Intensity of the throttle affects sounds

[HornetMaX]

>HornetMax

Okay, i'll try to find out all of this. Btw, i saw a "s2b_backfire" in a mod, is this working, or just a test ?

Here's how I think sounds are played, with a little example. Let's suppose we have:

• Layer0 (ON layer):
  
  
  • Sample_ON_0: with volume 1 between 3000 and 5000RPM
  • Sample_ON_1: with volume 1 between 5000 and 10000RPM
• Layer1 (OFF layer):
  
  
  • Sample_OFF_0: with volume 1 between 3000 and 5000RPM
  • Sample_OFF_1: with volume 1 between 5000 and 10000RPM

I've simplified the things just for clarity: only 2 samples per layer, samples envelopes are constant (=1) instead of the usual fade-out - fade-in, the samples do not have overlapping RPM ranges. Samples names are for clarity too.

Now let's say you're at 4000RPM with the throttle at X (between 0 and 1). What I do (and what I think GPB does, more or less):

• For each layer, find the volumes of each sample, given the current RPM: in this case we have:
  
  
  • ON layer: Sample_ON_0 with volume 1, Sample_ON_1 with volume 0
  • OFF layer: Sample_OFF_0 with volume 1, Sample_OF_1 with volume 0
• As the throttle is at X, weight each layer as: ON layer with weight (X), OFF layer with weight (1-X)

So in this situation we end up playing Sample_ON_0 at volume (X * 1) and Sample_OFF_0 at volume ((1-X) * 1)
The throttle level is, roughly, the engine load.

In reality: in each layer, you can have more than one sample with non-zero volume,just play all samples with volume > 0 (actually, I just play all of them all the time, if volume is 0, then the sample is muted).

Also, Samples typically have a fade-in / fade-out rpm range over which two consecutive samples overlap: this allows for smooth trnasition between one sampe and the next (and vice versa).  But that's not a problem at all: just determine for each sample its correct volume given it's volume envelope (described in the scl) and the current rpm (given by the physics).

In any case, as soon as my tool is ready you'll see all that visually: it will be much clearer

For the backfires: yeah it works, it's just not a smart way to do that. Instead of having some sort of logic that from time to time pops a backfire (under certain conditions), you have a sample that is always played when you are in the right RPM range. Making the sample long you give the impression the backfire doesn't happen always (as when you shut the throttle, the RPMs will go down and sooner or later the sample volume will go to zero).

MaX.

[Eagle]

I know that already. :\

The problem with simulators, is that they can't simulate perfectly the real time conditions of the engine irl. Hear an onboard video and ingame, even if it is similar, you hear the difference. And i think these kind of details makes the difference.

Examples:

Irl, driving in 6th gear at 2000rpm (the bike have the neutral limit at 1000) is making it look like it is drowning, you can feel vibrations on the bike. In gpbikes, you just drive at 2000rpm, normally..

The example of the "vrrrrrrooo" sound can be heard when you are stabilizing revs in a corner. In gpbikes, it only plays the usual sample.

These kind of things.. Of course i take examples from I4 4S engines, but it can be applied to other ones.

[h106frp]

I think your describing the different sort of sounds you get when either engine braking hard down through the gears with throttle shut off or labouring the engine with a large throttle opening and low rpm and too tall a gear.

[BOBR6 84]

Ok, if somebody can send me a few links via pm of some equipment il need (cheap as possible but good enough) il take a serious look. I have no idea what to look for..
I have free access to a rolling road.. For the laptop is that a case of having the right program etc?

Also.. If this is more than likely a waste of time and money. Please tell me lol

[h106frp]

Something like this will interface pretty much any mic via USB and give high quality recording (£37)
http://www.gak.co.uk/en/alesis-core-1-inline-usb-audio-interface/93391?gclid=CL_ty5T9tMMCFQbkwgodGy4A1Q
Comes with a free basic multitrack recording and editing package  Cubase® 7 LE which is the lite version of an industry standard

And for a simple mic cheap as possible something like (£25)
http://www.gak.co.uk/en/stagg-dm-5020/20869

And probably about £8 for a XLR mic cable

So you could get a start with a limited budget for <£70

This is a unidirectional (cardioid) pattern mic so will tend to pick up mainly the sound its pointed at and is probably best if pointing toward an exhaust or inlet to reduce contributions from other sources. This type of dynamic mic are easy to use in areas with high sound pressure levels. Condenser mics are nice and tend to have a wider/flatter frequency response but are a bit more finnicky to use due to the sensitivity but will be recommenced by a lot of people as being more 'professional'. If you stretch your budget a bit Behringer are accepted as producing good budget microphones.

If you really want to use condenser mics then you are best with an interface that has 'phantom power' this just feeds a voltage down the mic cable to power the microphone element and its associated electronics (or build your own battery powered in-line phantom supply if you are into that sort of thing and use the simpler interface above ).
http://www.gak.co.uk/en/focusrite-itrack-solo/74147

Ignoring mechanical noise the area over the tank for the intake and behind the exhaust would probably characterize the sound signature best. If you went to 2 channels you could mix this down to mono to taste, a stereo system is of course double the cost, or you could repeat the run with the single mic in different places and use the software to edit the samples.

Plenty of other software choices, the simplest to use and free is audacity.

If you really just wanted to just 'give it a go' and see if the results were worthwhile pursuing you could just get a simple battery powered condenser mic from somewhere like maplins, this will plug straight into the laptop and use its sound card and audacity. Mate of mine swears by this with his old minidisc recorder, currently on stock clear for £5


If you are going to try and engineer sounds for the sim the most important item is good source material, so a full set of samples as described by MaX would be a very worthwhile effort in my opinion and almost impossible to obtain any other way. You could spends hours extracting sounds from old video clips but because of the way the sim needs to blend them to cover the rev range they will never be 'right' unless they are captured in the correct rev range blocks.

If you could get a set from both the typical 'observer' position recording the exhaust and another from the 'rider tucked in' on bike position(head over tank) it would be really interesting to hear the difference in the replayed sets. Typically the sounds in game seem to be very exhaust biased which are behind the rider, but riding i personally always relate more to the bellowing sounds from the intakes which are just below your head.

[BOBR6 84]

Thank you very much h106frp!!

I can give this a go no problem! Two ''issues'' though.. If I record the sounds can all the data be sent easily enough to somebody who knows what to do with it? Maybe on this forum?
Also.. Using a rolling road, it would just be my bike used! Unless I can pursuade some friends to do it as well.. (not everybody is as keen as me for arm chair racing lol)

But hey.. It would be a start and a good test to see how it turns out, right?

Thanks again!

Ps. (SAS) feel free to have my posts moved somewhere else dude..

[h106frp]

We should be able to move them via dropbox or similar. I think you will certainly prove whether a decent sounding mod is possible if the source materials are available. With some processing you could probable produce several unique sounding sets from these recordings.

@MaX with your knowledge of the sound file can you suggest the best rev bands and sample lengths? (will need to be a bit longer time in the original recording to allow for loop trimming)

If no-one else wants to try i will have a go at making the loops although i'm a wrinkly analogue recording sort, the youngsters make this sort of thing look easy as its the way a lot of modern 'music'  is produced. Might need some help or pass it on to assemble the sim engine file.

Apparently audacity will do it as demonstrated here;
https://www.youtube.com/watch?v=ryLpfVecUDs

[HornetMaX]

Ignoring mechanical noise the area over the tank for the intake and behind the exhaust would probably characterize the sound signature best. If you went to 2 channels you could mix this down to mono to taste, a stereo system is of course double the cost, or you could repeat the run with the single mic in different places and use the software to edit the samples.

GPB uses mono samples, so anything would do (stereo samples will have to be mixed down to mono, easy).

If you could get a set from both the typical 'observer' position recording the exhaust and another from the 'rider tucked in' on bike position(head over tank) it would be really interesting to hear the difference in the replayed sets. Typically the sounds in game seem to be very exhaust biased which are behind the rider, but riding i personally always relate more to the bellowing sounds from the intakes which are just below your head.

I totally agree, would be nice to test that.

@MaX with your knowledge of the sound file can you suggest the best rev bands and sample lengths? (will need to be a bit longer time in the original recording to allow for loop trimming)

Warlock is probably better placed than me to answer that.
I'd say that we need one sample at idle, and then one each at low / mid / high revs, maybe something like 5,000, 10,000 and 15,000 RPM, each at thrrottle ON and then the same at throttle off.
Ideally the samples should be taken at constant throttle and you should mark down the RPMs of each sample.

Sample length: we probably need no more than 5 to 10 seconds, but capture a bit more to allow some room for tricky stuff.

But again, Warlock probably knows better.

Apparently audacity will do it as demonstrated here;
https://www.youtube.com/watch?v=ryLpfVecUDs

On an engine sample it may not be that easy.

MaX.

[Warlock]

I havn't time right now, ill try to answer this tonight guys 

[Warlock]

Ok, so for making loops this is basically what you have to do in audacity:

https://www.youtube.com/v/ryLpfVecUDs

I use Reaper or Adobe Audition, but is the same process so any software will work.
(Reaper is a great piece of audio software, and its free. At startup you will be prompted to buy it, but after 5 secs you can continue to use it FOREVER. Plenty of tutorials in youtube)

About the rpm range for recording, as Max said, Idle + 3 or 4 steps until max. rpms On and OFF throttle  (important take notes of the rpms for each step to name the wav files.)
I would say at least 4 steps , the more the better, because you have more options to play with. Some samples will fit better than others, and you can remove one of the steps  if you dont like it and play with the pitch in scl file to reach the next one.

Also the longer the better .wav. Any little variation in the sound if you loop a short sample with it , will be very noticeable after. So you probably need to cut that variation at some point and loop the rest of it. So record them until you think your bike is about to explode LOL 

You guys think is important what gear is selected when recording?    2nd-3rd gear 4th?

[Warlock]

By the way, in my opinion granular synthesis is the way to go, but we need a very good software that will cost a bunch 

Check this video and imagine a full rpm ramp  instead of star wars music.

https://www.youtube.com/v/ZEJiH3rVFB8


This is what i did for the moto2 sound but was a pita because bad quality of grains with free granular plugins for reaper. I tried everyone i could find..... , but no good enough.
What you hear in this video is amazingly good.
And this method is the way to get PERFECT loops, based on a single full rpm ramp recording.

[HornetMaX]

And this method is the way to get PERFECT loops, based on a single full rpm ramp recording.

Just to be sure I understand you correctly: if we had such a tool with a good sample (dyno sweep full ON, full OFF) you think we could create the samples to be fed into the .scl ?

MaX.

[Warlock]

Exactly

You can make it sound indefinitely at the point you want (rpms) like the video does with star wars song.
You can also select the width of the grain (grain size knob) lets say you select a full cycle of a 4t engine, and make it sound indefinitely 

I cannot undoubtedly ensure it will be perfect , because i didn't tried this specific tool, but i did the moto2 sound this way with other granular VST plugins i found like Kontakt 5, Granite or Density.
The difference is i didn't used 1 full cycle of a 4t engine, i had to select a wider range, and make it play slowly, as the guy did in the video plying the song in reverse.

I wish they had a demo to try it , but doesn't seems to
http://www.motu.com/products/software/machfive

[HornetMaX]

I wish they had a demo to try it , but doesn't seems to
http://www.motu.com/products/software/machfive

Well, it costs "only" 320-350 Euros no ? It's not really out of reach.

But the thing that buggers me is that it doesn't seem to do anything magical. I mean, what we need seems relatively easy: take a grain, vary its size, loop it, alter its pitch.
Or maybe I'm missing something ... I need to look a bit more into that.

MaX.

[Warlock]

Nothing magical, really. But other thing is quality.

Try to download Density demo from here and play a bit with it:

http://www.densitygs.com/

Download/Density Standalone (Mac/Win)