I actually agree 100% with everything that Jamexing has said.

Quote from AndRand :

as I understand sample is very short single pulse tone - i havent played with sound engine but - the pulse is singular blast so... have you tried clipped samples (in some frequencies)? just an idea

I have tried hundreds of methods of getting a sample, and taking one from white noise and running through different filters etc gives the best sounds in LFS. If by clipped you mean distortion, I've also tried that, and LFS cannot cope without rediculous in game clipping.

P.S. Any idea is always a good idea though.

The driver took the instruction "get a good launch off the line", slightly more seriously than intended.

New FZ sounds.

If these aren't an improvement I'll be suicidal.

"I couldn't face the long walk to the paddock... I wonder if he will notice.."

Quote from Lotesdelere :

Some tools you may find usefull:
http://socrates.berkeley.edu/~ ... ntary/NoiseGenerator.html
http://www.arteson.com/audionoise/
http://www.nch.com.au/tonegen/index.html
http://www.fleximusic.com/generator/overview.htm

Great finds mate! Thanks i'll have a look.

Edit: Can a mod please change "Thread" to "Log" please.

Last edited by DaveWS, Sun, 29 Apr 2007 13:14.

I'm not entirely sure but 0.1 seconds is too long and 0.08 is ok. Just edited the post where I said 0.8 seconds

Quote from XsX! :

hi dave and others, after readign this post i've tried to modify the files, and i'm came up with this after messing all the day.
they sound much similiar to the original ones, but are more lively and growly without being harsh or stressing. Another difference is that the inside and outside sound of the car now is more different and loud when in external view.
try the lx6 and 4 too, with closed roof too
these are not all cars, tried some of the ones i use most.

Sound nice and rough, but too much clipping, and imo the power off volume is too low, try adjusting the volumes a little.

Quote from Shotglass :

im not exactly sure if i get your point there

btw concerning white noise
keep in mind that the whiteness of the nosie doesnt determine the volume density so theres one more degree of freedom to play around with
in my tests with pure white noise i only got serious clipping issues with uniform white noise ... gaussion worked rather well

Look at the attachments. The first file is pure white noise cut to a pulse length. The 2nd one, is that white noise ran through a bandpass filter. The 3rd one is the same as the 2nd one except for the volume of the pulse goes from max at 0 seconds to 0 at the end. The last one has been through the same procedures as the 3rd one, but cut from a different place in the white noise I have. In theory, the 2nd and 3rd ones should sound different, and they sound the same. The 4th one should sound like the 3rd, and its completely different.

Quote from Shotglass :

i know i eventually gave up on trying to toy around with the sound engine after i figured that it more or less sounds the same regardless of the pulse sample
no idea what causes it but its definately a "problem"

If you read this:

Quote from DaveWS :

Now the interesting bit is the although the clip of (now filtered) white noise I took samples from was basically the same noise throughout, each sample I took from it sounded different each time in LFS.

The shape of the first few "waves" in the sample really determine how the engine will sound in LFS. So if you have a Edit: 0.08 seconds long sample, and you delete the first 0.005 seconds of it, the engine tone in LFS changes noticeably.

Another thing worth noting, is if you have a sample with a constant volume (so when you loop it you just get noise), it sounds the same as if you slope the volume to 0 in the sample (giving a popping noise when looped) in LFS.

@Lotesdelere: By using white noise, I can essentially change it to whatever "flavour" / type of noise I want. Strangely though, if you use a sample with a very broad frequency range, LFS doesn't like it at all, and you get constant clipping. This is why I played around with different bandpass filters to get the best range.

Last edited by DaveWS, Sat, 28 Apr 2007 22:00.

"Mark, this really is *groans with pleasure* nice you know, but there is a time and a place, and this is not the time nor the place..."

Edit: LOL good one tristan

Last edited by DaveWS, Sat, 28 Apr 2007 17:02.

Innovative SEAT driver finds new ways of putting the power down in the wet...

Or

mmmuuuuummmyyy!!!!

Quote from lalathegreat :

Too loud at low RPMS
then i noticed the car is actually too loud for the rpms its actually at in the lower range. if u listen to a real car drive by the engine is hardly audible(in the cruising RPM range) so is the exhaust(this is from the external view) the tires actually make more noise.

Loud IDLE
i noticed the idle is ridiculously loud i parked the XRT at the 50M marker at blackwood straight and i could still hear the car from 250M marker. also did the test at the start line on blackwood and i could hear the car at the pit entrance lane cones.

Power vs loudness
was curious to see if power played a role in how much noise the car produced so i lined a XRt up on the BLack wood straight took off did a replay and listened from the same spot the car left and noted when i could not hear the car anymore. i redid the test with the hatch and in terms of loudness they were pretty much the same. sound cut out at the same points. went back to an old lfs ver and tried the same test with a 7.5liter V8 (500Hp) same thing.

Attenuation
LFS dosent do this. basically as a sound source gets further away the higher frequencies of that sound are absorbed by air and don't travel as far as the lower frequencies. so stuff like gear wine should not be audible after a certain distance. another example of this is a helicopter u can hear the sound the blades make from further distances than you can hear the high pitched noised from the jet engine.

3D Sound would be nice
back Fires
would rather see this run off of some simulation rather than down shift and hearing a pop. rather have the sound and the size of the backfire dynamic. (love the sound of 4 or 5 cars going into a corner and all that random popping )
---
I don't think LFS sound is actually going to get better without hardware acceleration. i would suggest scawen look at OpenAL and EAX5 as 3d sound and attenuation are easier to implement using them. i can't really comment of all the benefits in terms of sound quality in using those because the documention for EAX5 is under NDA. You have to register with creative as a developer.

For the first two tests, this is already simulated to a certain extent in LFS. The cars tend to have a quiet idle volume unless you boost the "idle volume boost" tab.

In your 3rd test, what Bob said was correct, it has nothing to do with the volume of the car, LFS only plays the sound up to a certain distance away (which is not really a good thing imo).

For the rest of your points, yes, good ideas. I think backfiring wouldn't be too much trouble, it would be similar to the popping sound you get with engine damage / TC.

Quote from anttt69 :

Patch improvements were a huge step in the right direction thanks Davews
my suggestions:

When you crash into something it should sound like a skid, crunch & bang. Like a real accident. Not like a biscuit tin sliding across a car park.

Those horrible tin ey tapping nioses when you rub bumpers or doorhandles have got to go. Totally unrealistic imo.

So big crunch, wallop, bang sounds, curb rumbles, & race track atmosphere sounds (cars blasting down a straight in distance, cars roaring past the pits & gravel sound when you slide off into the kitty litter) even a pnumatic air gun (when pitting) sound might be an idea.

example of crash sounds: http://www.youtube.com/watch?v=G_6AyjDIqZA
just try to ignore limpbizkit & listen to all those 5's skiding & crashing
the cheers from the spectators says it all.

Yep, these too need work. But we need these re-recorded, and not just taken from videos, in order for better results. Skid sounds too.

Quote from Shotglass :

what i mean is not using one single sample for the entire time lfs is running but using a noise generator the output of which gets filtered to create the pulse "sample"
with the engine pulse sample essentially being noisy this might or might not add some depth to the engine sounds (either way it should be more realistic as no firing of the engine sounds exactly the same irl like it does in lfs)

I don't mind you chatting about other stuff if it's still basically on topic.

I now understand what you're saying. IMO, it's a great idea. Could be very interesting. Unfortunately I can't "preview" how it would sound though. One interesting thing you would probably like to hear, when I made the sample for the RAC from white noise, first I put it through a specifically set band pass filter to get a reasonable tone in LFS, then I tried cutting an adequate length piece of sound to use as a sample. Now the interesting bit is the although the clip of (now filtered) white noise I took samples from was basically the same noise throughout, each sample I took from it sounded different each time in LFS. So your idea could give some interesting results.

Last edited by DaveWS, Sat, 28 Apr 2007 11:31.

Quote from Ball Bearing Turbo :

Elaborate? Hm... There is a sound that goes tickety tack.

It's much more prominent off throttle, than on throttle. It's really audible in the FZ50 sound. All of the sounds sound good ON the throttle at very high revs. So, I suspect your are right, and that with hours of fiddling you could probably work some magic. Look what you did with the regular sounds!

If what you did somehow faded in based on torque/power output, and went away off the throttle, that might help. I suspect you already know what to do though.

edit: Why do you say it's going nowhere?

I get what you are describing now, that's the high frequency distortion. With more fiddling I could have the effect applied to certain volume levels and above (which would hopefully reduce the off throttle noise). The problem is most of what I'm doing is trial and error (and I really don't have enough time at the moment). I'm not as much as an expert as some of you think I am, I've learnt / am learning what I can though. But really, I need more of an understanding of how the sound system actually works. Anyway enough rambling, I have a huge amount of coursework due for Monday.

Quote from Ball Bearing Turbo :

What about that tickety tacking sound?

Erm, care to elaborate? I only posted these to see if there was potential, doing it this way. IMO it adds much more character and liveliness to the sound. With hours of fiddling I would think it could sound good. Its kind of a controlled distortion at specific volumes (making it different to clipping), but like I said could sound much more convincing..

Edit: My mistake was allowing too much high frequency distortion

Edit 2: This is going nowhere. I should let it rest until we have some more patches out of the way.

Last edited by DaveWS, Fri, 27 Apr 2007 23:36.

Quote from Shotglass :

which to my knowledge is a very crude system that stacks a few low passes on top of each other and thats it ... at least thats what it sounds like

Could be, although I wouldn't think so, based on how the sample affects the resulting sound in peculiar ways...

Quote from Shotglass :

how exactly did you create these ?

he should be able to implement my suggestion pretty quicky ... at least if he can figure out how to get a working pulse sample from noise

Basically I just edited the sound dynamics in an audio editor, to see what would come of it, and it requires very little processing power.

Do you think you could elaborate a little? Because I thought you were after a similar effect to what I used to produce those clips. Some of the engine pulse samples (such as the RAC) I made from white noise (heavily edited of course) btw.

Quote from Ball Bearing Turbo :

I just had a listen...

I'm not generally very critical, but I have to say that whatever that is sounds like the current sounds with some strange, VERY VERY artificial noise of some sort over top of it. It's much worse than what we have IMO.

You know I love your work, but whatever that was - I hope I never hear it again

I still love you though

Edit: In the attempt to say something constructive, I would note that at VERY high revs, it doesn't sound TOO bad. But at mid/low revs it's quite cacauphonous indeed. Whatever it is might bring a hint of the rawness needed at very high revs, but it ruins the rest of the rev range big time.

Edit2: I don't think we are going to get anywhere without Scawen doing... the reverse of what he's doing now with the exhaust side: induction noise needs to be simulated, not just exhaust noise. Induction is what makes up most of the vehicles engine sound when you're at high revs.... The mechanical clutter starts to overpower everything at some point. Botching the exhaust noise to compensate probably isn't going to be successful IMO.

You've been doing a lot of "stock car" work lately....

As far as I know the V8 in LFS is only flat plane crank... there needs to be a cross-plane crank firing order V8 to get the right sound, although whatever you're using sounds very nice.

Well, it was simply a quick attempt at fiddling with some sound dynamics editing (which can change the whole character of the sound, but its normally some form of distortion). I wouldn't expect it to sound good straight away, but I think there could be potential. I know there is a hissing effect, ignore that , it can be fixed.

I agree with you that this is a "fake" way of producing sound with more character, however I would think for Scawen to come up with a model of something else is a) going to take a while, and b) is going to require more CPU. However I agree it is the way to go eventually.

The stock car, yes... well I just think it sounds great that's all. As far as I know we WON'T be having one in LFS any time soon...

Edit: I don't mind criticism. If no one criticised we wouldn't get anywhere would we.

Last edited by DaveWS, Fri, 27 Apr 2007 23:22.

Some more.

Quote from Shotglass :

thats something a lot further down the line though whereas my suggestion can be done as a relatively quick hack job
going along the same trail of though you went what would help a lot is physical modeling of the pipes ... if im not mistaken currently the simulation is basically jsut a bunch of low passes stacked on top of each other ... if you were to add some delay between those with the possibility for waves to travel both ways things should change dramatically for the better

one thing which i think is way off is that the transmission whine doesnt change at all with the gear settings and final drive

That's not how the sound system works as far as I know. It puts the pulse sample through an exhaust simulation.

Quote from Shotglass :

like i said ... ditching the samples and going for true noise as excitation should add a whole bunch of distortions which sound better and are relatively realistic

I guess you are thinking along the lines of these, which would not be too much work for Scawen to implement into the sound system (I would guess). I think there is potential here, they aren't perfect though, have a listen anyway.

Edit: Actually they are pretty crap, but I STILL think there is potential.

Last edited by DaveWS, Fri, 27 Apr 2007 23:52.

Victor really must be wetting himself by now..

Another mix for you guys to judge.

Ok guys, here is another dual recording of the BF1. In other words a MIX of two recordings of the BF1 with different settings. IMO this is really getting somewhere now. Edit: The more I listen to it the worse it gets.

P.S. Again ignore the reverb/echo like effect where the mix isn't perfectly matched.

Last edited by DaveWS, Wed, 25 Apr 2007 22:52.

There is no hope..

BF1 any better?