directional speakers

@treepmeyer, did Theoretica give you any specific recommendations or general guidelines? Did they tell you why they recommend speakers that are "more rather than less directional"? My guess is that they want to minimize early lateral reflections, but I could be wrong.

@treepmeyer posted, quoting from Theoretica’s website:

"An ongoing investigation of speaker directivity at Princeton University’s 3D3A Lab, has shown that dipole speaker designs, electrostatic speakers, as well as speakers with horns and waveguides offer significant advantages in 3D imaging with BACCH® 3D Sound in highly reflective rooms..."

Very interesting. These are pretty much the same conclusions I arrived at in my own explorations of loudspeaker design and room interaction (for unprocessed two-channel stereo). I like some of the attributes of dipoles and some of the attributes of horn/waveguide type speakers.

Duke

It looks like there is a divergence of opinion about how wide the speaker’s radiation pattern should be.

Given that the BACCH-SP system is using the spatial information already on the recording to present an exceptionally wide soundstage, rather than relying on strong early same-side-wall reflections to do so, I agree with Theoretica’s recommendation of speakers that are "more rather than less directional".

I pursue spatial realism in my speaker designs by a different route than what Theoretica is using, and in the course of product development I conducted quite a few controlled blind listening tests. I find a trade-off relationship between soundstage width enhancement due to strong early same-side-wall reflections on the one hand, and image precision, soundstage depth, and clarity on the other. I prefer the latter package of attributes, hence my preference for relatively narrow-pattern speakers, BUT my approach DOES give up some soundstage width relative to wide-pattern speakers.

The BACCH-SP arguably offers "you can have it all, and better than before", and in the context of what their processing does, strong early same-side-wall reflections are counter-productive.

Duke

@treepmeyer, in my experience interaural crosstalk cancellation works; it significantly expands the soundstage in width and depth, and enhances the sense of being within the acoustic space on the recording.

I owned an electronic interaural crosstalk cancellator in the early 80’s and it worked but colored the sound, and Theoretica has addressed the coloration issue.

In the mid-80’s I made a passive interaural crosstalk cancellator using large foam wedges which worked but was impractical. Theoretica has addressed the practicality issue.

Polk Audio made, and I believe still makes, speakers that use a secondary array of drivers to provide interaural crosstalk cancellation. Ime their approach works too.

But Theoretica’s processing goes beyond interaural crosstalk cancellation; it also uses de-correlation, which ime further improves perception of the acoustic space already on the recording. And since the BACCH processing is extracting information from the recording itself, the sense of spaciousness won’t have the "sameness" from one recording to the next that can arise with synthesized spaciousness contributed by upmixing to delayed rear channels.

To the best of my knowledge, techniques for interaural crosstalk cancellation only work within a fairly small listening area, typically one person wide and maybe two people deep. So there is that limitation. My understanding is that some versions of the BACCH system do head-tracking and therefore work from a wide range of listening locations, but still only for one (or two) people at a time.

(I have no affiliation with Theoretica.)

Duke

@erik_squires, when the day comes that it's MY bad memory in the spotlight, I hope to remember your example and be half as gracious as you.  Mad respect.

Duke

Regarding the radiation pattern of ATC speakers which use that big 3" dome midrange on a shallow waveguide -

As a general ballpark rule of thumb, a horn or waveguide starts to lose directional control below the frequency where its dimension in a given direction is less than 1/2 wavelength, though it still has some directional control down to the 1/4 wavelength frequency.

It looks to me like the round waveguide is about 3" in radius (6" in diameter) and maybe 1.75" deep. The 1.75" depth is the limiting factor. Assuming these numbers are correct, theoretically the waveguide will start losing directional control around 3.8 kHz, and it won’t have much effect below about 1.9 kHz. The directional control won’t fall off a cliff because the radius of the wageguide is still large enough to impart some control, but it’s being largely short-circuited by the shallowness of the waveguide.

You can see the off-axis response out to 40 degrees in the measurements made by Troels Graveson, scroll down about 1/3 of the way:

http://www.troelsgravesen.dk/ATC-SM75-150.htm

It looks to me like useful pattern control is maintained down to about 1.6 kHz, so a little bit lower than my calculation.

In the SCM40 the upper crossover point is 3.5 kHz, so the shallow waveguide is narrowing the radiation pattern from there down to about 1.6 kHz, a range of a little over an octave.

I would expect the direct radiator dome tweeter that takes over north of the 3" midrange to have a wide radiation pattern in the crossover region, narrowing of course as we go up in frequency and the wavelengths become shorter. So even if the midrange dome’s waveuide was large enough to exert serious directional control across its range, the tweeter would still have a wide pattern.

I used to be a dealer for ATC, and remained one up until the time I started manufacturing loudspeakers. I think very highly of ATC, and in particular I think very highly of that dome midrange, but "directional" is not a term I would have used to describe ATC speakers.

Duke

dealer/manufacturer