Binaural demo of ambiophonics. - YouTube
This is a binaural demo of ambiophonics.
Don't judge on sound quality. it's highly compressed (youtube) and I used a Kodak playtouch pocket recorder.
Notice the depth of field and how the center image is the same size as the R+L.
One guitar should be far left and one guitar should be just right of center.
Of course, this all depends on your own HRTF, I used my own head and ears for the recording.
Also, there are no treatments in this room, so it will sound a little wet to you, but perfectly natural to me because I'm used to my room signature.
Notice the airplane that comes in at 1:00.... that's in the recording, and sounds like a real airplane overhead in my room.
Notice how the cereal sound pans from R to C at 1:35....The HF certainly gets crisper due to correct pinna cues. This is the fault of the prototype speaker. I placed the R+L drivers way to close, 10" apart. This is something that will be remedied in my current build.....The "Supernova" found under the full range forum.
This is a binaural demo of ambiophonics.
Don't judge on sound quality. it's highly compressed (youtube) and I used a Kodak playtouch pocket recorder.
Notice the depth of field and how the center image is the same size as the R+L.
One guitar should be far left and one guitar should be just right of center.
Of course, this all depends on your own HRTF, I used my own head and ears for the recording.
Also, there are no treatments in this room, so it will sound a little wet to you, but perfectly natural to me because I'm used to my room signature.
Notice the airplane that comes in at 1:00.... that's in the recording, and sounds like a real airplane overhead in my room.
Notice how the cereal sound pans from R to C at 1:35....The HF certainly gets crisper due to correct pinna cues. This is the fault of the prototype speaker. I placed the R+L drivers way to close, 10" apart. This is something that will be remedied in my current build.....The "Supernova" found under the full range forum.
Hi Barleywater
If by “phase coherence” you mean the acoustic phase as shown, it is the phase relationship between the input signal and the acoustic pressure the speaker produces. To get acoustic phase (per Heyser) you need to remove the apparent phase added due to the time of flight between the speaker and microphone. What is left when you remove all of the fixed delay is the acoustic phase of the source itself, that acoustic phase and the amplitude response govern how well a loudspeaker can reproduce the waveshape of a complex input signal, like a square wave or say music.
Fwiw, the SH-50 only reproduces a square wave from about 240Hz to 2900 Hz, it does so anywhere from very good looking on oscilloscope to fair depending. Over a broad band, the speaker radiates a portion of a sphere about 50 by 50 degrees.
There is very little change in appearance of the square wave until one is nearing the pattern edges, there radiate as if there had only one driver so there aren’t any lobes or nulls at crossover or anywhere else, just one forward lobe.
It does this because anywhere two or more drivers / sound sources interact; they are less than a quarter wavelength apart and so add coherently.
An odd ball crossover and the front to back spacing allows the elimination of the phase shift (time off set) the “named” crossover slopes impose resulting in one source in time as well as space, over a broad band..
This is the design approach essentially all of our full range loudspeakers use, the SH-50 and these being one of our oldest products.
Here are more based on the same acoustic approach;
Danley Loudspeakers | Danley Sounds Labs | Danley Sound Labs, Inc.
http://www.facebook.com/DanleySoundLabs?ref=ts
While a standard, perhaps what matters more than the response at one meter is the mag and time response at the listening position where you are when you experience the loudspeaker.
Best,
Tom Danley
If by “phase coherence” you mean the acoustic phase as shown, it is the phase relationship between the input signal and the acoustic pressure the speaker produces. To get acoustic phase (per Heyser) you need to remove the apparent phase added due to the time of flight between the speaker and microphone. What is left when you remove all of the fixed delay is the acoustic phase of the source itself, that acoustic phase and the amplitude response govern how well a loudspeaker can reproduce the waveshape of a complex input signal, like a square wave or say music.
Fwiw, the SH-50 only reproduces a square wave from about 240Hz to 2900 Hz, it does so anywhere from very good looking on oscilloscope to fair depending. Over a broad band, the speaker radiates a portion of a sphere about 50 by 50 degrees.
There is very little change in appearance of the square wave until one is nearing the pattern edges, there radiate as if there had only one driver so there aren’t any lobes or nulls at crossover or anywhere else, just one forward lobe.
It does this because anywhere two or more drivers / sound sources interact; they are less than a quarter wavelength apart and so add coherently.
An odd ball crossover and the front to back spacing allows the elimination of the phase shift (time off set) the “named” crossover slopes impose resulting in one source in time as well as space, over a broad band..
This is the design approach essentially all of our full range loudspeakers use, the SH-50 and these being one of our oldest products.
Here are more based on the same acoustic approach;
Danley Loudspeakers | Danley Sounds Labs | Danley Sound Labs, Inc.
http://www.facebook.com/DanleySoundLabs?ref=ts
While a standard, perhaps what matters more than the response at one meter is the mag and time response at the listening position where you are when you experience the loudspeaker.
Best,
Tom Danley
Yes, it is interesting.
..The minor differences could be down to reflections, though again, when I tested them the second time it was well away from walls.
As for the L channel discrepancy vs. the R - that could be a problem with my sound driver (or something else).
Anyway - it's was enjoyable and I'm glad Tom took the time to try it out and comment.
Tom.wav is clearly more focussed than 1 or 2.5 - with no real difference between 1 and 2.5 for me. I can't comment on depth, since 1 and 2.5 sound a bit duller than Tom.wav. That could suggest more distance.
One personal peculiarity to speak of: Tom.wav travels at (somewhat) equal angles from left to right. But in 1 and 2.5 the central "tom" and the next two "toms" to the right form a sort of cluster with reduced angles between them and a large angle to the third "tom" from center.
My right ear canal is kind of "un-normal", since no standard earplugs or earbuds keep hold in it. I reckon that 1 and 2.5 lack some information, that stabilizes the angular separation in tom.wav - at least for me.
View attachment 314878
My right ear canal is kind of "un-normal", since no standard earplugs or earbuds keep hold in it. I reckon that 1 and 2.5 lack some information, that stabilizes the angular separation in tom.wav - at least for me.
..makes me think I need to have my hearing checked
- it's always a possibility that something is wrong there. By phase coherence I am referring to phase and magnitude of frequency components of signal being maintained in output. Square wave is convenient example even with band limiting imposed by sampling rate.
The direct sound waveform received at listener position will correlate highly with output of speaker along design axis for smoothly produced wavefront. This is basis for how mind sees rest of room with ears.
Picture is worth a thousand words; frequency response, phase plot, and knowledge of minimum phase all pass at 300Hz and 1200Hz precludes stable square wave across most of operating band.
Attached is zip of swept square wave from about 100Hz to 3kHz captured on design axis of DSP implemented Pluto type speaker over 10 seconds. File is intended to look at with editor in lieu of numerous pics.
Diffraction as off axis frequency/angle dependent radiation leads to splotchy illumination consistent across source material, leading to identifiable sameness imparted to listening experience of system/space.
Similarly, baffle edges, sound reflecting off magnets and baskets, along with insides of cabinets that comes back through driver(s) impact resolution of system and identity as projector.
The panning experiments with modified bandwidth and uncluttered sound field reveal classic difficulties in achieving repeatable reference monitoring regarding sound image.
Regards,
Andrew
It's all nice, BUT, what will happen to the harmonic phase coherence when you put 2 loudspeakers in a triangle and get terrible stereophonic comb filtering at the listening position ?
There should be at least one possible solution to reproduce sound over loudspeakers while the above requirements remain achievable: Cross talk cancelling
Hi Barleywater
I am not sure you looked at the phase part of the curve, there is no “all pass” crossover phase shift like Butterworth, linkwitz riley and other named slopes produce.
The high pass hf crossover alone for example has a 4th order ultimate slope which normally would put 360 degrees phase rotation centered around 1200Hz which is not present. The phase response is as if it were one driver, not 7 with a 3 way passive crossover.
Here is an old link to some square waves if curious;
http://www.diyaudio.com/forums/multi-way/71824-making-square-waves.html
DSP correction has been commonplace in commercial sound for a long time and for one single location, it can perform miracles and can appear to be “perfect acoustic bondo”, but does not fix root or acoustic source problem.
The problem is that if one has separate sources producing the same signal (like a multi-way speaker at crossover), one has separate path lengths to the ear.
DSP can easily fix time and can correct a multi-way loudspeaker but the larger the spatial problems it’s is accounting for (the larger the acoustic spacing relative to a quarter wl), the smaller the “corrected zone” is and usually the result is worse everywhere else.
A multi-way speaker that produces an interference pattern (lobes and nulls) will still be producing an interference pattern with DSP correction because it can’t fix problems in horizontal or vertical placement unilaterally, only locally for one point in space.
Think on the large scale like concert sound, it usually sounds good or fair at the mix booth and different everywhere else.
The SH-50 (Synergy horns) are unusual in that they can be very largely phase correct without DSP and by radiating a segment of a sphere over a wide band act / measure /sound like as if it had one crossover-less driver and so it presents a very uniform spectrum across the listening area.
While the coherent addition of drivers makes it an ideal thing to unilaterally correct the remaining time/phase with DSP (and that works great!), the market demand is much more for a passive speaker “if” it sounds good without being active.
I guess I would say the goal here was to fundamentally fix the acoustic source radiation and you fix many problems like lobes and nulls etc all at once.
The speakers I mentioned are much more powerful than hifi speakers and are normally used say 30-100 feet from the audience or larger distances. As you probably have observed, most loudspeaker problems become larger, faster than the signal you want so if your audience is 19 times farther away, you need 100X the acoustic power to reach the same level.
The audience plane is usually as wide or wider than the coverage of a single box and so, the object is to have the sound field be as identical as possible everywhere within the pattern.
Also, confining as much of the total acoustic power to the design radiation angle is a very important thing, the larger the room, the less natural absorption there is and the more of a limiting factor reflected and reverberant sound is so far as understanding words so a constant/ high directivity design was also needed. In a living room, the directivity means the mag/phase response and impulse response at the listening position is much more like the 1 meter measurement. When you can confine most all of the energy into the front pattern, the direct field and intelligibility zone is maximized in large spaces..
If I recall the picture of your speakers, they would be nearly / mostly onmi directional up to the higher frequencies and would radiate from nearly a single point in space with a simple pattern.
If they only had a tweeter that was smaller still but covered the entire frequency range, then the radiation would for the most part be a simple sphere and then portion of a sphere, largely a single point in time and space.. From that (the shape of the point source radiation envelope) it would seem that your point source speakers are actually similar to the Synergy horns which is curious, those being more or less the whole pie, while the Synergy horn is as much as possible just a slice of pie..
While the company is so far not interested in hifi, I am and that is what drives the designs, why I mentioned it here..
The Synergy horn radiates like that except the angle is bounded by a solid wall (tangent pressure boundary forms an acoustic mirror image of the rest of the sphere) with a confined angle less than a flat baffle.
What radiates begins at highest frequencies at the apex at a dimension too small for the sound to have any directivity and so the angle is set by the horn wall.
AS it progresses out of the driver into the horn and at a dimension suitable for lower frequencies, the mid energy is added in synchrony with the highs and then further down, again according to the acoustic dimensions, the low frequency energy is added.
That time /phase offset the crossover normally causes is offset by the drivers inverse physical locations so the result is what is in time and space like one driver and not a 3 way system.
What radiates is a segment of a spherical radiation, not the entire sphere and as if it were a single (but impossible) horn driver.
The “spherical-ness” is audible too when you place two of them side by side hard packed, you cannot hear a seam where one box ends and the other begins, only possible by radiating as if it were a point at the very back of the cabinet (in fact from a point where the horn walls would intersect).
Best,
Tom
I am not sure you looked at the phase part of the curve, there is no “all pass” crossover phase shift like Butterworth, linkwitz riley and other named slopes produce.
The high pass hf crossover alone for example has a 4th order ultimate slope which normally would put 360 degrees phase rotation centered around 1200Hz which is not present. The phase response is as if it were one driver, not 7 with a 3 way passive crossover.
Here is an old link to some square waves if curious;
http://www.diyaudio.com/forums/multi-way/71824-making-square-waves.html
DSP correction has been commonplace in commercial sound for a long time and for one single location, it can perform miracles and can appear to be “perfect acoustic bondo”, but does not fix root or acoustic source problem.
The problem is that if one has separate sources producing the same signal (like a multi-way speaker at crossover), one has separate path lengths to the ear.
DSP can easily fix time and can correct a multi-way loudspeaker but the larger the spatial problems it’s is accounting for (the larger the acoustic spacing relative to a quarter wl), the smaller the “corrected zone” is and usually the result is worse everywhere else.
A multi-way speaker that produces an interference pattern (lobes and nulls) will still be producing an interference pattern with DSP correction because it can’t fix problems in horizontal or vertical placement unilaterally, only locally for one point in space.
Think on the large scale like concert sound, it usually sounds good or fair at the mix booth and different everywhere else.
The SH-50 (Synergy horns) are unusual in that they can be very largely phase correct without DSP and by radiating a segment of a sphere over a wide band act / measure /sound like as if it had one crossover-less driver and so it presents a very uniform spectrum across the listening area.
While the coherent addition of drivers makes it an ideal thing to unilaterally correct the remaining time/phase with DSP (and that works great!), the market demand is much more for a passive speaker “if” it sounds good without being active.
I guess I would say the goal here was to fundamentally fix the acoustic source radiation and you fix many problems like lobes and nulls etc all at once.
The speakers I mentioned are much more powerful than hifi speakers and are normally used say 30-100 feet from the audience or larger distances. As you probably have observed, most loudspeaker problems become larger, faster than the signal you want so if your audience is 19 times farther away, you need 100X the acoustic power to reach the same level.
The audience plane is usually as wide or wider than the coverage of a single box and so, the object is to have the sound field be as identical as possible everywhere within the pattern.
Also, confining as much of the total acoustic power to the design radiation angle is a very important thing, the larger the room, the less natural absorption there is and the more of a limiting factor reflected and reverberant sound is so far as understanding words so a constant/ high directivity design was also needed. In a living room, the directivity means the mag/phase response and impulse response at the listening position is much more like the 1 meter measurement. When you can confine most all of the energy into the front pattern, the direct field and intelligibility zone is maximized in large spaces..
If I recall the picture of your speakers, they would be nearly / mostly onmi directional up to the higher frequencies and would radiate from nearly a single point in space with a simple pattern.
If they only had a tweeter that was smaller still but covered the entire frequency range, then the radiation would for the most part be a simple sphere and then portion of a sphere, largely a single point in time and space.. From that (the shape of the point source radiation envelope) it would seem that your point source speakers are actually similar to the Synergy horns which is curious, those being more or less the whole pie, while the Synergy horn is as much as possible just a slice of pie..
While the company is so far not interested in hifi, I am and that is what drives the designs, why I mentioned it here..
The Synergy horn radiates like that except the angle is bounded by a solid wall (tangent pressure boundary forms an acoustic mirror image of the rest of the sphere) with a confined angle less than a flat baffle.
What radiates begins at highest frequencies at the apex at a dimension too small for the sound to have any directivity and so the angle is set by the horn wall.
AS it progresses out of the driver into the horn and at a dimension suitable for lower frequencies, the mid energy is added in synchrony with the highs and then further down, again according to the acoustic dimensions, the low frequency energy is added.
That time /phase offset the crossover normally causes is offset by the drivers inverse physical locations so the result is what is in time and space like one driver and not a 3 way system.
What radiates is a segment of a spherical radiation, not the entire sphere and as if it were a single (but impossible) horn driver.
The “spherical-ness” is audible too when you place two of them side by side hard packed, you cannot hear a seam where one box ends and the other begins, only possible by radiating as if it were a point at the very back of the cabinet (in fact from a point where the horn walls would intersect).
Best,
Tom
talking of crosstalk, I had last night my first go ever at ambiophonics. Olasonic speakers, VST pluggin on foobar. WOW!!!!
Wow! It's really starting to catch on now!
Hahaha!
Hello Tom
This problem is the same for "classic" IIR designs: good design will try to have phase coherency between two drivers troughout their crossover (say one octave around the crossover point) by targeting perfect symmetrical Linkwitz-Riley acoustical slopes (for both amplitude, and most importantly phase), by resorting to EQ or asymetrical electrical slopes. But of course the further apart the drivers and the worst it will get off axis (the coherency will be lost). In this regard your design is of course vastly superior!
Concerning DSP correction (convolution) I think there is two ways of doing it:
- the automated way (DRC) which if not done correctly (ie multiple measurement points and great care with what to correct and what to leave untouched) can lead to what you describe (sound get better for one location, but worse anywhere else). I think too aggressive and poorly made automated corrections (like measure, inverse, convolve...) have given this approach a bad reputation, but with care it can be quite efficient (cf DRC-FIR)
- the manual way, base on linear phase filters and/or phase correction of an existing (and already phase coherent) IIR system (including bass reflex phase shifts, etc.). If done correctly the coherence (between drivers) will be the same as an IIR system, but the phase will be linear throughout the whole spectrum. I have developed a software that tries to follow this approach (manual corrections and/or linear phase filtering):
http://www.diyaudio.com/forums/mult...hase-linearization-eq-fir-filtering-tool.html
What Tom describes as "a slice of the pie" has other benefits over the whole pie (omnidirectional ) approach for stereo image in a listening room. A controlled directivity allows for time-intensity trading to provide an image that is relatively immune to the position of the listener. You can'ttoe-in an omni speaker! (see Article about setting up controlled-directivity waveguide type speakers)
So, Tom, how are people to get the hifi advantages of the Synergies with the company doing only the pro market? The DIY approach is a difficult path on those.
So, Tom, how are people to get the hifi advantages of the Synergies with the company doing only the pro market? The DIY approach is a difficult path on those.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.