I see. Is the principle not the same? For example, here I'm showing a high passed and inverted, correlated signal along with the resultant partially cancelled response for three cases.. same level, +3dB and -3dB. The level variations also show frequency based variations for the same reason as above...
You didn't factor in the delay here... plus there would be another roll off on the top end of the inverted signal.
Ideally I would have made the cancellation signal a flat phase corrected (though DSP) solution. Such a signal could
quickly be made with RePhase to test in a dry run like this. I've used 270 uS as well as 540 uS of delay in my setup...
Ideally I would have made the cancellation signal a flat phase corrected (though DSP) solution. Such a signal could
quickly be made with RePhase to test in a dry run like this. I've used 270 uS as well as 540 uS of delay in my setup...
Last edited:
So the signals aren't in time with each other at the ear? (Only speaking of the signal components in question)
No, the inverted opposite channel signal is added with delay.
Look at the graph I posted of the measurement at the ear position (in this case with a dummy head in place).
We're trying to correct the leak of the right channel to the left ear and vice versa.
At least, that's one way to use cross talk. One could also do it like Linkwitz once did with up close speakers placed near to the head
and get away with re-enforcing the left and right signal that way or canceling the opposite channel. (see Patrick Bateman's text)
He called it the Watson project I believe...
Look at the graph I posted of the measurement at the ear position (in this case with a dummy head in place).
We're trying to correct the leak of the right channel to the left ear and vice versa.
At least, that's one way to use cross talk. One could also do it like Linkwitz once did with up close speakers placed near to the head
and get away with re-enforcing the left and right signal that way or canceling the opposite channel. (see Patrick Bateman's text)
He called it the Watson project I believe...
By having the right channel at the left ear partially cancelled by a signal added at the left channel, to reach the left ear together in time but inverted?
Yes, added beside the left channel (extra tweeter) to be able to adjust level and delay. (quite a bit like Polk SDA)
I've done it within the channels themselves (DSP manipulation) but never really kept it for music playing.
It wasn't working out in longer listening sessions for me. Instead I use simpler mid/side EQ which has similar benefits.
Technically, there's no end to it, as the extra added channel will leak to the opposite ear as well. 😉 That's how those RACE
algorithm's work within https://www.ambiophonics.org/ but they change the setup to better cater to this (speakers placed close together)
One of the supposed ultimate solutions is Prof. Choueiri's Bacch. I haven't heard it so I can't comment on that.
I've done it within the channels themselves (DSP manipulation) but never really kept it for music playing.
It wasn't working out in longer listening sessions for me. Instead I use simpler mid/side EQ which has similar benefits.
Technically, there's no end to it, as the extra added channel will leak to the opposite ear as well. 😉 That's how those RACE
algorithm's work within https://www.ambiophonics.org/ but they change the setup to better cater to this (speakers placed close together)
One of the supposed ultimate solutions is Prof. Choueiri's Bacch. I haven't heard it so I can't comment on that.
Last edited:
Hey Ronald, I looked that up here https://www.izotope.com/en/learn/what-is-midside-processing.html
is that the same as the matrix that jaddie mentioned?
is that the same as the matrix that jaddie mentioned?
Well, Jaddie used those basic principles to come up with the matrix components.
Mid/side EQ is often used in Studio work. The Mid component (which still has some Side information!) is Left + Right channel.
The Side component is Left minus Right channel (the opposite of that is Right minus Left, which is the same as the Left minus Right inverted).
You'd have to be careful with mid/side EQ and manipulation, as it will change the channel separation negatively.
As mentioned in that Phantom Center thread, one can achieve a more equal tonal balance across the left phantom right stage with some mid/side EQ.
Boosting the side channels just a little bit, also works as a cross-talk cancellation effect for the side channels, making them appear a bit wider.
-- back to the garage to cut more Scan Speak 10F frames, sorry, prioritizing 😀 --
Mid/side EQ is often used in Studio work. The Mid component (which still has some Side information!) is Left + Right channel.
The Side component is Left minus Right channel (the opposite of that is Right minus Left, which is the same as the Left minus Right inverted).
You'd have to be careful with mid/side EQ and manipulation, as it will change the channel separation negatively.
As mentioned in that Phantom Center thread, one can achieve a more equal tonal balance across the left phantom right stage with some mid/side EQ.
Boosting the side channels just a little bit, also works as a cross-talk cancellation effect for the side channels, making them appear a bit wider.
-- back to the garage to cut more Scan Speak 10F frames, sorry, prioritizing 😀 --
jaddie ... to be clear, those remarks in my last post were Floyd Tooles' not mine. That linked thread to AVS may be of interest to others as well as the OP'er for more of Tooles' remarks on this very subject.
What he terms "flat response" means to me that response doesn't deviate greatly from a flat drawn line .. a line not necessarily equal to SPL left to right. As for coloration, that's what separates good loudspeakers from bad. Loudspeakers without resonances/coloration don't draw attention to themselves and produce a more realistic sound stage. Those first reflections are important to the overall stereo illusion and the flatter loudspeakers response is, on/off axis, the better. That way less/no coloration is produced at our ears. Again, flat as in no excessive peaks/nulls in the frequency response.
What he terms "flat response" means to me that response doesn't deviate greatly from a flat drawn line .. a line not necessarily equal to SPL left to right. As for coloration, that's what separates good loudspeakers from bad. Loudspeakers without resonances/coloration don't draw attention to themselves and produce a more realistic sound stage. Those first reflections are important to the overall stereo illusion and the flatter loudspeakers response is, on/off axis, the better. That way less/no coloration is produced at our ears. Again, flat as in no excessive peaks/nulls in the frequency response.
Last edited:
Staying out of the theorising/semantic debates, as gripping as they are, I'd recommended trying something like a pair of Allison Acoustics CD-6's if you can find a pair at a good price with good woofer surrounds, or are willing to replace them.
but I like vintage stuff
but I like vintage stuff
jaddie, thanks for the writeup on interaural cancellation.
It seems the interaural cancellation route has to make for an even smaller sweet spot than normal stereo... no?
That would be deal breaker for me personally.
I'm thinking i want to work on a 3-speaker LCR setup, with the goal of both a more solid image and a wider listening zone. Hopefully not mutually exclusive goals. Ala Gerzon or something similar...
(i've played with multichannel extensively in the past (SACD and DVD-A) in a large studio type room carefully setup ....no desire to go past LCR again...heck, i often prefer single side or dual side mono, on less than well done stereo tracks....)
It seems the interaural cancellation route has to make for an even smaller sweet spot than normal stereo... no?
That would be deal breaker for me personally.
I'm thinking i want to work on a 3-speaker LCR setup, with the goal of both a more solid image and a wider listening zone. Hopefully not mutually exclusive goals. Ala Gerzon or something similar...
(i've played with multichannel extensively in the past (SACD and DVD-A) in a large studio type room carefully setup ....no desire to go past LCR again...heck, i often prefer single side or dual side mono, on less than well done stereo tracks....)
No. Mid/side is essentially using two signals, a L+R sum and a L-R difference, to re-process. The matrix used for developing the cancellation signal doesn't use the mid, a L+R sum.
...and attenuated and filtered. It's not a prefect process. Cancellation only occurs when the two signals have exact phase and amplitude. Phase is dependent on time, and there's a time difference between the left and right ear. Amplitude is affected by HRTF. So the cancellation signal that arrivces at the left ear that attenuates the right siganl won't be correct to have the same effect at the right ear.
It depends on a few factors. To be technically correct, no, the sweet spot for stereo has no lateral dimension, it's very small. That becomes more noticeable with crosstalk correction that uses inverted/delayed/filtered signals from the opposite chnnel. The matrix method is somewhat less critical because there is no processing of center-panned sources.
Preferences aside for the moment, the problem with LCR has always been deriving a good center from two channel stereo. There are many good active processors today that do this, and they're in AVRs. They use active non-linear processing to get a hard center signal. A passive, or at least linear L+R tends to narrow the stage while providing a solid center, which isn't always desirable.
Subjectively I’ve found that placing the mains close on each side of the 60” plasma just fwd and towed in slightly (between 5-10 deg) really supports a phantom center well…..not just for video but plain music also, even at loud volumes. Which in turn opens up the entire soundstage.
Of course your tv must be centered eye/ear level along with tweeters.
Of course your tv must be centered eye/ear level along with tweeters.
Thx. Pragmatically, that says to me, that an already too small spot (regular stereo), just gets smaller.
Yep, i use open architecture processors that allow roll-your-own processing designs for just about any task conceivable (or inconceivable !) 🙂
Currently Implementing this from Gerzon's paper http://decoy.iki.fi/dsound/ambisoni...for Multispeaker Stereo (TRIFIELD)_Gerzon.pdf.....
Need to hurry up and finish 3rd speaker for center, to start testing it and listening.
I only have one Michael Gerzon story, I only met him once at an AES in LA. We sat across the table from each other at lunch with a mutual friend. He went on endlessly about Ambisonics and it's virtues in capturing and reproducing a 3D sound field. He talked about a demo he wanted to record, with the microphone in the convicted's point of view in a guillotine execution. I vividly recall his eyes lighting up as he rubbed his hands together with glee as he described how "you could hear the blade coming down and passing right through your own neck!" He was absolutely giddy.
That's it. That's all I've got.
That's it. That's all I've got.
OP
Stockholm syndrome
“Temporary psychological disorder that appears in the person who has been kidnapped and that consists of being understanding and benevolent with the behavior of the kidnappers and progressively identifying with their ideas, either during the kidnapping or after being released.”
Stockholm syndrome
“Temporary psychological disorder that appears in the person who has been kidnapped and that consists of being understanding and benevolent with the behavior of the kidnappers and progressively identifying with their ideas, either during the kidnapping or after being released.”