equalise the pressure in the cylinder. Same for the exhaust, but with a "positive" wave: positive pressure relative to atmospheric, gas flows out of the cylinder.
The best way to get the best picture of all the acoustic phenomenon that affect that "pressure equalisation" is to use a real engine cylinder, and the easiest way to get the dynamic situation of all the overlapping pressure waves in the ducting and the cylinder(s) is just to run it. Anything short of that dynamic and spatial accuracy is not going to give the full picture.
I think you could probably just motor the engine (e.g. with an electric motor) and it'll work with less noise, but the cylinder conditions at IVO are partially determined by the combustion process (and the exhaust tract), so it shouldn't be ignored. I.e. you need to record a real engine, running normally, for reference.
If you're using a recording for reference, and not for playback in-game, the absolute isolation of all sources is not strictly necessary (not that PD manage it with their recordings anyway), but it would be useful for development of the model itself. You could use a few examples whose intake sound is easily isolated, or use manufacturers' data.
Regarding the model itself, automotive engineers use simulation to quickly determine a starting point for all the ducting, cam timing etc. for the engine, including silencing requirements. That might be a good place to start looking. If you want to measure the acoustic response of ducts and their ancillaries, the work has already been done and there are tomes full of empirical data as well as empirical and physical models intended for design work in engineering, and which could be coerced into a model fit for a game.
However, for some reason I found it very hard to model the induction sound properly when I didn't much struggle with the exhaust situation. I'm still not sure exactly why that was - I applied the same process to both, but the results just came out better for exhausts. Maybe there is an inherent difficulty with induction that PD is also facing (likely related to non-linearity).
Still, the basic induction noise I ended up using made enough of a difference, and it should be possible to cheat with it if it won't model itself right...
Raceroom has set the bar where trans whine and trans modelling is concerned!
www.youtube.com/watch?v=Y5SMi6FfyyI
I think iRacing still has the crown:
It's not just an effect, it's part of the physics simulation:
buildnotes
Flexible Drivetrains
- All cars now have flex modeled in their drivetrain math, instead of being perfectly rigid. This gives all kinds of subtle and interesting audible and physical engine rpm wobble effects.
Source.
Most other sims use scripted effects, or at most a primitive, single mass-spring approximation, rather than the (likely still simplified) full system representation iRacing uses.
