| AudioREVIEW's Forum Archives - Home Theater |
|
|   For virtual surround, how does Psycho-acoustics Effects work | Smokey
Apr 29, 2003 7:41 AM | | There are virtual surround sound algorithms that make use of only left and right speakers and psycho-acoustics effects to emulate true surround sound formats (like for TVs). So my question is what are psycho-acoustics effects and how is two channels is used in conjunction with it to create "virtual" surround sound?
Thanks. |
| Combination of several things | Sir Terrence the Terrible 1
Apr 29, 2003 1:09 PM | | Smoke,
I will try and keep this as digestable as I can. The process of virtual surround involves timing, cross vector cancellation, and eq to make it work.
First, you have to be sitting exactly in between the two speakers to make sure that the signals from each speaker reach the ears at the same time. A few inches to the left of right and you disturb the simultaneous arrival of the processed signals.
Secondly cross vector cancellation is applied to each channel to reduce crosstalk. When our ears hear signals from spaced speakers we hear two components. The signal from the speaker closest to the corresponding ear, and a partial signal of the speaker from the opposite ears as it travels around the head. This reduces the size of the soundstaging, and reduces the detail within the boundaries between the two speakers. To combat this you must add a cancelling component to each channel. This means placing a bit of crosstalk cancelling signal from the right channel into the left speaker and visa versa. By doing this you eliminate the crosstalk and open up the soundstaging.
Eq is then applied to cancel any frequency response errors cause by suspending phantom images between speakers. This includes the center, and surrounds signals.
All of this processing can be done either via DSP algorythms, or a set of indentical drivers on the speaker itself like Polks Stereo Deminsional array that actually feeds the cancelling signal directly from a set of drivers on the outer edge of the speaker.
Nobody uses this technique because it is expensive to do.
I hope this helps you understand the process and I know my best student will have questions. Fire away smokester.
Sir Terrence |
| Sir TT, you completely lost me :) | Smokey
Apr 29, 2003 2:50 PM | | Hey TT, nice to hear from you.
First two paragraphs I understood, but I got lost during third paragraph starting with "Secondly cross vector cancellation is applied...", and definitely lost on the fourth paragraph also. Can you please explain it in simpler terms and how does all this make perception of sound coming from surround?
I read some where in your past posts that if some type of out of phase signal is fed simultaneously to front right and left channels, our perception is that sound coming from behind us...or something like that. As you can see, I am lost.
Thanks for clarification. |
| I don't know if I can make something so complex easy... | Sir Terrence the Terrible 1
Apr 29, 2003 4:08 PM | | Smoke,
This is such a complex issue, I don't know if I can simplify it. I'll try though.
Smoke when you listen to two speakers the sound from the left speaker reaches the left ear. However it also travels around the head and reaches the right ear delayed a couple of milliseconds. This is called crosstalk. You follow me? This crosstalk contracts the soundstage, and prevents the L-R information(out of phase component)in each channel from being heard as recorded. Are you following me so far?
To help eliminate crosstalk, a anti phase signal from the opposing channel has to be implemented. So, if I put the left channel information in the left channel, I am also going to put a crosstalk cancelling signal in the right channel as well. So as the left channels information travels around the head, it will meet the crosstalk cancelling signal in the right channel which cancels out that crosstalk. Prologic and prologic II both use this process to increase apparent signal seperation between channels. This process allows all recorded information including the L-R component(signals that are ambient in nature and are directed to the rear channels)to be heard clearly from both speakers. This widens the soundstage significantly, and with phase and frequency correct speakers can place images anywhere in a 360 degree soundfield when mixed(or encoded) as such.
As you know vocals coming from a "hard" center sound very different than phantom images suspended between two speakers. When It comes to the surround it is even more different because of reflections from the pinna. The frequencies between 2-4khz will be dulled because of HRT cancellation. To deal with that, eq is applied to correct the cancellations brought on by head and ear reflections. If you don't do this as sounds travel across the soundstage they will sound significantly different as it passes through the center, and the rear. Eq is to correct this and smoothen out channel transitions.
And your quoted me correctly. If you place any L-R(or out of phase)component in the Left and right channels, sit right in between the speakers, it will sound like it is coming from behind you. This is partially how virtual surround works. But without cancelling crosstalk, none of this information will be heard.
Any questions? ooops stupid question!!!
Sir Terrence |
| I don't know if I can make something so complex easy... | Smokey
Apr 30, 2003 7:40 AM | | Beside your post, I did some research on this subject, and oh man, this thing can get complicated quick. But now I have better understanding of it.
As you said, the main thing was to get rid of cross talk [distortion] first as to be able to do any kind of surround sound projection and sound staging. I found this illustration that explain it visually:
..........<img src="http://music.northwestern.edu/classes/3D/images/17xtlk.gif
.
And here is how it eliminated via DSP algorithm:
......<img src="http://music.northwestern.edu/classes/3D/images/19xtlkCn.gif">
.
Any way, I have much better understanding of this subject and pretty fascinating stuff.
Thanks. |
| |
|
|
