I was thinking, what might next generation surround sound be like? With 2.1, the sound can only appear to be coming from somewhere between the two speakers. 5.1, 6.1, and 7.1 solve that by adding speakers all around to get a 360 degree sound image. The only problem is that their is an empty void above and below the listener's head. Everything is located in a horizontal plane running through the listener's head. While movie producers can do some neat tricks to make it seem like something is travelling over head without speakers at different heights, these definitely don't give any real sense of immersion. If, however, there were speakers at different hieght levels (and recordings that could take advantage of that fact) the listener would be able to close their eyes and easily form a 3d image (not planar like with current surround sound formats).

http://img284.imageshack.us/img284/6319/84surroundbg3.png

There are eight woofer/midrange/tweeter speakers each located in each corner of the room. The four speakers with the red dots are subwoofers (to allow for more effective directional effects). The listener is in the exact center of the room (in width, length, and height) and the room is as close to a cube as possible.

In the future, more standards such as 14.6 (an aditional surround on each of the four walls and on the ceiling and floor and a subwoofer directly above and below the listener) could be added, making the speakers closer and closer to forming a sphere.

So, what do you think? Should this be adopted as a new standard? What should be different about it?

Would love a height channel!!! But, let's get to real 7.1 first before proposing even more speakers. BTW, I think you mean 'immersive'.

Simon

I think a 4.2 system would be doable if you were in the center of a tetrahedron.
http://en.wikipedia.org/wiki/Image:Tetrahedron.gif
Speaker in each corner with 2 subs to cover the lower and mid bass ranges.
Now all i have to do is find the right shape room:)

http://img153.imageshack.us/img153/9673/209814oz6.png

what about 20.4?

scratch that.. 20.4.4

4 mini surrounds/jetpacks for the chair

So, what do you think?Remember, you asked.

The centre of the room is probably one of the worst locations to place the listener (very uneven frequency response). Likewise, a cube is one of the worst shapes for a room (will re-inforce peaks and dips around the same frequencies rather then spread them out).

As for speaker placement: you have to keep in mind that human hearing changes with direction. A symmetrical speaker layout (like the one you proposed) doesn't take that fact into account. For example: our ability to create phantom images is worst at our sides (because the opposite ear is in acoustical shadow and isn't helping with localization), so it really helps imaging stability to have speakers directly to the sides of the listener. Your set-up doesn't have any speakers at those location. A centre speaker up front is beneficial for soundstage depth, stable central imaging, better vocal tonality, minimizing comb filtering, etc. You don't have a speaker at that location.

Placing speakers in the corners might be OK behind the listener, where muddying the sound will create a diffuse and enveloping surround field. But in the front soundstage, where the goal is articulation, the speakers should be away from room boundries. Ironically, you've placed all the front speakers where three boundries meet, which is probably the worst location for them. To help articulation and clarity, the front speakers should be at or near ear level, which none of your speakers are.

Most of the localization in your set-up is achieved via phantom imaging, which is a bad idea since it is inherently unstable. Locations like your sides or directly in front of you should have a hard sound source (for imaging stability) rather than rely on phantom images.

One thing about your set-up that will work very well is the subwoofer placement. Putting them at the midpoint of each wall will help create smoother bass response that is consistent across multiple seats. Unfortunately, none of the other speakers are good locations, especially up front.

If you take into account how our hearing works and build upon already established speaker configurations, you can come up with something that will be a better standard than current 5.1-speaker set-ups.

Good Luck,
Sanjay

Don't forget the butt kicker or bass shaker as part of the chair

4 mini surrounds/jetpacks for the chair
ooooo, I hadn't thought of that. I thought we were going to have to start buying furniture with cables. Although the sling approach might provide a more realistic experience when watching porn.

Just use headphones :-)

**ducks**

IMHO, the most glaring omission in current MCH technology is proximity. Even in the THX 20.4 demos, the sound are around and above but never RIGHT HERE! Adding that capability would drastically transform the ability to recreate a live event.

Just use headphones :-)

**ducks**

Actually, I was thinking about that. The listener could have high quality headphones hooked up to a gyroscope and motion sensors. As the person rotates their head and walks around the room (which, by the way, is the only REAL way to tell the difference between stereo and surround sound) a computer could caclulate in realtime what they would actually be hearing. Since people only have two ears, this could surpass even the best surround sound system.

The channels in the recording would represent instruments in different locations (specified in the recording), not just left, right, surround, etc. speaker channels.

Imagine this:
You put on an blindfold and a large pair of headphones. You hear a bell to your left (your brain knows this because the sound hit your left ear first). You start hearing more and more bells ringing up. You turn around, dazed, and eventually you are able to pinpoint the exact location of all the bells. You walk towards one of them. You know you are right next to it because it is much louder than the rest of them. You then walk out of your living room. As you walk out and the sound of the bells is quiet behind you, you take off the blindfold. You gaze in looking for the invisible bells you hear ringing, suspended in mid-air in your living room. You can look at the exact point where you hear a bell ringing, but you see nothing.

Click on the attached image. :)

This shows what I am talking about. The red boxes are channels. Each of them represents a virtual sound source which can be located anywhere in the sound stage. The sound stage is marked by the blue IR emmitters which help the headphones know where they are. The large green box is the audio processor which takes the information gathered by the headphones to figure out what to do with the audio channels. As the person moves though the room, this audio processor figures out what the persons ears should be hearing. Think of it as a sound hologram.

The important thing to remember is that the red boxes don't represent speakers like traditional audio channels do. They represent sound sources (instruments, people, objects). There can be as many of them as the audio processor can handle. Unlike with speakers, they can be located anywhere within the sound stage.

If this is actually is developed, people could save tons of money on speakers, and get much more holographic sound.

I did some research, and the headphone technology I thought of has already been patented by Sony: http://www.google.com/patents?vid=USPAT6108430&id=RE0EAAAAEBAJ&printsec=abstract&zoom=4&dq=6108430#PPA1-IA1,M1

They haven't developed it yet though.

Edit: Nevermind. This is more of a Virtual Surround type technology to make sounds appear as though they are coming from outside the head. It does have headphone rotation, but not much more. My idea is still not patented AFAIK. So, who wants to help me work on this? :)

I have never been clear on how binaural hearing deals with altitude. Someone suggested the shape of the ear canal allows the brain to get some sort of altitude spacial clues.

The shape of the ear might change the way audio sounds from different directions. This would be a thing to add to the audio processor if this is true.

The brain is an amazing thing. It is tuned perfectly to be able to tell as much information as it can from just the sound coming into the ear. The human brain and ear have evolved together to be able to make certian assumptions. For example, high pitched sounds are instantly assumed to be at a higher altitude than low pitched sounds (high pitched sounds would be from things like birds). There is a part of the ear that is normally associated with balance that lets people know which direction is up. Using these two pieces of information along with the delay and volume difference and muffling from one ear to the other, people can tell the direction of high pitched sounds very well. Movie makers can use this assumption of altitude based on pitch to let the viewer know whether a plane is flying under then or over them in surround sound, even though all the speakers may be at ear level.

When people hear a high pitched sound, however, they don't just keep this assumption. Just look at what birds do when they hear other birds singing, they turn their heads. This is a precise movement that the birds use as a reference. The bird's brain uses the information of how much the birds head turned and how the sound changed in each ear to determine the location for sure. This is one of the things left out by normal binaural imaging. Long story short, people can't tell the altitude of a sound for sure with normal binaural sounds, they just base it on assumptions which are tuned to be as accurate as they can be. My idea fixes this.

Through some of my own tests, I have found out that almost all of the information you get about sound is independent from what you see/feel/know. That means that even though you might know for a fact that a sound is coming from two speakers, you will still hear the sound coming from what your ears and the aural seciton of your brain determine is the sound source. This means that you can completely replicate any audio you want with just headphones and a way to tell head position and rotation. Just knowing that a person is on your right doesn't affect very much where you hear the sound coming from.

Binaural audio is very related to this. After listening to tons of binaural audio online (you need headphones, not speakers), I have say that using headphones instead of speakers is probably the best idea if you want to be able to pick out the location of every single sound. Speakers just can't compete with headphones in this way.

Play this with headphones (good quality) and then on speakers: http://upload.wikimedia.org/wikipedia/commons/3/36/BinauralPaper.ogg
Speakers just don't compare to headphones with this.

Now the question is how to simulate this with headphones, IR tracking, and mono recordings using the position and rotation of the head relative to the sources.

Too much wall of text for me to follow along but how did you make it to 8.4? Did you double
up the speaker wire on one binding post or what? I always wanted to get that more than
the normal speaker arrangements. Plus, I know this is a stupid question but is it matrix
sound or playing mono on 2 speakers?

Too much wall of text for me to follow along but how did you make it to 8.4? Did you double
up the speaker wire on one binding post or what? I always wanted to get that more than
the normal speaker arrangements. Plus, I know this is a stupid question but is it matrix
sound or playing mono on 2 speakers?

I haven't actually "done" anything yet besides research. The original thread was trying to think up a better surround format to beat todays sound formats. The problem with the sound formats today is that they don't convey the height of a sound source very well. They are pretty much limited to a disk of sound which they can't go above or below. The orange disk represents the height that all sounds seem to come from with current surround sound.
http://img157.imageshack.us/img157/4215/unhappylo2.png
I originally thought about 8.4 surround where there would be speakers above and below the listener to give a much better sense of the height of sound. This would be a very easy solution.

But then someone talked about how you also need to have proximity of sound. Speakers across the room can't emulate something whizzing past your ear like headphones can. One solution would be to just have more speakers closer to the ear for those times when a sound is supposed to be closed to you, but that just wouldn't be the same as it actually being there. That's when I came up with the idea of a headphone surround type thing. This is how my headphone surround would work:

(this is the part you should read)

1. During recording, microphones are placed next to each sound source and the position of the microphone is recorded.
2. All the microphones have their own discrete channel, so each instrument has their own mono track (instead of each speaker like with today's surround sound)
3. During playback, the listener wears headphones with gyroscopes and IR to track the location and rotation of the person's head.
4. Using the position and rotation of the person's head, a computer calculates in realtime what each ear would be hearing of the different mono tracks (the farther away the ear, the longer the delay and the quieter the sound and vice versa)
5. The processed audio is amplified and sent to the headphones.

Basically the sound sources are like virtual speakers, and there can be as many of them as the computer can handle. Since we only have two ears, the computer can give positional audio with just the two speakers in headphones using binaural imaging, and there is no way to tell the difference.

This doesn't exist yet, but you can try listening to some binaural audio for the next best thing. (binaural audio is recorded differently than normal stereo audio and is meant for headphones)

I haven't actually "done" anything yet besides research. The original thread was trying to think up a better surround format to beat todays sound formats. The problem with the sound formats today is that they don't convey the height of a sound source very well. They are pretty much limited to a disk of sound which they can't go above or below. The orange disk represents the height that all sounds seem to come from with current surround sound.
http://img157.imageshack.us/img157/4215/unhappylo2.png
I originally thought about 8.4 surround where there would be speakers above and below the listener to give a much better sense of the height of sound. This would be a very easy solution.

But then someone talked about how you also need to have proximity of sound. Speakers across the room can't emulate something whizzing past your ear like headphones can. One solution would be to just have more speakers closer to the ear for those times when a sound is supposed to be closed to you, but that just wouldn't be the same as it actually being there. That's when I came up with the idea of a headphone surround type thing. This is how my headphone surround would work:

(this is the part you should read)

1. During recording, microphones are placed next to each sound source and the position of the microphone is recorded.
2. All the microphones have their own discrete channel, so each instrument has their own mono track (instead of each speaker like with today's surround sound)
3. During playback, the listener wears headphones with gyroscopes and IR to track the location and rotation of the person's head.
4. Using the position and rotation of the person's head, a computer calculates in realtime what each ear would be hearing of the different mono tracks (the farther away the ear, the longer the delay and the quieter the sound and vice versa)
5. The processed audio is amplified and sent to the headphones.

Basically the sound sources are like virtual speakers, and there can be as many of them as the computer can handle. Since we only have two ears, the computer can give positional audio with just the two speakers in headphones using binaural imaging, and there is no way to tell the difference.

This doesn't exist yet, but you can try listening to some binaural audio for the next best thing. (binaural audio is recorded differently than normal stereo audio and is meant for headphones)


... I still don't know what you are talking about. Did you read my question or is
it just me? :p

... I still don't know what you are talking about. Did you read my question or is
it just me? :p

It's not matrix, and it's not just playing mono on two speakers. It is an entirely new format I am proposing.

The key difference is that audio is stored by sound source, not by speakers.

Typical stereo, 5.1, and 7.1 store the exact information that each speaker will play. Each of the channels represents a speaker (left-front, center, right surround, etc.).

What I want to do is have each channel in the recording represent a sound source (lead singer, bass guitar, drum 1, drum 2, guitar, etc.). Then your audio system at home will determine how to emulate that using the speakers you have.

Understand?

Concerning your original diagram:

- why the glaring omission of center channels?

- Isn't subwoofer non-directional?

- How do you get the chair to hang in the air like that?

- How do you get into the chair hanging up in the air?

- Won't the huge hole in the corner of the room be a problem?

Concerning your original diagram:

- why the glaring omission of center channels?
I hadn't really though about that. I just assumed the four speakers already in the front would be able to make up for the lack of a center, but now that you mention it, a center probably would help.

- Isn't subwoofer non-directional?
To an extent. People certainly have a much tougher time figuring out direction with low frequencies, but I think targeted subwoofers wouldn't hurt, if not help with LFE in movies. For music, it probably wouldn't matter too much.

- How do you get the chair to hang in the air like that?
In reality, the chair would be on some sort of platform. If you don't want a platform affecting the sound, it could be suspended with steel cables.

- How do you get into the chair hanging up in the air?Ladder?

- Won't the huge hole in the corner of the room be a problem? I needed to have that to make the drawing clear. Without it, it was hard to recognize the positions of various things.

It's not matrix, and it's not just playing mono on two speakers. It is an entirely new format I am proposing.

The key difference is that audio is stored by sound source, not by speakers.

Typical stereo, 5.1, and 7.1 store the exact information that each speaker will play. Each of the channels represents a speaker (left-front, center, right surround, etc.).

What I want to do is have each channel in the recording represent a sound source (lead singer, bass guitar, drum 1, drum 2, guitar, etc.). Then your audio system at home will determine how to emulate that using the speakers you have.

Understand?

If it's not matrix and if the source isn't 8.2 surround.... then how could it be discrete?
I mean I understand your explanation but I am just confuse if it's matrixing it or the
source has a discrete 8 channel. So to me what you are saying is emulating means
Pro-Logic to me and it must be matrix.

If it's not matrix and if the source isn't 8.2 surround.... then how could it be discrete?
I mean I understand your explanation but I am just confuse if it's matrixing it or the
source has a discrete 8 channel. So to me what you are saying is emulating means
Pro-Logic to me and it must be matrix.

Typical recordings, are discrete in terms of playback.
With this, it is discrete in terms of the recording.

With both, there are discrete audio channels (unlike prologic), but the meanings of those audio channels are different.

The format I am proposing would have an unlimited number of discrete channels. When something is recorded, each sound source would have its own discrete channel.

Let's compare a 6-channel recording using my new format, a normal 5.1 recording, and a prologic recording all of the exact same event:
http://img248.imageshack.us/img248/2390/comparekz3.png
The colored circles are the sound sources. The black boxes with the pipes are the recording medium. The ovals to the right are speakers. The gray boxes indicate audio processing.

The first one is the one I am proposing. Each sound source is recorded discretely. The mixing is done in realtime, not before recording, so the colors (different sounds) remain in tact on the recording. In this example, a person is listening using headphones. If someone wanted to, they could bypass the audio processor and use an amplified speaker for each discrete channel. With proper speaker placement, the event could be exactly replicated.

In the second one (normal 5.1) all the sound sources are mixed together differently for each channel making discrete channels of mixed audio. Each discrete channel has information from many sound sources (which is why the recording has brown pipes). The channels are discrete, but the sound sources aren't, and there is no way to undo this.

The third one is prologic. You can see why it is different and bad.

Here is the other way to set up a system with my proposed format: http://img265.imageshack.us/img265/5992/imagemb0.png
Notice the lack of all audio processing and mixing.
The speakers would need to be placed differently for each recording, but it would give superior sound.

Do you see how my format would be better now? If you don't, consider this:

With normal 5.1 recordings, the location of every singe sound is emulated by multiple speakers playing a slightly different version of the same thing. The sounds are rarely supposed to be coming from where one of the speakers is actually located.

With my format, this emulation is completely controlled by the equipment of the listener. The information is isolated and preserved longer than with typical surround sound.

I found out that the if the distance away from a sound source is multiplied by n, the Pa of sound pressure is divided by n. When the Pa of sound pressure is divided by two, the dB-SPL decreases by 6.02.

The speed of sound through air at STP is about 340m/s. If the distance between the two eardrums is 1/6 meter, then the maximum delay from eardrum to eardrum would be about 0.5ms.

When sound travels through the head, the high frequencies will be impeded more than the low frequencies.

These three things will be useful to figuring out what each ear should hear.

OK, I understand now.

After talking with some people about this and thinking it over more, another major affect on the sound would be how the head affects the waves of high and low pressure. As the waves hit the head, they go around it more than through it. This means that the ear that isn't facing the sound source will hear a version much different than the ear facing the sound source.

One of the major effects would be high frequency attenuation.
http://img372.imageshack.us/img372/9081/sounday2.png
With low frequencies, the pressure wave has a chance to envelope the head and get to the ear facing away from the sound source, but the high frequencies, the wavelength is too short to do much of anything to the ear drum facing away from the sound source. I hope my picture helps.

I made another diagram: http://img337.imageshack.us/img337/1192/speakereo1.png
The darker gray is high pressure, and the lighter gray is low pressure. The white region is where the pressure change from room pressure would be the least. If you imagine the bars of pressure being larger than the head, the ear in the white region would get more of that pressure change than if the bars were thinner.

Does anyone know where I can find more solid information on this type of stuff?

Most of what I have been doing is just BSing (theorizing), but I would like to see if other people have actually done experiments to determine if some of the things I think are true actually are true (like high frequency attenuation for the ear opposite the head).

Actually, I was thinking about that. The listener could have high quality headphones hooked up to a gyroscope and motion sensors. As the person rotates their head and walks around the room (which, by the way, is the only REAL way to tell the difference between stereo and surround sound) a computer could caclulate in realtime what they would actually be hearing. Since people only have two ears, this could surpass even the best surround sound system.

The channels in the recording would represent instruments in different locations (specified in the recording), not just left, right, surround, etc. speaker channels.

Imagine this:
You put on an blindfold and a large pair of headphones. You hear a bell to your left (your brain knows this because the sound hit your left ear first). You start hearing more and more bells ringing up. You turn around, dazed, and eventually you are able to pinpoint the exact location of all the bells. You walk towards one of them. You know you are right next to it because it is much louder than the rest of them. You then walk out of your living room. As you walk out and the sound of the bells is quiet behind you, you take off the blindfold. You gaze in looking for the invisible bells you hear ringing, suspended in mid-air in your living room. You can look at the exact point where you hear a bell ringing, but you see nothing.

This technology exists, and it exists in two channels. It's called a binaural recording. It's done by using two omnidirectional microphones inside a dummy head, typically. The best example I've ever heard is called Kunstkopfumgang, a German/English recording. If you'd like a sample (it's not copyrighted anywhere as far as I know), and you've got some decent headphones to give it a listen with, PM me for a copy. It really is out of this world.

I have never been clear on how binaural hearing deals with altitude. Someone suggested the shape of the ear canal allows the brain to get some sort of altitude spacial clues.

Actually, sound location in all dimensions, including altitude, is more a function of the pinna (the outer ear) than the auditory canal, although it is true the arrival time differences is another important factor. Reseach in audiology has indicated that people whose outer ear is lost experience diminished sound localization abilities much like people who lose one of their eyes experience diminished depth-of-field perception.

Doug

I think you are right YerDugliness. This really supports what you are saying: http://www.waisman.wisc.edu/hdrl/take_Test.html

This technology exists, and it exists in two channels. It's called a binaural recording. It's done by using two omnidirectional microphones inside a dummy head, typically. The best example I've ever heard is called Kunstkopfumgang, a German/English recording. If you'd like a sample (it's not copyrighted anywhere as far as I know), and you've got some decent headphones to give it a listen with, PM me for a copy. It really is out of this world.

I know what binaural audio is, I was trying to go a step beyond. Instead of just capturing binaural audio, I want to make a format that will allow you to create it in realtime based on the position of your head.

With normal binaural audio you are stuck hearing exactly what the dummy/person heard at the time of the recording. If you get up and walk around the room, you will still hear exactly what was recorded, and that won't be what you would hear if you were to have walked around the room that the performance took place in.

If I can find a way to closely model what happens to audio from the source to the brain, I could make a program which would take any number of mono sources and change them in realtime to match what someone should be hearing based on the position and rotation of their head.

I think you are right YerDugliness. This really supports what you are saying: http://www.waisman.wisc.edu/hdrl/take_Test.html

Thank you, greeniguana00. M.S. in Speech Pathology with an emphasis and real interest in audiology. Too bad my university didn't offer a M.S. in audiology at the time.

I think I understand what you're trying to accomplish. Imagine being at a concert, 8th row center seat. You're watching the vocalist who is located center stage and the sound appears to be spread out in front of you. Then someone throws a bra at the stage from your left, so naturally you turn to look at the originator of such a unique action. Now the sound seems to be coming from your right rather than from in front of you. The location of the sound origin hasn't changed, only the orientation of the listener's head.

Is this the effect you're trying to achieve? If so, it appears to be an audio variation of the "virtual reality" visual experience. Perhaps the researchers in that field have already developed positional sensors (probably accelerometers) and an algorhythm that can be of use for your purposes and with binaural recordings it might be possible with good headphones.

Good luck--this sounds like a very interesting quest!

Doug

I know what binaural audio is, I was trying to go a step beyond. Instead of just capturing binaural audio, I want to make a format that will allow you to create it in realtime based on the position of your head.

With normal binaural audio you are stuck hearing exactly what the dummy/person heard at the time of the recording. If you get up and walk around the room, you will still hear exactly what was recorded, and that won't be what you would hear if you were to have walked around the room that the performance took place in.

If I can find a way to closely model what happens to audio from the source to the brain, I could make a program which would take any number of mono sources and change them in realtime to match what someone should be hearing based on the position and rotation of their head.


So let me get this straight; what you want to do is create a virtual environment, and have it sense the position of the speaker, and make sound accordingly, located where it should be in the room? If that's the case, all you would really need is a computer, a regular pair of (preferably wireless) headphones, a really good motion sensing system, a binaural recording, and a program that would accomodate for adapting the recording. Unless of course, you want there to be things to walk around and bump into that will also make sound, which would require the computer to be able to generate those sounds on its own.

The best equivalent would be the 3d sound in some computer games. The sound is recorded by source, and then what actually gets sent to the sound card is created in realtime based on where you are relative to that sound in the game. So the recordings would not be binaural (they could actually have any number of channels), but what you would hear would be binaural after it goes through processing.

Basically, what I am proposing is very similar to the sound in games besides the listener wearing motion sensing headphones instead of controlling a character.