Interesting, VERY different speaker system I helped to design.

[Harvey]

Some of you may know that I'm a musician and an electronic design engineer, mostly pro audio, and mostly a keeper of the black art of analog circuit design. Recently, I had the pleasure of being part of the team that designed and built a unique piece of ear candy called the Space Station, and it's a truly amazing box because it creates a huge stereo sound field from a single enclosure.

Note - I'm not spamming. I designed the analog signal path circuitry from the input jacks to the input of the power amp section, but I don't make anything from sales of the product. I just wanted to share this piece of my work with my fellow members. I think it's about as cool as any audio product I've seen in a long time.

Here's a pic:

It's intended for single musicians playing instruments with stereo voices and it's an interesting way to get a good PA blend on vocals, either with or without the instruments in the mix. One of the coolest parts about it is, there's no "sweet spot" between left and right speakers. The effect works all over the room. As a keyboard player, I can hear the same balance on my instrument as the rest of the band and the audience.

Link to some cool videos deomonstrating the effect. They were mixed in stereo, what you're hearing is the sterio playback through the Space Station, and recorded using an M-S (Mid - Side) microphone.

This is one of my favorites. If you want to get the full effect, try listening in headphones.

Note - The first few minutes of every demo is the owner, Aspen Pittman, talking about it. Aspen also founded Groove Tubes, who sold pre-tested, matched sets of vacuum tubes for guitar amps, etc.

Once you heard his spiel, you can skip the same few minutes on the rest of the tracks, but they're well worth hearing. Hope you like this.

 

[DrPizza]

Nice!

 

[Perknose]

:thumbsup:

 

[darkswordsman17]

Yeah that's quite cool. I've been saying for years there's no reason we can't get good spatial representation using a single speakers if they're designed properly. And then you'd just put all the money into making the best speaker (technically or subjectively) that you can instead of trying to compensate with more and more speakers (of course with individual speakers being able to present great spatial audio, adding more should then just increase the output versus needing to worry about placement as much).

Now, what about the recording side? I've been wondering why audio hasn't moved on from channels to make recordings be agnostic of your setup, such that it has say spatial data so that then the device you listen to it on would then place sounds according to your sound setup. I guess you could argue that's what channels is doing, but with how much processing power there is now, it seems like it'd be worth it to make a next step. Or maybe I really don't know recording technicalities all that much?

 

[Harvey]

Yeah that's quite cool. I've been saying for years there's no reason we can't get good spatial representation using a single speakers if they're designed properly. And then you'd just put all the money into making the best speaker (technically or subjectively) that you can instead of trying to compensate with more and more speakers (of course with individual speakers being able to present great spatial audio, adding more should then just increase the output versus needing to worry about placement as much).

Now, what about the recording side? I've been wondering why audio hasn't moved on from channels to make recordings be agnostic of your setup, such that it has say spatial data so that then the device you listen to it on would then place sounds according to your sound setup. I guess you could argue that's what channels is doing, but with how much processing power there is now, it seems like it'd be worth it to make a next step. Or maybe I really don't know recording technicalities all that much?

Good thoughts, but it doesn't work quite that way. Starting with the recording side, recording a stereo image from a single point is called "M-S" or "Mid-Side" miking." It was devised by EMI engineer Alan Blumlein, an early pioneer of stereophonic and surround sound. Blumlein patented the technique in 1933 and used it on some of the earliest stereophonic recordings.

It uses a front facing direcctional (cardiod) mic, which records the mid signal (L + R) and a side facing figure 8 mic. In this configuration, the two capsules of the figure 8 mic are essentially out of phase so it records the differnce (L - R) between the left and right sighals.

M-S mic configuration​

In this pic, the cardiod mic is on top, and one side of the figure 8 mic is facing the camera. The front of the sound field is to the left, and it produces this sound field pattern:

Polar pattern of M-S mic setup​

To derive a left - right sterio output, the two signals are mixed through a summing circuit which produces:

(L + R) + (L - R) = 2L

The signals are also mixed through a differencing circuit which produces:

(L + R) - (L - R) = 2R

One advantage of this technique is that it always produces a good center channel information which is always in phase because it was recorded by a single microphone. That means, you don't get weird losses and cancellations when the stereo signal is mixed to mono.

Another advantage is that the recorded information contains a wealth of phase information that can actually reproduce height and depth images from two well placed stereo speakers. You'd have to hear an M-S recording of sound illusions in space, such as a grand piano, a small chamber group or a rotating Leslie speaker (think Hammond B3, etc.) to appreciate the illusions of depth and motion.

The Space Station uses the same idea in a speaker system and the same sum and differnce matrix to convert left and right stereo signals to M S.

L + R = Mid
L - R = Side

It uses a front firing speaker for the Mid signal and a side firing speaker for the Side signal. The enclosure is open on both sides so the Mid speaker reproduces the required difference output.

Unfortunately, the accuaracy of the stereo illusion is limited by physics. To recreate an accurate stereo illusion, the speakers would have to be closer together than 1/2 of the shortest wavelength (highest frequency) signal. This is easy with tiny mic capsules, but not so with physically large speakers.

The resulting illusion can never be an accurate reproduction of an original stereo sound field, but for musicians in a live playing environment, it creates an amazing, dynamic spacial illusion for musical instruments with stereo voices, such as grand piano, rotating Leslie speakers, etc. And, as I mentioned in my OP, another advantage is that this illusion works evenly throughout the listening area, rather than in a restricted "sweet spot" between left and right speakers.

We can talk ABOUT audio all day, but the bottom line is, you have to hear this effect to believe that a huge 3D sound field can be coming from a single box.

As an audio electronic design engineer and musician, it was a pleasure to participate in the design of the product. Just wanted to share this with my fellow audiots and musicians.

 

[jpishgar]

Way, way over my head. But still very much cool.
-JP

 

[Harvey]

Way, way over my head. But still very much cool.
-JP

Thanks. The illusion is actually IN your head, not OVER it.

Athur C. Clarke's (2001) A Space Odyssey) third law states, "Any sufficiently advanced technology is indistinguishable from magic."

I've been present watching skilled, knowledgeable musicians and recording professionals scratching their colllective butts and getting the "Holy shit!" effect when experiencing a huge stereo illusion coming from a single box.

You don't necessicarily have to understand the magic to get it, but actually listening to it works.

 

[darkswordsman17]

Holy carp I wrote a lot, hope it made some semblance of sense.

Interesting, so is that in a mono mix (that equipment can then process into individual sound fields for stereo), or is it actually recorded as a stereo mix right off the bat? I might not be getting what is going on there.

Of course now if you want a stereo or surround mix it's much more relatively simple, cheap, and easy to just setup mics (especially if you have a mix of equipment that would be suitable for different mics).

I know there's some binaural mics (I believe some company was even developing a 360 degree mic). Beck had a setup (and I believe you can buy it for yourself) that used a cube with 2 "ears" on each side (such that each corner would act like the front of a persons head as opposed to each side having separate but flat ears). I believe it used 8 channels and then they'd have the soundboard for mixing and output.

I was kinda thinking that maybe there would be ways to reduce the overall size of the recording if you could somehow put things down in a computational manner, versus having say however many mics pointed in varying directions (and not locked to any certain number of channels). And then your device would take your setup and adjust the factors to make it sound right. So for headphones it could adjust the timing and extreme separation, and then for any number of speakers adjust to fit spatial and timing data, which would enable higher quality than say expanding a stereo recording for surround, or downmixing for fewer than the channels the recording is. This way you have one single recording with all the data that wouldn't be like having a recording with a bunch of different channels, but you'd get better quality regardless if you were listening on 1 speaker or 10 (compared to saying having 2 or 6 channel recording where it'd have to be extrapolated up or down).

Or maybe some way of splitting certain data of the recording into computational output, like making spatial and timing data so like it could give instrument/sounds dimensional plot points while having the frequency response/harmonics/distortion/etc have separate data (hopefully at maximum level). You could use more mics (and possibly other equipment that aren't about offering max fidelity of audio signals, but maybe could offer other data) to gain more precise inputs, and then end up with a smaller file that computers could process easily (and not be limited to the channels it was encoded into).

Not sure if any of that makes sense, and is really likely just complexity without much real benefit (but likely significantly more computational needs). And unless the data was efficiently lossless could actually cause larger file sizes than just having a bunch of channels that an audio processor could then determine placement for your speaker setup. Plus it could cause coherency issues in comparison to having a bunch of channels to mix with (in many ways, what I was saying would be like in a recording studio where individual sound components, like each instrument, has its own channel, I guess I'm saying I'd love to have a format that offers up all the data, and then your equipment can process it to how you see fit). The fact that we can get so much information from relatively simple recordings is quite cool on its own.

The Space Station just effectively has like a simple mixer board then? I was going to ask how it worked (but realize smaller audio shops might not always like to share their details so other companies can rip them off and mass produce them in China for cheap), if it had just one speaker (and had like a baffle that could split the output somehow, like there's some speakers that put a cone shape in front of the woofer in order to disperse the sound outward giving it more dispersion) or more (so say something like the Bose 901, some front facing some pointed to reflect off walls), but watched one of the videos and it seemed like it was just one. But from what you say, I'm guessing the mid woofer is front facing, and then the tweeter is side facing? I see there's the grille over the front (that looks like the midwoofer), and then the opening on the side (but don't see a woofer there, so I'd assume the tweeter). Or are there multiple drivers just with different placements (so say 2-3 woofers, facing different directions and then a tweeter as well)?

Not to get too (much more) off tangent, but thinking about say planar speakers that have strong dipole output. You could make it so there's multiple "speakers" but all are around a single center beam that you could then turn each speaker to face wherever you want. With a planar speaker you could just have two and then have them turned perpendicular to each other (so one would face forward like the cardioid mic, and then the other would be like the figure 8 mic). On the bottom of the center beam (to serve as a sturdy base, pun intended) you could have a woofer for the low end (which I believe planars typically don't excel at lower frequency output).

Even on just the Space Station, would be curious about adding a passive radiator for more low end feel too without need to bulk up amp output for more and more speakers (which probably isn't well suited for some of that music, but maybe it could help some keyboards hit lower octaves and mimic pipe organ).

 

[Harvey]

Holy carp I wrote a lot, hope it made some semblance of sense.

Interesting, so is that in a mono mix (that equipment can then process into individual sound fields for stereo), or is it actually recorded as a stereo mix right off the bat? I might not be getting what is going on there.

This system does not convert a mono signal to stereo. Like left + right stereo recording, it uses two channels of amplification. See my previous post about "M-S" or "Mid-Side" miking" that records mid and side information and mixes these signals through a sum and difference circuits to derive left + right stereo.

This speaker system does the inverse. It starts with left + right stereo signals and mixes them through the same kind of sum and difference circuits to derive mid and side information. The speakers are arranged with one speaker facing forward (mid) and one speaker facing sidways and open on both sides (side).

I also described why this speaker system is not able to reproduce an accurate stereo image due to the physical size of speakers compared to microphone capsules. What it does is create is a very large, dynamic spatial illusion that works very well for musical instruments and any other kinds of signals where an exact replication of a recorded stereo image is not the objective.

The advantage of the Space Station is that it creates the illusion from a single box, and the illusion works across the entire listening area, not just a small "sweet spot" between left and right speakers, so everyone, including the musician playing the instrument, hears the same mix.

The bottom lins is, it's an interesting and useful illusion. If you lke the illusion, you get what we were after.

.The Space Station just effectively has like a simple mixer board then?

No, just the sum and difference matrix I descibed and a couple of controls for volume and the balance between the levels of the mid and side channels. Because the mid channel is tri-amped, it also has controls to set the balance between the woofer, mid and high frequency drivers.
.
Forgive me for not addressing the rest of your post because it goes into other subjects that could be interesting discussions, but they're not really related to what the Space Station does or how it works.

 

[MongGrel]

I'm a bit skeptical, but of course I've never been in a room using one.

Does seem interesting.

:thumbsup::thumbsup:

 

[Harvey]

I'm a bit skeptical, but of course I've never been in a room using one.

Does seem interesting.

:thumbsup::thumbsup:

You and almost everyone else who hasn't yet heard it in action. I've been in rooms full of experienced recording engineers, audio gurus and musicians walking around the thing and scratchiing their heads and other more socially awkward parts of their bodies, wondering how it could do what it does.

It doesn't violate any laws of physics in this universe, but it reminds me of Arthur C. Clarke's (2001: A Space Odyssey) third law - "Any sufficiently advanced technology is indistinguishable from magic." :sneaky:

 

[Psilosounds]

Harvey! I have been looking for something kind of close to what you helped design, but not have not seen anything that quite fits the bill. I have come up with a pretty approximate physical design, but I am very interested in some of your mid side processing knowledge. I had decided that was how to approach it before finding that box you worked on, that’s actually how I found it, and this thread. Please hit me up
If you are interesting in possibly working on another similar project.