5.1 to Binaural

[Black5Lion]

Hi guys!
I'm a headphone user, as I'm sure many of you are. Except I personally prefer in-ear headphones
I'm currently using Audio Technica's CKS55 which sound AWESOME! but they're only stereo.
I know there are some "virtual surround" voodo software out there, but I want the real thing, or at least something that sounds real.
Something like this. I've read about the topic a bit.
This topic piqued my interest. From my understanding, you could theoretically turn any 5.1 recording you have into binaural format by playing it on a real 5.1 speaker configuration and recording it with a dummy head that has two mics positioned in the center of the room.
Here's what I've been thinking, we have game engine, and we have games that have things that sound in them. (like the radios in portal)
We should be able to theoretically make a "virtual room" with a "virtual 5.1 system" that basically received the 5.1 audio from a file, and played each channel in it's respective "virtual speaker". We'd then have a "virtual dummy head" or even an entire "virtual person" with two "virtual mics" inside it's two ears, it's then record the audio and save it in a 2.0 binaural format.

Is that possible?
If so why did no one make it? Can I make it with something like unity?

 

[spaceman]

Either the SB Z series or Smythe Realizer are 2 options. Pretty far apart price wise. Both do a good to amazing job of surround emulation.

 

[Nenu]

We should be able to theoretically make a "virtual room" with a "virtual 5.1 system" that basically received the 5.1 audio from a file, and played each channel in it's respective "virtual speaker". We'd then have a "virtual dummy head" or even an entire "virtual person" with two "virtual mics" inside it's two ears, it's then record the audio and save it in a 2.0 binaural format.

Is that possible?
If so why did no one make it? Can I make it with something like unity?

This is the voodoo that you dismissed.

 

[AVT]

This is the kind of thing that, if possible, will require a team of people with PhDs and a lot of money and resources to accomplish.

 

[Nenu]

Or just get a Sound Blaster Z/Zx/ZxR
http://www.legitreviews.com/creative-sound-blaster-z-zx-sound-card-review_155705/4

Games

Creative’s hardware virtual surround sound downmixing for headphone has already impressed us which we shared in our Sound Blaster ZxR review. Movie watching and gaming gain the most out of this capability since obviously these are the few applications that produce surround sound audio. Compared to other virtual sound implementations, the unnamed technology used in the Sound Blaster Z-Series cards produces the most convincing illusion of simulated rear channels. Of course headphones and headsets can’t compete with real surround speakers as far as positional audio is concerned.

 

[Skripka]

This is the kind of thing that, if possible, will require a team of people with PhDs and a lot of money and resources to accomplish.

Not really.

Last house I worked at, our audio engineer experimented with recording live acoustic shows with 5-7 microphones and creating 5.1 or 7.1 performance recordings. Even ghetto'd a mic armature together to mount them on. Getting the output product to sound decent took a crapton of individual mixing work...then you had the overhead of storage, as when you go from 2 channels to 6 or 9 channels your increase the scratch space for lossless performance recording in ProTools by, surprise surprise, 6X or 9X. Then finally when the end product got produced, it didn't sound much better as an archival recording of a performance than stereo on a good set of cans or full-size speakers. Our main problem was we always tried to keep 1-years worth of performances on HDD, which with 200 or 300 2 hours shows a year becomes very problematic when the higher ups don't want to buy HDD storage to even store stereo.


At the end of the day...for applications like just using headphones all the computation doesn't get you a whole lot compared to a good stereo implementation. IMHO.

 

[AVT]

Not really.

Last house I worked at, our audio engineer experimented with recording live acoustic shows with 5-7 microphones and creating 5.1 or 7.1 performance recordings. Even ghetto'd a mic armature together to mount them on. Getting the output product to sound decent took a crapton of individual mixing work...then you had the overhead of storage, as when you go from 2 channels to 6 or 9 channels your increase the scratch space for lossless performance recording in ProTools by, surprise surprise, 6X or 9X. Then finally when the end product got produced, it didn't sound much better as an archival recording of a performance than stereo on a good set of cans or full-size speakers. Our main problem was we always tried to keep 1-years worth of performances on HDD, which with 200 or 300 2 hours shows a year becomes very problematic when the higher ups don't want to buy HDD storage to even store stereo.


At the end of the day...for applications like just using headphones all the computation doesn't get you a whole lot compared to a good stereo implementation. IMHO.

But for things like accurate positioning in video games, you could in theory get a lot. See this: https://www.youtube.com/watch?v=IUDTlvagjJA

A lot of games turn the right headphone speaker completely off when the sound is coming in directly from the left, which is wrong, because the sound bounce off the opposite wall and then hit the right ear. I'm simplifying, but you get what I mean. No speaker should ever be fully off.

The problem is, how do you write an algorithm to do it in real time that doesn't suck? And, how do you convince players to tell the game whether they have speakers or headphones on?

 

[Black5Lion]

Either the SB Z series or Smythe Realizer are 2 options. Pretty far apart price wise. Both do a good to amazing job of surround emulation.

I'm thinking of a software solution.

This is the voodoo that you dismissed.

I don't think so, I mean it doesn't just downmix stuff to 2.0 with emulated surround, but it would account for the sound bouncing of the walls/objects in the room.

This is the kind of thing that, if possible, will require a team of people with PhDs and a lot of money and resources to accomplish.

Not really.
*snip*

Or just get a Sound Blaster Z/Zx/ZxR
http://www.legitreviews.com/creative-sound-blaster-z-zx-sound-card-review_155705/4

I''d really prefer a software solution even if it's not in real time.

But for things like accurate positioning in video games, you could in theory get a lot. See this: https://www.youtube.com/watch?v=IUDTlvagjJA

A lot of games turn the right headphone speaker completely off when the sound is coming in directly from the left, which is wrong, because the sound bounce off the opposite wall and then hit the right ear. I'm simplifying, but you get what I mean. No speaker should ever be fully off.

The problem is, how do you write an algorithm to do it in real time that doesn't suck? And, how do you convince players to tell the game whether they have speakers or headphones on?

I want this but not necessarily in real time.

 

[mindbomb]

The asus xonar sound cards with dolby headphone, and all those dolby 7.1 gaming headphones work on this principle.
https://www.youtube.com/watch?v=6wFdQPGxKIQ