Sound Engine infos

Discussion in 'General Discussion' started by Fonsecker, Oct 7, 2010.

  1. Max Angelo

    Max Angelo Registered

    Joined:
    Oct 5, 2010
    Messages:
    4,958
    Likes Received:
    10
    Wait a moment ... may be i am writing things you know very well :), but pasting samples is not easy.

    You have to check the wave shape and paste the two samples where the wave is "smooth", like it was one new smooth wave, without interrupt points.
    Two waves pasted without a continuity in the shape, generate a noise, using any software ... well, my Sound Forge version is very old, may be in the years the technology is improved and now loops and pasted samples are easier to do, and the tools automatically smoothing the wave shapes now work better :)
     
  2. mianiak

    mianiak Registered

    Joined:
    Oct 5, 2010
    Messages:
    691
    Likes Received:
    46
    If you use a spectrum analyser, you can aim to get every frequency equal, this will work well with almost every system. I did live sound for most of my life (till I went half deaf in one ear :( ). The whole Idea about producing a good sound is to get it 'flat', as in a flat frequency response. Then the enhancements are done by the end user. You get the urge to turn a certain frequency up, but you have to discipline yourself not to and never go above 0db on an EQ, always turn down what's bad, never turn up what's missing.

    In live sound, you use the 30 band EQ to remove the dominating frequencies in the room/venue you are in and make the PA sound flat, then once it is flat you can enhance individual instruments with the mixing desk.

    There is another thing I have learned and it is a fact, different races of people hear sound different, Asian hear midrange, African hear bass and European hear high's, you can see/hear for yourself by checking out different music from around the world (not commercially produced music, check out local unsigned artists).

    So getting the sound as flat as possible is your best bet, it might not sound good to you, but that's because of your system. Unless you have a couple of thousand dollars worth of mastering monitors in a purpose built mastering studio, you will never here it like it should be. That's why using a spectrum analyser is good for the home user.
     
  3. Carbonfibre

    Carbonfibre Registered

    Joined:
    Oct 5, 2010
    Messages:
    70
    Likes Received:
    0
    That's interesting, because I'm of European origin and I consider music with crystal clear highs to be of better quality. -also, using a EQ to boost highs help me increase perceived quality, and well as tightening up the bass frequency which results in a standard bell curve EQ setting that I nearly always apply by default where I can.

    (I'm no sound expert, but I can throw some words out to describe it.)

    rFactor audio (even on the best mods) always tends to be a bit lose on the frequency representation imo. Maybe that's why people keep incorrectly praising GT5 sound for the quality of it's synthesizers because of excellent frequency separation of the highs while maintaining a punchy bass, with any noise being in high frequency range. I know rFactor often has a crushed bass with noise that often spreads all over the mids because tight bass is loss when the samples pitch is raised.

    So if rF2 could have something like a dynamic range, EQ and lowpass filters implemented to make sure there's always a deep rumble from road-noise/growling engines below 100hz while having sharp detail at 15,000 Hz or higher at the same time... well that would be a great treat! :cool:
     
  4. MaXyM

    MaXyM Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,774
    Likes Received:
    29
    It's not so easy. Pasting two waves to get no sound crackles is not about joining to digitized waves (means, neighbour samples with most same value). It must be done in frequency domain. Mean, sample must be de-synthesized into frequency waves (ie FFT), then find closest points and after all synthesize sound again. Faster method is X-fade. Means recordings are mixed with folding volume.
    WIthout it you will get co crackles but you may hear periodical rumbling.

    Remember quality sound processing is not cheap (good algos + powerfull resources).
    Finally don't know what you mean 'flat'. Flat in frequency range (which sounds like nonsense) or beaing the same as original sound (without postprocessing).
    Finally to get the same result as in real, you should play recordings with the same dynamics and volume as real one, which is often unacceptable)

    If you want to have punchy bass you have to record sample with good quality or process it using some software. Of course this bass will be heard only for drivers equipped with required sound-rig (especially sub-woofer).

    Again: rf sound samples quality comes from 2 reasons:
    - hard to record such dynamic sound like car with amateur hardware
    - rf low quality samples processing algos *)

    *) last time I worked on samples for some not-sporting car. Using given recordings I wanted to get quite muted, specially on hi-freq sounds. After preparing sample into sound editor I put into rf and get some strange added noise. It wasn't aliasing noise. Some kind of sinus wave about 500Hz modulated by current playback frequency (engine rpm).
     
  5. mianiak

    mianiak Registered

    Joined:
    Oct 5, 2010
    Messages:
    691
    Likes Received:
    46
    Yes, flat as in flat frequency response. It's the first thing every sound engineer learns. Call it crawling before you walk, you cant mix good unless you have a flat sound to start with. What I was advising was for the home user who doesn't have the expensive equipment. We are not talking about Abbey Road studios, if we were I'd explain it a completely different way, but by using this technique, you will get it as close as possible to what it needs to be.

    I honestly don't know how to react to your reply. In a way I'm offended, but in another way I want to help you understand more about sound.
    You have a fair idea of what's going on and it seems as though you have a natural talent for sound editing. But accusing me of talking nonsense is not very nice, it'd be best if you listened to people who really know what they are talking about that take the time to explain things. You will only benefit from it in the end.

    I think your meaning harmonic distortion, it happens when 2 frequencies are not in the same 'scale'. You might have heard a guitarist tuning up, as he turns the tuning keys on his guitar you hear the sound quiver and when it comes smooth you know it's in tune. The quiver is harmonic distortion. Harmonic distortion can sometimes be magical, it can create certain overtones that give a certain sound a unique characteristic, like a car engine for example.

    Maybe it's a language barrier issue, but I think you mean sine wave, not sinus wave.

    One more thing, blending sounds together is not that hard once you know what your doing. I said this to a friend in another place, It's like making a seamless texture, you find parts that look similar to the parts in the join and you clone those bits onto the joining area. Sound is much the same, you just find similar bits. But I must admit, joining music together is a bit easier than joining an engine sound together. But still the same principle applies.
     
  6. MaXyM

    MaXyM Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,774
    Likes Received:
    29
    hmm. I cannot see I wrote nonsense accusing you.
    Your idea is great but imo impossible to realize in today's products on game market. There are other (faster, cheaper) solutions, call it 'shortcuts' to get the same result. Especially if it would require the knowledge about sound mastering (from common modders).

    Thanx for giving me an idea with harmonic distortion

    btw: I used to be music composer years ago. So I have some basis but never reached sound producers skill.
     
  7. mianiak

    mianiak Registered

    Joined:
    Oct 5, 2010
    Messages:
    691
    Likes Received:
    46
    Its very simple to do, you just open the spectrum analyser and a 30 band EQ, then look in the spectrum analyser for the peaks and turn down those frequencies in the EQ. Turning down 1 freq. doesnt always work, that's because of the overlap between frequencies in the EQ, so if you find your turning down 1 frequency completely, you will find that the frequencies next to it need turning down, that way you don't have to completely remove 1 freq.

    If your talking about using a hard limiter or compressing the sound, you don't want to do that, it will kill the dynamics.

    yw re the harmonic distortion.

    You used to be a music composer :D that explains it, sound guy v's musician is a never ending battle ;) We find it best to just let the musician think they are right :p
     
  8. Carbonfibre

    Carbonfibre Registered

    Joined:
    Oct 5, 2010
    Messages:
    70
    Likes Received:
    0
    Well I hope rF2's sound engineer(s) is a proper expert on the subject since they would be the people/person working on the problem.

    But I also hope there's a good comprise between calibration and creativity in the end, since in my opinion, audio should also be delivering the correct emotional response just as much as the correct frequency response. ;)
     
  9. Ronin Hammer

    Ronin Hammer Registered

    Joined:
    Oct 8, 2010
    Messages:
    28
    Likes Received:
    0
    i agree with what the lot of you say about the current rF sound-engine, it's not bad.. actually it's not bad at all, but the lack of really good material to work with
    and in many cases, modders who doesn't take the time to really refine the SFX file to make the most out of the samples they have, is what makes people say "the sound engine is bad" but just as we heard with the porsche before, you can make it sound really sweet.

    However, there are some things in the sound engine that are missing, for instance the trackside cameras does not reproduce the road-noise a car produces, if you've seen youtube videos of high-speed runs along some straight all you can hear is roadnoise untill the tailpipes are facing the camera.

    another thing that's missing is simple reverberations, in car and when seen from trackside cameras, it's not like it's a tricky thing to put in.. simple EAX effects
    (wich btw has been around for ages... nascar 2003 anyone?) isn't that hard to implement -WHICH is why i have good hopes for these kind of effects in rF2.

    The mixture of samples and generated sound is what would solve the dynamic frequency-respons problem, with parts of the enginesound being generated as in
    LFS, blended with carefully pitch-matched engine sounds (so no double-sounds are audible through out the rev-range) again, this is not all that difficult to recreate
    so lets hope for the best but mostly, lets hope they create a sound-engine that is MODABLE =)

    My 2 thoughts. Rock on! Btw, Good thread =)
     
  10. Pandamasque

    Pandamasque Registered

    Joined:
    Oct 5, 2010
    Messages:
    160
    Likes Received:
    2
    Here's another suggestion for a new but rather simple and absolutely necessary sound parameter.

    SFX should have a line like this:
    In rF1 you can hear opponents that are like 2 seconds behind (unless the external sounds are very quiet). In reality you can't hear the car behind even when it touches your rear bumper. When the opponent's exhaust pipes are directed at you and very close, only then you'll hear his engine noise over your own.
    Of course that effect may depend on the type of car. Hence the need for an adjustable parameter.
     
    Last edited by a moderator: Jan 2, 2011
  11. Cavallino

    Cavallino Registered

    Joined:
    Jan 6, 2011
    Messages:
    14
    Likes Received:
    0
    I like the sounds of the latest F1 1991 mod, but I am curious to know if it is possible to take/record more contemporary F1 car sounds ( i.e. onboard/ and onboard downshifting), from taped TV broadcasts or internet, to be recorded and used in Rf F1 games. It seems that the modern F1 mod sounds don't capture that wobblely, shaky engine/ transmission tone when F1 cars launch off lower gears and the down shift sounds are not authenic, as they would be say from TV broadcast onboard. I am intrested in this area. also, I am especially disappointed in CM F1 2010 sounds.
     
    Last edited by a moderator: Mar 9, 2011
  12. Carbonfibre

    Carbonfibre Registered

    Joined:
    Oct 5, 2010
    Messages:
    70
    Likes Received:
    0
    The answer to that is either we're going to have to continue putting the warbles into the samples like we do currently, which would be disappointing depending on whether rF2 improves nothing else. Or ISI is going to have to figure out how to make the warbles happen dynamically through the sound engine, as well as a few other things.

    The general consensus is; removing the amount of canned effects would be improving things. E.g. in other fields of expertise; micro-bumps are better than sine wave bumps because sine waves are fake bumps. Also, 3D kerbs beat having flat kerbs from a graphical point and a physics point.

    Warbles in the sample are to simulate the transmission effects which aren't happening in rFactor, therefore the sound is canned/fake. Transmission modelling in rF2 would have to be overhauled first I imagine, before the sound would start to respond properly.

    It's a bit of a double edged sword for modders though. On one hand, modders will no longer have to put the warble in the sample, however, if transmission is to be truly separate in rF2, having samples made with the engine noise, transmission, and turbo/super charger sound all embedded into one stereo recording loop would become problematic if they all warble together.

    rF2 would need to come packaged with about 5 or 6 new solo transmission samples to cover all the different types of vehicles, plus turbo and super charger samples. In rF1 we got just 2 "tranny" samples I think... Enduracers introduced a new sample with the warble embedded which changed everything for the better a few years ago. In rF2, modder's may possibly have to record new race car engines using recordings of their production model variants somehow (without the transmission sound), so that those things can be mixed in later.
     
    Last edited by a moderator: Jan 6, 2011
  13. Cavallino

    Cavallino Registered

    Joined:
    Jan 6, 2011
    Messages:
    14
    Likes Received:
    0
    Anyway, let's hope it will sound close .... Thanks.
     
    Last edited by a moderator: Jan 7, 2011
  14. Pandamasque

    Pandamasque Registered

    Joined:
    Oct 5, 2010
    Messages:
    160
    Likes Received:
    2
    With all the rF2 dedicated interview over the years I don't think ISI ever spilled a single bean about rF2 new sound features. In the latest wet track preview you can hear snippets of F1 cars driving away. It sounds absolutely like rF1 including the samples.

    If we can't have sound previews at this point can someone from ISI please tell us something about what to expect?
     
  15. Flaux

    Flaux Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,012
    Likes Received:
    422
    +1 I'm with you pandamasque
     
  16. Cavallino

    Cavallino Registered

    Joined:
    Jan 6, 2011
    Messages:
    14
    Likes Received:
    0

    Love these sounds, I think if there was more bounce / vibrato sound in tranny when launching off low gears it would be absolutely perfect.
     
  17. Cavallino

    Cavallino Registered

    Joined:
    Jan 6, 2011
    Messages:
    14
    Likes Received:
    0
  18. Antony

    Antony Registered

    Joined:
    Apr 2, 2011
    Messages:
    1
    Likes Received:
    0
    we would prefer this one

     
  19. 88mphTim

    88mphTim racesimcentral.net

    Joined:
    Sep 23, 2010
    Messages:
    10,840
    Likes Received:
    314
    Yes. Yes.

    :)
     
  20. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    Great and great! :D
     

Share This Page