OK Scawen, I understand you fully now. Just need DaveWS to redo his pack and all will be great again, or for now I can reduce the volumes manually. Perhaps I just have sensitive ears for clipping, it sounds horrible to me, one of the main reason I can't enjoy music at clubs, aside from the barrage of volume, is that it's so distorted and clipped I can barely make it out.
With what you were saying about the ear naturally soft clipping loud sounds, I always thought racecars sounded better from a small distance, so it could well be more realisitic to have soft clipping when incar to simulate the volume, although personally I'd rather do that with the volume control on my amp.
XsX! - it's all personal oppinion really, although higher end kit will show up any distortion and clipping in the signal more easily.
i know this is a bit of a thread hijack but it might still help scawen to work out the sound system a little better
is there anybody here who can explain why ears would soft clip ?
The ear is an analog system, with (effectively) a inverse log (however you would describe it) response to sound pressures at different frequencies. Therefore, the 'signal' recieved by the brain has very little differentiation between "loud" and "very loud" as compared to "quiet" and "normal".
Very loud is mostly not heard, but felt throughout the body.
Clipping is an artifact of representing an analog (boundless) signal in a digitally limited space. This obviously doesn't apply in a biological system except in the most extreme cases.
There also seems to be some misunderstanding of what "soft clipping" actually refers to.
Hard clipping can be represented in the worst case as a sound looking like:
__ __ __ __ _ __ __ __
Note that the transitions are vertical. This causes a large amount of additional high frequency 'transition' noise in itsself - the speakers literally stop with a bang. Also, it causes reverberations in the speaker cone itsself. In the worst case, this can destroy speakers.
Soft clipping makes the above picture look like:
_ _ _ _ _/ \_/ \_/ \_/
Which is much less hard on the speakers and creates less distortion, through 'curving' the top and bottom of the digital signal slightly - The actual value used is usually 1-2 bits of resolution, so has little impact on the actual sound.
i havent really found anything relating to clipping from skipping though the wiki article
but what you said sounds mole like compression than clipping to me ie the ear doesnt clip sounds but normalizes what you hear
which means that lound noises are heard quieter than they are but without distortion
With the ear working as described above, it does not / obviously cannot reduce the amplitude of incomming sound & therefore is not normalizing, rather it's more sensitive to quiet(er) sounds despite what's arriving at the ear canal. This is somewhat analagous to soft clipping, and likely greatly affects how we "perceive" very loud sounds. This, from what I can tell, is why Scawen went down this path. Obviously it's backwards, because the dyanmic range we can produce on a PC is so limited in comparison to the experience we want to reproduce - but it's definitely the correct approach. Your ear being analog has infinite ability to "soft clip", even though technically is not "clipping" at all per se, but rather more like an infinitely adjustable response curve to varying amplitude.
How else could one produce the illusion with such a limited dynamic range that we have to work with?
the ear itself being a mechanical/hydraulic instrument of course cant
what i actually meant is that the hair cells together with your brain reduce the sensation of louder sounds and amplify quieter ones which is in fact normalization
no its not cause clipping is an effect on the waveform itself not on the level
ive yet to see anything that causes distortion to the perceived waveform and not just the perceived level of the sound (other than masking but thats not the subject here)
theres the problem everything ive read so far relates to the level and not the amplitude
Actually that would be compression, since some dynamic range is lost in what you described. Normalizing would be just increasing the max amplitude to a some "normal" level, while preserving difference in volume in relation to other sounds.
I think you're describing the perception rather than the phenomenon that causes the said perception, at least in the latter case - no amplification can occur... I'm looking at it from the perspective of your ear being an "IO device".
The same sound will be perceived totally different at very high SPL in comparison to a reasonable SPL, which means it's being interpreted differently. Whether this is due to the brain interpreting it differently, or the ear passing the brain different information is irrelevant because the end result is the same. So it is semantics, but in a broad sense, your ear "is modifying the waveform" - note the quotes because it's physically impossible to literally do that, and yet the end result - at least in terms of perception - is similar with loud sounds... At least from my limited understanding.
I don't understand what you mean by that, could you explain it? (thought those were synonymous)
Clearly you would "pwn!!11" me with an acoustics discussion (i've been here awhile and read plenty of your posts) so I am here to learn as well, as I find this topic very interesting.
edit: Just to be clear, I think we should decide what we are going to focus on - perception, or rather what is literally happening. It would be easy to confuse the two and misunderstand each other. My basic contention is that what Scawen mentioned regarding how he's processing the sounds now is correct because it will help to create some of the perceptions that were previously missing, if you were not running full volume in LFS.
This has an easy work around.
Assign the Shift Up and Shift Down to 2 keyboard keys also. Then use the profiler of your wheel to send these keystrokes whenever you click the paddle shifters or the suquential chifter and it's done.
Yes and no. You´re of course right but its a matter of how you look at it. Normalisation usually refers to something you do on a a single track. So if you consider conversations with speakers talking at different loudness as sperate tracks to be normalised, the effect will be compression as related to your whole lifetime of hearing being a single track.
Either way I´m pretty sure this is what happens in your ear/brain. Maybe it´s just me but when I whisper I sort of know that the speaker is very quiet, but I don´t really perceive him as being quiet, as long as he´s still loud enough to be intelligable.
Googling turned up this:
The inner hair cells (IHC) of the cochlea, which convert sound to nerve impulses, have a dynamic range of less than 50 dB. But we can hear over a 120 dB dynamic range! How is this possible? It turns out that the ears have a built-in sound level compression system, created by the outer hair cells (OHC) of the cochlea. In the most active region of the cochlea basilar membrane, a 4 dB increase in sound pressure at the eardrum increases the membrane motion as little as 1 dB, due to mechanical action of the OHC.
Actually it is physically possible, by either limited space for the membrane to move or the hair cells themself limiting that movement.
But I´ve yet to find anything, that really supports limiting distortion below the point where the ears receive permanent damage, which could be seen as some very very hard clipping but thats not really the point of this discussion.
All non linearites of the ear I´ve read about so far, are intermodulations of fequencies close to each other, caused by the "spacial dispersion" one tone causes along the length of the membrane.
Level is a measurement of power so it basically refers to the mean of the waveform, whereas amplitude is the maximum of the wave.
Level limiting will squeeze and expand the waveform to a certain (quadratic) mean, without any change to the shape of it so the only distortion/loss is in dynmaic range (effectively what the volume slider does with v2).
Opposed to that amplitude limitation just cuts off everything above a certain threshold, causing a clearly visible change in the waveform as well as lots of unwanted high frequency components.
I´m not sure if those two are clearly sperateable.
But I can feel a loss of 10 FPS since I applied this patch on any cars any circuits, may change my graphical card soon (6600GT Max Details, no AA or aniso, 1280x1024)
Soo nice to see the hi res mirrors
I didn't see any FPS loss...at most it was 1 or 2 fps, and thats with my crappy old xp2400(@2266)/9700pro rig.
Im still amazed at how well this sim runs on my old machine. Even with Hi-res skies and loads of hi-res textures it flies. (1152*864 ,2*AA, 8*AF everything turned up full)
With a full grid it will drop to 25fps on the start line, but generally it keeps up a steady 50-60 fps up to over 100fps on certain tracks with a small grid.
Amazing bit of programming IMO
The wonder of directx 8, tbh i found making games in directx8 hella easier than directx9 with all its defining and declaring... how i will laugh when directx10 games are in full swing... when ill be on my linux box enjoying opengl and my 500 fps
Well, not completely ... I had to unlock my 2nd installation of LFS in my office (tssst ... dont tell anybody !!!) while I had no problems with my home's one.
I know, you are going to tell me that I shouldn't play LFS ant my office ...
Noticed that the previous screen shots in \shots folder are overwritten if you restart LFS and start capturing images again. Can you please make it so that it just adds up?
(6600GT Max Details, no AA or aniso, 1280x1024)
) while I had no problems with my home's one.