I really enjoy the sound of ambiophonics. If you use a spreadsheet to calculate your delays, and you're listening to a good recording, a miniambio can make the room seem to disappear.
There's a new array available in the new Cadillac, and while they haven't revealed how it works, I believe it's a variation on ambio.
If anyone wants to give this a try, here's my hypothesis on what they are doing, and why you might want to tinker with it.
If you are already familiar with crosstalk cancellation, skip the next two paragraphs.
The way that crosstalk cancellation works is that a signal is played that's out of phase with the opposite channel. The reason that we do this is that some sound from the left channel reaches your right ear, and vice versa. Crosstalk cancellation uses an inverted signal to cancel that out.
Subjectively, the soundstage gets wider. When everything is working REALLY well, it's almost like you're transported to where the original recording was made. Crosstalk cancellation is very very VERY room dependent, and this is the reason that a lot of speakers that do it are very directional. (Both Polk and the Ambiophonics institute used vertical arrays; Choueri(sp?) uses Gedlee speakers with waveguides.
In the Polk SDAs, the crosstalk cancellation happened at the source of the sound. In other words, the cancellations speakers are in the same enclosure as the "main" speakers. There is a speaker in the loudspeaker cabinet that plays the left signal, and there's a speaker in the cabinet that plays the right signal (but inverted.) And vice versa.
I think that the Cadillac is doing the same thing, but they've moved the cancellation speakers to where the listeners are. The reason that this is practical in 2015 is because amps and DSP are cheap. In order to do what they're doing in the Cadillac, you must take the signal and delay it.
I think that we could do the same thing at home. Yes, it would quite odd to have a headrest on your listening chair, but I think it would work.
With mini DSP, it is easy to do:
1) You have four channels of sound
2) The left and the right channel are plain ol' stereo
3) There is a second set of outputs. One channel is the left, but inverted. The second is the right, but inverted.
4) Because the speakers by your head are closer than the speakers far away, you have to delay the speakers that are far away. For instance, if the speakers on the far side of the room are two meters away, then you would delay them by 5.92 milliseconds.
Any questions?
There's a new array available in the new Cadillac, and while they haven't revealed how it works, I believe it's a variation on ambio.
If anyone wants to give this a try, here's my hypothesis on what they are doing, and why you might want to tinker with it.
If you are already familiar with crosstalk cancellation, skip the next two paragraphs.
The way that crosstalk cancellation works is that a signal is played that's out of phase with the opposite channel. The reason that we do this is that some sound from the left channel reaches your right ear, and vice versa. Crosstalk cancellation uses an inverted signal to cancel that out.
Subjectively, the soundstage gets wider. When everything is working REALLY well, it's almost like you're transported to where the original recording was made. Crosstalk cancellation is very very VERY room dependent, and this is the reason that a lot of speakers that do it are very directional. (Both Polk and the Ambiophonics institute used vertical arrays; Choueri(sp?) uses Gedlee speakers with waveguides.
An externally hosted image should be here but it was not working when we last tested it.
In the Polk SDAs, the crosstalk cancellation happened at the source of the sound. In other words, the cancellations speakers are in the same enclosure as the "main" speakers. There is a speaker in the loudspeaker cabinet that plays the left signal, and there's a speaker in the cabinet that plays the right signal (but inverted.) And vice versa.
I think that the Cadillac is doing the same thing, but they've moved the cancellation speakers to where the listeners are. The reason that this is practical in 2015 is because amps and DSP are cheap. In order to do what they're doing in the Cadillac, you must take the signal and delay it.
I think that we could do the same thing at home. Yes, it would quite odd to have a headrest on your listening chair, but I think it would work.
With mini DSP, it is easy to do:
1) You have four channels of sound
2) The left and the right channel are plain ol' stereo
3) There is a second set of outputs. One channel is the left, but inverted. The second is the right, but inverted.
4) Because the speakers by your head are closer than the speakers far away, you have to delay the speakers that are far away. For instance, if the speakers on the far side of the room are two meters away, then you would delay them by 5.92 milliseconds.
Any questions?
Hi,
No, you have to delay the speakers that are near.
Cancellation from source speakers is some jiggerypokery
that depends on HTFs (head transfer functions) and feeding
out of phase left from the right and vice versa.
There is no sound (sorry) basis to crosstalk cancellation.
It goes back to early Hafler implementatations and is
at best a one trick pony you quickly get tired of.
Its simply based on a wrong assumption, listen to headphones
for for no cross channel crosstalk, and the cross channel bleed
circuits for headphones that generally improve matters a lot.
rgds, sreten.
No, you have to delay the speakers that are near.
Cancellation from source speakers is some jiggerypokery
that depends on HTFs (head transfer functions) and feeding
out of phase left from the right and vice versa.
There is no sound (sorry) basis to crosstalk cancellation.
It goes back to early Hafler implementatations and is
at best a one trick pony you quickly get tired of.
Its simply based on a wrong assumption, listen to headphones
for for no cross channel crosstalk, and the cross channel bleed
circuits for headphones that generally improve matters a lot.
rgds, sreten.
Yes, you can delay the ones near your ears but, they would have to be cut at 1khz at least to avoid confusing pina cues.
Even then, I'm not sure it would work.
I often thought of this with SL's WATSON.
Even then, I'm not sure it would work.
I often thought of this with SL's WATSON.
By the way, ambiophonics is not something to be disregarded as sreten implies.
And its far from a headphone sound.
The only way to get frontal pina cues is by utilizing the smith realizer.
Or, of course with binaural recordings.
And its far from a headphone sound.
The only way to get frontal pina cues is by utilizing the smith realizer.
Or, of course with binaural recordings.
Holy ___, this works astoundingly well. Bose is really onto something here.
I'll post some results here soon. Here's a preview:
1) This images better than Danley SH-50s
2) This images better than a good miniambio setup
3) This images better than 95% of everything I've ever heard, without the weird 'swishiness' you get sometimes with ambio
I bet it sounds different! But when we think about all those reflections in a room I start to wonder what is actually cancelled and at which point in space... isn't it just a mess?
By the way my sister had a stereo FM/cassette portable by JVC in the early 80's, that had ambiophonics emulator! From stereo source it made weird sound from speakers, quite good with phones. Those days JVC and Sony produced some binaural/ambio recordings and they wanted to promote it. This is a part of history now.
StereoTimes -- Ambiophonics, 2nd Edition: Replacing Stereophonics to Achieve Concert-Hall Realism
StereoTimes -- Commentary: Ambiophonics, 2nd Edition
Modern ambiophonics http://www.ambiophonics.org/
By the way my sister had a stereo FM/cassette portable by JVC in the early 80's, that had ambiophonics emulator! From stereo source it made weird sound from speakers, quite good with phones. Those days JVC and Sony produced some binaural/ambio recordings and they wanted to promote it. This is a part of history now.
StereoTimes -- Ambiophonics, 2nd Edition: Replacing Stereophonics to Achieve Concert-Hall Realism
StereoTimes -- Commentary: Ambiophonics, 2nd Edition
Modern ambiophonics http://www.ambiophonics.org/
Last edited:
Actually, it turns out my hypothesis was wrong!
Bose isn't doing crosstalk cancellation with the headrest speakers; they're doing noise cancellation.
Nonetheless, putting crosstalk cancellation speakers near your head seems to work shockingly well. I never would've considered it if it wasn't for that Car and Driver article.
I've tried regular stereo, I've tried various flavors of crosstalk cancellation. This is working better than anything I've tried before. It basically sounds like you're wearing headphones, but the soundstage isn't "in your head", the soundstage is a semi-circle that's as deep as the loudspeakers.
Well this is all pretty interesting... so just to check what you are saying - you have used the DSP4YOU/ambio4you ambiophonics unit (2-channel or 4 channel?) or some other ambiophonics processor?
The reason I ask is that I am currently designing a large multi-driver Linkwitz-style dipole (with even-order-harmonic cancelling woofer positioning). I'm using twelve 12 inch drivers, and this makes the thing about 36in (90cm) wide and 40" (1m) tall even when the drivers are packed closely together. I'm thinking of placing an open-baffle panel on top of this that will contain the rest of the drivers, located on each end (at the left and right extremes of the panel).
According to what I have read about the DSP4YOU unit, I would not need to make the OB panel all that wide if I used the ambiophonics unit and I would still be able to get a nice stereo image or soundstage because of the processing. Does that sound correct?
Last edited:
I have an ambio4you but I am not using it.
I am using a plain ol' miniDSP. I am doing the same thing that Polk was doing, except with two differences:
1) the crosstalk cancellation speakers are about a foot from my head
2) the crosstalk cancellation speakers are delayed and attenuated, due to the fact they're closer
That's it!
I like the sound of ambio, and I've tried a lot of variations. The problem I had with ambio was that bad recordings sound totally "mono", and most recordings are bad. Also, ambio nukes the bass.
This setup sounds more enveloping than ambio. It sounds a LOT like wearing headphones. Except you're not.
Hi Patrick, interesting work!
1. What is the angle between the cancellation speakers, zero degrees would be both directly in front of you? (a diagram would help!)
2. What is the frequency response of the cancellation speakers, and relative SPL at the listening position?
Thanks,
Mike
1. What is the angle between the cancellation speakers, zero degrees would be both directly in front of you? (a diagram would help!)
2. What is the frequency response of the cancellation speakers, and relative SPL at the listening position?
Thanks,
Mike
Interesting. I would have thought that you should only be feeding the difference between left and right channels to the canceling speakers?? what if you have things that are equally left and right... if you send an out of phase version of the left channel to the right channel cancelation speaker but the right channel has the same sound then don't you lower the overall level of those sounds that are equal between left and right?
I can imagine singers becoming further back and softer, as they are often smack bang in the middle of the stereo image...
Tony.
I can imagine singers becoming further back and softer, as they are often smack bang in the middle of the stereo image...
Tony.
At the frequencies where crosstalk cancellation works, angle doesn't matter really. For instance, 500hz is 0.7 meters long. At these lengths the speaker could be pointed up, left, right... It doesn't really matter.
What *does* matter is getting the wavefront from the cancellation speaker in sync with the wavefront from the stereo speaker. A shift of even ten centimeters (4") will make a difference. This is because we're working to 'null out' the crosstalk, and in order to do that, we need to get the wavefronts to line up.
I got good results with a tape measure, but even better results using a mic.
Basically you just set up a mic, and measure the stereo speaker, then the crosstalk speaker. And then you delay the crosstalk speaker until they line up.
If I'm not mistaken, you need to do that with the crossover filters in place. For instance, I'm doing a second order lowpass at 750hz on the crosstalk speakers. That lowpass will introduce a delay, which will affect how much delay you use with the DSP.
If that last paragraph is confusing, keep in mind that you can get quite good results with a tape measure. (sound travels 34cm in one millisecond.)
As for the frequency response, I believe that the crosstalk cancellation speakers and the stereo speakers should have the same frequency response (and ideally the same phase) in the passband where both are operating. I am running the crosstalk cancellation speakers over a two octave bandwidth, from 188hz to 750hz. Frequencies above 1khz do not benefit from crosstalk cancellation, so I'm not doing any cancellation above 750hz. (Obviously, you can play around with the low pass and see what difference it makes.)
Last edited:
This sounds easy enough to try, my side surrounds are about 4 feet away on each side, though...do you think that's too far? I've got a spare MiniDSP, this seems really trivial to try.
Give it a try, but I'd recommend moving them closer. The further away they are, the more likely you're going to have play them loud. And the louder they play, the more they'll interfere with the mains.
Here's an example:
Let's say your main stereo speakers are two meters away and they're playing at 90dB. If your crosstalk cancellation speakers are one meter away, then they can play at 87dB. (Because they're half the distance.)
Move 'em 25 centimeters away, and you can drop the level to 81dB...
Basically we can use delay and attenuation to insure that the soundstage is anchored to the mains. I think that one problem with multichannel setups is that the additional channels become distracting. We can avoid that via delay and attenuation.
I might try today and see what happens. I'd move them closer, but they're wall/ceiling mounted and angle downwards, so I can't really move them. I might just toss up an extra set of spares and rew/target curve them to the mains.
I wonder how effective this would be with the cancellation speakers sitting to the sides of the mains?
I wonder how effective this would be with the cancellation speakers sitting to the sides of the mains?
All of this is subjective of course. But here was my experience:
When running conventional ambio, with crosstalk cancellation occurring at the same location as the "stereo" speakers, there was a noticeable improvement with good recordings.
For instance, I would occasionally hear a soundstage that was wider than the loudspeaker locations. And the center was always rock-solid, because the mains are close together.
For me, the problem is that 90% of the recordings out there are basically mono.
With a conventional stereo triangle, mono recordings still have width, because the loudspeakers are a few feet apart. So even if the recording is utter crap, there's still some "size" to the image.
With ambio, bad recordings sound exceptionally bad. And most recordings are bad.
With the setup described in this thread, even bad recordings sound fairly spacious, because the crosstalk cancellation speakers are quite a distance from the stereo speakers, about one or two meters away.
The easiest way to describe it is that it sounds like you're wearing headphones. But you're not. And the depth of the soundstage isn't limited to your head; it extends to the depth of the stereo speakers and extends to the left and to the right of you. (As you get with headphones.)
When running conventional ambio, with crosstalk cancellation occurring at the same location as the "stereo" speakers, there was a noticeable improvement with good recordings.
For instance, I would occasionally hear a soundstage that was wider than the loudspeaker locations. And the center was always rock-solid, because the mains are close together.
For me, the problem is that 90% of the recordings out there are basically mono.
With a conventional stereo triangle, mono recordings still have width, because the loudspeakers are a few feet apart. So even if the recording is utter crap, there's still some "size" to the image.
With ambio, bad recordings sound exceptionally bad. And most recordings are bad.
With the setup described in this thread, even bad recordings sound fairly spacious, because the crosstalk cancellation speakers are quite a distance from the stereo speakers, about one or two meters away.
The easiest way to describe it is that it sounds like you're wearing headphones. But you're not. And the depth of the soundstage isn't limited to your head; it extends to the depth of the stereo speakers and extends to the left and to the right of you. (As you get with headphones.)
Aha. It seems as though a weekend murdering amount of experimentation is in order!
Edit: I've already got a recording picked out to try this with as well, there's a Ry Cooder with Mohan Vishwa Bhatt (sp?) called "Meeting by the River" which used a very simple blumlein microphone setup...sounds damn good, even if the music isn't really most people's cup of tea.
Edit: I've already got a recording picked out to try this with as well, there's a Ry Cooder with Mohan Vishwa Bhatt (sp?) called "Meeting by the River" which used a very simple blumlein microphone setup...sounds damn good, even if the music isn't really most people's cup of tea.
Last edited:
Here's what I'm using for the stereo speakers:
Here's what I'm using for the cancellation speakers:
For the first round, where I did the calculations with a tape measure, it took all of two hours to get everything going. And most of that was just running the wires
If anyone wants to try this, but doesn't have four satellites laying around, just take a look on Craigslist. People practically give away computer speakers and home theater in a box setups. I've measured the frequency and phase on these satellites and they're excellent. It's amazing how flat you can get the frequency response and the phase when you're using 2" and 3" drivers.
An externally hosted image should be here but it was not working when we last tested it.
Here's what I'm using for the cancellation speakers:
For the first round, where I did the calculations with a tape measure, it took all of two hours to get everything going. And most of that was just running the wires
If anyone wants to try this, but doesn't have four satellites laying around, just take a look on Craigslist. People practically give away computer speakers and home theater in a box setups. I've measured the frequency and phase on these satellites and they're excellent. It's amazing how flat you can get the frequency response and the phase when you're using 2" and 3" drivers.
- Status
- Not open for further replies.