Well: You could try the RightMark 3DSound: Positioning Accuracy test and see how that works out for you. You'll be able to compare the effects of DS3D Software, DS3D Hardware, and the various attributes of EAX with regards to how they impact the sound positioning abilities of your hardware and speaker setup.K.I.L.E.R said:
HRTF and soundcards
- Thread starter K.I.L.E.R
- Start date
Here's a good example of what very good HRTF sounds like(it's recorded with a dummy human head with microphones at the location of the ears): http://www.holophonic.ch/archivio/testaudio/Cereni - Holophonic.mp3
Sweet :smile:
holy shit.. depth...
uh ya.
You have 2.1 then, not two speakers only.. the .1 part is important for movies
Whats your setup anyway?
I have some shitty logitech 5.1 setup.. the z5300.
My parents 2 10+ year old dsm time window 3 towers hit lower.
Kinda strange.
They are ported though.
holy shit.. depth...
K.I.L.E.R said:
uh ya.
You have 2.1 then, not two speakers only.. the .1 part is important for movies
Whats your setup anyway?
I have some shitty logitech 5.1 setup.. the z5300.
My parents 2 10+ year old dsm time window 3 towers hit lower.
Kinda strange.
They are ported though.
Last edited by a moderator:
You are forgiven.K.I.L.E.R said:
I think you still have a fuzzy idea about the nomenclature here. I'm willing to bet that you can hear both loudspeaker with both ears. That is, if you only play through one speaker, both ears will still hear sound. If so, HRTF is irrelevant for you, because your head/chest/hair is already shaping the sound that your ears pick up. No need for simulation.If you play through the speaker on the left for instance, your right ear will pick up the sound, but with (frequency dependent) attenuation, chest/shoulder reflections added and a phase difference between the left and right ear, all of which help you determine where the sound originates.
Where HRTFs make sense is with headphones, where the sound that is picked up by for instance the left ear is not affected by your head and body attenuating and reflecting the sound depending on sound source position. The HRTF simulates this to some degree, preferably in an adjustable manner, because people differ, even their hair style and clothing has a significant effect, not to mention the shape of their ears.
So if you are talking about loudspeaker reproduction, forget about HRTFs. It's not really relevant. There are ways of faking 3D soundfields with speakers, but none of them are very good, because the reflections of the space you are in will always be superimposed on the direct path sound. This is true for multi speaker setups as well. With stereo speakers, you can add reverb to simulate depth, you can play little games with phase to disconnect the sound from the speakers a bit but.....honestly you'd be better off simply forgetting about trying to simulate 3D sound with stereo speakers and instead go for good stereo sound, and let your mind, your vision and the game environment help you position the sounds correctly. The mind and eyes are very active in how we humans percieve nd position sound.
A 2-channel binaural recording carries all the information needed (in fact, all the information possible) for accurate sound space positioning. I've done binaural recording on an amateur basis, and the result can be startlingly realistic reproduction of soundspaces in a way no loudspeakers in rooms can do. But such recordings sound narrow/flat (due to their low stereo separation) and somewhat coloured on speakers, and since speakers are the accepted standard for sound reproduction we're stuck.
Don't sweat HRTFs. In fact, if you don't use headphones, my recommendation is that you just forget about it completely.
Last edited by a moderator:
cloudscapes
Regular
Graphics are easily marketable with screenshots and video clips. It's apparent to the client (the gamer) that improvements need to be made to graphics hardware just by glancing at them. Audio on the other hand, is really difficult to market. Magazines won't work, TV programs would require living room setups to be perfectly balanced (for positional audio) and websites need to excpect the user to have good speakers/headphones. It's a snowball effect right now. The craze is graphics and more recently physics. Hopefully truly revolutionary audio will follow eventually.
I'm one of those types who are extremelly dissapointed in Creative for what they've done to Aureal and their tech. Even software emulation now wouldn't be that hard, would not require a huge hardware redesign, and it certainly blows EAX out of the water. The whole legal issue is Creative want's to push their own format and convince the world that there's nothing better than EAX, and they're succeeding.
The other snowball effect is that devs are stuck on EAX. I don't knwo if it's the easiest implementation, but if a game has the EAX 7.0 SUPER HD EXTREME GAMER WITH FREE BLOWJOB EDITION logo on the box, it's profitable for the publisher. Devs aren't always willing to take the risky road and try something different. The best solution would be to support multiple technologies.
Hopefully by spreading binaural MP3s that show "what could've been" around, and educate people, a big enough movement can be created to force either Creative or another competitive hardware manufactuer to pick up where A3D left off and blow our minds.
I'm one of those types who are extremelly dissapointed in Creative for what they've done to Aureal and their tech. Even software emulation now wouldn't be that hard, would not require a huge hardware redesign, and it certainly blows EAX out of the water. The whole legal issue is Creative want's to push their own format and convince the world that there's nothing better than EAX, and they're succeeding.
The other snowball effect is that devs are stuck on EAX. I don't knwo if it's the easiest implementation, but if a game has the EAX 7.0 SUPER HD EXTREME GAMER WITH FREE BLOWJOB EDITION logo on the box, it's profitable for the publisher. Devs aren't always willing to take the risky road and try something different. The best solution would be to support multiple technologies.
Hopefully by spreading binaural MP3s that show "what could've been" around, and educate people, a big enough movement can be created to force either Creative or another competitive hardware manufactuer to pick up where A3D left off and blow our minds.
Chalnoth said:
Again, I think there is a problem in nomenclature.
As I said, and you are aware, there are things you can do to simulate a 3D sound space from stereo speakers. But using HRTF as a label to describe it is not very appropriate, since the term is well defined for headphone reproduction, but doesn't really make sense in the case where the sound is affected by your whole body. A HRTF simulates this response, which, for someone listening to speakers, is already there. Now the stuff you can do (and that I feel shouldn't be referred to as using HRTF simulation) to simulate 3D response from speakers is iffy at best. Since you are not using headphones, the phase information is going to get screwed up if you move your head or turn it, and the sound space simulation (reflections/reverb) is going to clash with the natural reflections of the actual room you are sitting in and not mesh with the visual iformation you pick up. While admittedly you can fake some semi-decent effects, they are just too damn sensitive to the practical particulars. You are not sitting with your head fixed in a dentist chair in an anechoic chamber. There have been experiments with DSPs with mics that pick up the room reflections and try to cancel out the first order reflections. And that's a step forward, but then this cancellation does not take into account the second and higher order reflection, only works in one specific spot, and messes up the sound proportionately everywhere else, is subject to mic and looudspeaker nonlinearities... It's a nightmare.
You can use these tricks, fiddle around with them a bit, say "wow, this is cool" and game on, or you can just ignore it and game on. To me, the result of the second alternative will typically be better all things considered. Our positional perception is very dependent on visual context and expectation, so things will tend to work out just fine gaming wise under any circumstances.
As I said, except for headphones, I'd simply ignore it, and if the HRTF algorithms don't do a decent job of simulating your particular ears, you might well be better off not using the processing even with headphones. I have the X-Fi Elite (curse Creative three times and throw salt on their grave for their actions with Aureal and their tech), which is as good as it gets at this point, and I typically don't use HRTFs even for headphones, even when gaming.
YMMV though. For something like this, people should do their own experimentation and decisions based on preference.
No there isn't. HRTFs can be used very well with loudspeakers, in principle only crosstalk cancellation needs to be added compared to headphone listening. HRTF stands for head-related transfer functions and the bulk of the effects comes from the head, including the earlobes. The effect of one's body is minuscule compared to the effect of the head and I don't know if such an effect is really incorporated to HRTFs used in sound cards. If it is, then it can be easily disabled for speaker listening, just as crosstalk cancellation is added for speaker listening.Entropy said:
Well, you're right, Entropy, in that the environment is much easier to control for headphones, and thus the experience will always be better with the same amount of work put into the HRTF with headphones as with speakers. But I distinctly remember being able to think I was hearing things happen behind me from a 2-speaker setup on my Vortex 2. Granted, it wasn't a perfect effect, but it was vastly better than anything I've seen since.
Bolloxoid said:
Hold on.
Yes crosstalk cancellation is the first step if you want to try to emulate a soundfield using speakers, but the scheme fails before it even reaches first base. You cannot achieve crosstalk cancellation with speakers in a room. Cannot. Don't take my word for it, prove it to yourself by feeding a mono white/pink noise signal to both loudspeakers, connected out of phase. If it were possible to achieve even step one, you should be able to find a spot which was silent. You will fail. You will find places where the noise is a bit lower. Fine. Now move your head. Notice how the noise not only gets a bit louder but will typically vary in timbre. And this is with perfect out of phase feeds that takes into account the frequency and phase response of the specific speaker you use.
Sobered up yet?
Electronic cancellation typically tries to cancel only the direct sound, which it can do a (crummy) approximation of in one spot only. Move or turn your head (and the distances are quite short), and you will get a cancellation or amplification of the sound depending on frequency and position. Yummy. And then we still haven't taken into account reflected sound - against walls, floor, roof and your body. These reflections just get added on top, but will be disproportionately strong if you were at all successful in cancelling the direct sound, creating an unnatural effect for the room you're in. (Canton once demonstrated a very expensive system that attempted to cancel the first order wall/roof/floor reflections. By and large a failure, for many reasons.)
On top of this mess, you will now try to superimpose frequency/phase information that may or may not correspond well to the particular listeners body, head and ears. Good luck. As I mentioned, I have done binaural recordings on an amateur basis (obligatory name dropping: I almost bought an Aachen head from Ken Kantor of NHT fame. Couldn't justify the outlay, unfortunately.) I've spent a fair amount of time trying to decently replicate the binaural experience electronically. I was largely unsuccessful. As are application specific DSPs from AKG and Sennheisser. I've plowed the Journal of the Audio Engineering Society from the 70s forward. I've talked to people, listened at trade shows and in the homes of DSP wizards. Recreating a 3D soundfield with headphones is difficult enough, but there you have some theoretical hope. Trying to reproduce a 3D-soundfield with a pair of loudspeakers in a room is so impossible I just can't find words. It cannot be done.
That is not to say you can't produce some phasey sound that your brain will interpret as spacious, and adding visuals will greatly aid you in judging directions. But in no way, shape or form is this "correct", and I'd say that in most cases you will do just as well if not better by simply reproducing stereo sound over your stereo speakers. Your brain and eyes will still help you with positioning, and the sound will be a better match for your equipment.
If someone who reads this is at all interested in 3D soundfields, I can't recommend anything higher than to buy a couple of stealth omnis ($80-250) that you can clip to your glasses close to our ears, and go out and record the world. Experiment. Compare. Think. Read. It's very, very interesting and educational. You will be using your own "dummy"
head, so there will be no mismatches, and the recording will take place outside the pinna, so the sound will only go through one set of outer ears, not two as in most binaural demonstration recordings (first that of the dummy head, and then yours).I sound very authoritarian in this post, more than I'd like, but then I have really gone through all of the above myself, and please see it more as a sign of my still strong appreciation of binaural recording methods. With that background it's impossible for me not to regard 3D-sound via speakers as a very distant second best. Sorry about that.
This is too simple a situation. Of course it will fail because absolute cancellation only occurs at one point, but we have two ears.Entropy said:
Yes, there are problems with using HRTF's with speakers. But that doesn't mean it's a pointless exercise. The Vortex2 already showed tangible 3D positioning from two speakers, and that was developed quite some time ago.
Similar threads
