Announcement

Collapse
No announcement yet.

Another Cube Cab

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Another Cube Cab

    Here's another version of the cube box recently completed. The 2 foot cube has a corner cut off at 45 degrees, the hypotenuse is 2 feet. A 2 foot square front panel of 16 five inch 8 ohm (the blue Sony "dynamic woofers" from the buy outs section) covers the corner, and a back panel of four 8" woofers (Studio Pro WH8) tucked behind. The space behind the front panel is less than the opening on the sides, so it has a bit more horn design to it, but still maintains the detuned design because the back panel is vented into the space between ther panels and the front panel is open back. Two piezo tweeters give it a good high end. Response 60 Hz to 27 kHz across the drivers. Sonys and piezos at $2.50 each, the WH8s at $12.50, this box cost $95 to populate. I put a bit more into the construction on this one, including using rubberized car undercoating paint, so the total cost was about $125.

    8 ohm network, 750 watts handling, tested at 500 watts with a Behringer iNuke 1000 bridged. Due to a previous comment about bedroom playing and pointing it at the wall, it occurred to me that this design might be very well suited for the testing site. So I turned it on its side with the face pointing up at 45 degrees and set it so it was aimed at the center of the movie screen, a slightly concave wall 40 feet high by 60 wide leaning forward about 10 degrees. The sound level behind the speaker was very even across the entire 454 parking space lot, and vocals were clearly audible a quarter mile away and 45 degrees off axis. Between the speaker and screen there was a lot of echo interference. Set upright and facing away from the screen, the sound level on axis and up to 45 degrees to either side was as consistent with the reflected sound. More than 45 degrees off axis there was a noticeable drop off across the frequency range, a bit more so in the high midrange, like lead guitar solos. At 90 degrees the sound level was about half that at the same distance on axis. This was as expected from the design, an open front backed by a reflector, to focus both front and back outputs of the drivers in a pattern somewhat like cardioid. In the intended configuration, two of these with one at each side of the presentation area, they can be turned to reinforce that focus in a listening area or turned more outwards for a wider coverage.

    A bit about the venue: The Moonlite Drive In, open since 1949, is one of only three drive in theaters in the National Registry of Historic Places. It is located along The Crooked Road, a collection of sites in southwest Virginia that present traditional country, bluegrass and mountain music. As such, in times past, musicians would come and play from the roof of the concession stand during intermission between the double features. Now over 60 years old, it's showing its age. Efforts are underway to get a grant to help with upkeep since for many years now its intake was only enough to keep it open. Also in the planning stage is to reassociate it with music, both during intermissions and as separate presentations. If you happen to be in the Bristol VA/TN area (Abingdon is 13 miles north of Bristol along I-81) on a weekend between April and October, stop by. Check the website at http://www.moonlitetheatre.com/ for details, including the possible addition of music.
    Attached Files

  • #2
    Re: Another Cube Cab

    I'm sorry, but nothing about that cab is of any value, a comb filtering nightmare is the only way I can describe it.
    www.billfitzmaurice.com
    www.billfitzmaurice.info/forum

    Comment


    • #3
      Re: Another Cube Cab

      Originally posted by billfitzmaurice View Post
      I'm sorry, but nothing about that cab is of any value, a comb filtering nightmare is the only way I can describe it.
      Well sir, I definitely respect your opinion, as I'm familiar with your experience. However I disagree with it, specifically because of experience. Your knowledge of theory can't be criticized, but theory is not the same as experience, in this case direct experience. I'm a professional scientist and make extensive use of theory, up to the point of a specific hypothesis. After that comes testing, preferably with measurement, collection of pertinent data, and analysis of the data in order to support or fail to support the hypothesis.

      At this time I have one data point, that being my own subjective experience. My experience was that it sounded pretty darn good. With another's opinion, it'd be less subjective, but by no means objective. If you can describe for me how I might test for such effects, I'd be glad to have your assistance. I have access to a great deal of signal analysis equipment and could very well do the testing. Still, there's no direct experience on your end. Along those lines, if you could suggest a design or a specific loudspeaker for comparison, I'd be glad to do some A/B testing and make a recording of it. Yes, the recording will alter things some, but the same recording of the two side by side should at least give a sonic comparison to experience. I can record digital at 96000 Hz sampling and provide a 32 bit WAV file, which should make for a fairly realistic reproduction. Or if anyone interested is in my area, I'd be glad to provide a demonstration and let them provide an opinion if not their own measurements. The sound reinforcement pro at the local store hasn't heard them but was impressed with the build and wants to try it out, and I can tell you what he thinks.

      Comment


      • #4
        Re: Another Cube Cab

        Originally posted by drmcclainphd View Post
        After that comes testing, preferably with measurement, collection of pertinent data, and analysis of the data in order to support or fail to support the hypothesis.
        Show us polar plots.
        www.billfitzmaurice.com
        www.billfitzmaurice.info/forum

        Comment


        • #5
          Re: Another Cube Cab

          The original concept was called "Sweet 16" in the January 1961 Popular Electronics magazine I read as a kid.

          "Not a Speaker Designer - Not even on the Internet"
          “Pride is your greatest enemy, humility is your greatest friend.”
          "If the freedom of speech is taken away, then dumb and silent we may be led, like sheep to the slaughter."

          Comment


          • #6
            Re: Another Cube Cab

            Originally posted by billfitzmaurice View Post
            Show us polar plots.
            I most certainly can. But of what? A polar in its simplest respect is a full band out put volume measure at enough variations from axis to give a good curve of apparent output around it as center. That'd measure the wide band dispersion but wouldn't speak to the hypothetical "comb filter" problem.

            That problem would be exacerbated by the 2D output of a set on 4 x 4 drivers, and should be measured in 3D if it's to be meaningful, as failure to show it in 2D makes the 3D measurement trivial. To ignore this would be to accept theoretical point source solutions as adequate to describe real world 3D sound reinforcement devices in real time and ignore variance in that effect due to different frequencies, which I'm sure you're aware of from the misapplication of "impedance" as if it were frequency independant.

            If you want a polar of multiple frequencies at various off-axis measures, I can do that too. I'd prefer to provide it as a 4D stack of time/frequency analysis plots done in continuous wavelet analyses as it's confidence level increases exponentially over Fourier. Of course you're need the Matlab signal analysis package to visualize the output.

            I have access to an anechoic chamber at Virgina Tech and could arrange this testing. I'd appreciate it if you could tell me the data format would best suit you and precisely what measurement parameters would facilitate this.

            Also, I've already addressed the extreme interaction/diffraction "problem" as should be expected from this design elsewhere in this forum. I too follow some theoretic directions when testing such things. Mine comes in large part from Richard Feynman's "no-one has ever been able to define the difference between interference and diffraction satisfactorily. It is just a question of usage, and there is no specific, important physical difference between them."

            Since a measurement based demonstration (too few data points to be called an experiment) should have a proven design to compare against, I again request your assistance with regards to a baseline to compare against this. I did develop a simplified version of a multiply folded horn utilizing a single driver at the base (plus additional wide dispersion piezos for high end) that had adequate transformational characteristic. But the design I took as inspiration also made use of non-optimal geometry within the horn (various lengths of rectangular cross section segments, the non-circularity of which would produce frequency dependent intereference), almost certainly resulting in the sort of effect you initimate, perhaps there's a better design against which I might compare these. Or I could compare these with the original inspiration directly and see if perhaps one or the other, if either, proves superior.

            Comment


            • #7
              Re: Another Cube Cab

              Originally posted by Sydney View Post
              The original concept was called "Sweet 16" in the January 1961 Popular Electronics magazine I read as a kid.

              Precisely. an array of identical drivers produces a planar presentation.

              The 3D interaction of the drivers produces, at the front of the transducer, such an enormous interaction that the "comb filter" problem receeds into microscopic differences.

              This design is the Sweet 16 with the addition of lower and higher frequency drivers to fill the spectrum. Plus the directionality.

              Comment


              • #8
                Re: Another Cube Cab

                Originally posted by drmcclainphd View Post
                I too follow some theoretic directions when testing such things. Mine comes in large part from Richard Feynman's "no-one has ever been able to define the difference between interference and diffraction satisfactorily. It is just a question of usage, and there is no specific, important physical difference between them."
                To my understanding:
                Feynman included: The best we can do, roughly speaking , is to say, when there are only a few sources, say two, ( as in Young's slits, ) it's called interference, but with a large number of sources, the process is labeled diffraction.
                That seem to be the consensus of traditional useage.
                "Not a Speaker Designer - Not even on the Internet"
                “Pride is your greatest enemy, humility is your greatest friend.”
                "If the freedom of speech is taken away, then dumb and silent we may be led, like sheep to the slaughter."

                Comment


                • #9
                  Re: Another Cube Cab

                  Originally posted by drmcclainphd View Post
                  I have access to an anechoic chamber at Virgina Tech and could arrange this testing. I'd appreciate it if you could tell me the data format would best suit you and precisely what measurement parameters would facilitate this.
                  .
                  You can save yourself a lot of time and effort and just use the applets here to see what happens when multiple point sources are employed, compared to a single point source:
                  http://www.falstad.com/wavebox/

                  There's this also, which shows how poorly even two point sources sum, and the more point sources one employs the worse the result:
                  http://homepage.mac.com/randyhyde/we...udioStuff.html
                  www.billfitzmaurice.com
                  www.billfitzmaurice.info/forum

                  Comment


                  • #10
                    Re: Another Cube Cab

                    Originally posted by billfitzmaurice View Post
                    You can save yourself a lot of time and effort and just use the applets here to see what happens when multiple point sources are employed, compared to a single point source:
                    http://www.falstad.com/wavebox/

                    There's this also, which shows how poorly even two point sources sum, and the more point sources one employs the worse the result:
                    http://homepage.mac.com/randyhyde/we...udioStuff.html
                    Well, amongst other complicating factors, they're not point sources. They're full size driver cones. The HF beaming inherent to them actually assists in integration, and ought to be closer to a planar wave than a point source sim would indicate.

                    I agree with you that it wouldn't be for me due to the multipath behavior and resultant poor transient response, but certainly an array like that could have massive output and high efficiency.

                    Comment


                    • #11
                      Re: Another Cube Cab

                      Originally posted by badman View Post
                      Well, amongst other complicating factors, they're not point sources. They're full size driver cones.
                      You must not know what a point source is. It has nothing to do with the size of the drivers. In this case each individual driver is a separate point source at those frequencies above where the center to center distance between each adjacent driver is one wavelength.
                      www.billfitzmaurice.com
                      www.billfitzmaurice.info/forum

                      Comment


                      • #12
                        Re: Another Cube Cab

                        Yeah what Bill is saying is that below a certain frequency (probably 700hz or so) these drivers will couple together mutually and not have really any bad side effects. But above that frequency they each 'beam', and this along with it's interaction with the other drivers causes severe 'lobing' in the polar response and 'comb filtering' in the frequency response. The response just gets ugly and inconstant throughout the coverage.

                        That is to say that if the drivers were crossed over low enough, and used with another HF element, this configuration would be totally fine by our standards.

                        Comment


                        • #13
                          Re: Another Cube Cab

                          Originally posted by killersoundz View Post

                          That is to say that if the drivers were crossed over low enough, and used with another HF element, this configuration would be totally fine by our standards.
                          Adequate maybe, but not necessarily fine. Even if mutually coupled throughout their bandwidth horizontal dispersion will be far narrower with a block driver arrangement than with a taller/narrower arrangement. For best results the total width of the radiating plane should be no more than 1 wavelength. At 100Hz that's a very easy task; at even 1kHz not so much, let alone 10kHz.
                          www.billfitzmaurice.com
                          www.billfitzmaurice.info/forum

                          Comment


                          • #14
                            Re: Another Cube Cab

                            I'd offered to produce objective measurements regarding some designs and their output characteristics. Two semesters of short notice substitute teaching at 2 different colleges and caring for my wife after 2 strokes slowed me down, but did not stop me.

                            My colleagues declined to perform measurements in their anechoic chamber foremost because it was engaged in someone's funded dissertation regarding acoustics at low pressure such as on Mars, having more to do with low pressure CO2 burning catalytic jet engines than anything sound-wise, but there it is. Three days to bring it down and stabilize it at working temp/pressure made the offer unfeasible.

                            But they also declined on some prima facie evidence that made the argument absurd, not only from their engineering stand point, but also from mine as a neuroscientist specializing in the physiology of the auditory system, from ear to cortex. I'd be ashamed I hadn't thought of it, but two people I taught about the relevant phenomena did think of it, and they were professors, so I reckon I did my job.

                            The comb filter effect occurs when "multiple sources" produce the same signal at different distances from the measurement device. It exists and can be measured. The more sources, the more it occurs. It also occurs when the source is not a point, hence any source not a point (ie. having a diameter) is a "problem" in and of itself.

                            Multi-driver columns are a popular design right now. Their theory is that as the multiple waveforms spread, they form a columnar wavefront, a means of focusing a signal in space. They do so because of the constructive interference side of the comb filter thingy. But, the vertical columns produce a vertical column wave. Maybe the disrtortion doesn't show up until the distance between them is greater. Fine. So turen your head sideways. Do you suddenly hear massive dropouts in certain frequencies? Me either.

                            The Karlson designs made use of frequency dependent exit venting to produce a reverberation effect from a single source, effectively producing a psychoacoustic depth effect that would eventually be reproduced in electronics with cross channel phase shifting. It worked because the ears and brain are tuned to differences in frequency and phase with much more sensitivity than you are used to experiencing directly. They use it to produce source localization; this is not only where something seems to be, it is the very essence of stereo imaging. One point source and two point ears can do it. Two point sources and one point ear can do it. Neither ear nor source can be a point, which is good because the effect produced makes it easier to perceive localization and stereo imaging. Got more than one speaker and/or ear? Me too.

                            The blurring of comb filtering depends on (1) the area of the source(s), the distance from source to head and (3) the distance between the ears. It also is frequency dependent in that these will occur at different distances according to the frequencies involved, but will occur regardless. What area of driver surface or combination of driver and matrix appears as a point source? In reality, none, regardless of theoretical maneuverings to exclude natural effects from nice clean theory. So let's just say for the sake of argument that you DO have a point source, you DO have a single ear with a point opening and you have NO impedance effects to contend with due to intervening wave carrying medium, because your ear is perfectly impedance matched (ie. directly connected) to the source. Just for the sake of argument, mind you. What planet would you be from? Not this one.

                            Your auditory system not only takes this comb filtering "problem" into account, it absolutely requires it for perception. The brain reacts to the variance in sound arrival by capturing, recognizing and filtering out latter identical signals. The brain recognizes the sounds initially in he mid-latency processing in evoked potentials called N30 and P50. This process is when your brain determines which came first and which came how much later. How much you end up paying attention to such things can be predicted by watching how these signals produce a negative swing to the rest of the auditory processing from 100 to 150 msec post stimulus, the "mismatch negativity". Note that none of this requires your attention because it all happens before your attention. You become aware at 250-350 msec. This all produces what you might become aware OF, but most definitely produces perceptual effects whether or not you are aware of them. If you did not detect the comb filter effect at the perceptual level you could not possibly hope to detect distance and range of a sound source, and if you could not do so fast enough to react without thinking about it, your brain would tend to become, over time, a member of that class of organisms less suited to evading predation -- extinct. The "problem" isn't, outside of theoretical discussions removed from reality.

                            So let's take the Sweet 16 design and examine it. Just as a column produces a columnar wavefront, a 4 x 4 array produces a planar wavefront. Is comb filtering present? It has to be, not just because theory says so, but because the comb filtering interference makes the indiviual source differences interfere each other out of existence leaving constructive interference producing a planar wave.

                            So what's the "problem"? There's a couple parts to it. One is neglecting the evidence of listening enviromnment. Does the ear hear it? You'd best hope so, and if and when it does, it won't come out as a problem, but rather as a useful perceptual cue. So you've got a microphone that's sensitive enough to pick up theser pockets of interference drop outs near the drivers. How near? Nearer than you listen. And is it relevant? That depends on whether you have ears and a brain, or a microphone and a measurement device. The latter can detect the phenomenon. Does that make it a problem? To the population of microphone and measurement device equipped organisms, possibly. To us, emphatically no. The other part of the problem is the "knows enough to be dangerous" effect. When one accumulates a depth of technical understanding without a concominant breadth within which to place things in context, one can only arrive at conclusions such as Neils Bohr spoke of: "The experiment does not allow us to be able to say anything about nature. It only allows us to say something about the experiement." I ran across this problem working with people on amateur rocketry projects. One such was the common knowledge that supersonic vehicles should have conical noses. They all knew the names of the rockets that had them. Unfortunatly common knowledge didn't include the fact that those noses were as they were because they were the easiest to manufacture. And the common knowledge presenting itself as authoritative prevented them from looking up freely available scientific evidence that the conical nose was one of the worst designs for breaking Mach. Similarly, kit was "well known" that high airspeed stresses could rip rockets apart if they weren't glued together well. But nobody bothered to examine the strength of glue joints compared to the materials glued. Even the cheapest plain white Elmer's produced glue joints stronger than the materials it was used on. Therefore, if something failed, it was not the joint, or if it was it happened because the glue wasn't used properly. Even after explaining this to them, the "experts" continued on their merry way arguing the minutia of things of which they knew everything about half of, and nothing about the other half.

                            Let's drop back briefly and re-examine the concept. Let's say we do have a point source, and microphone measurements suffice. Now we want to use an impedance driven acoustic amplifier -- a "horn" -- to make it louder. The optimum shape is a exponentially expanding cylinder. Optimum, not perfect. The wavefront cannot expand evenly because of friction at the edges, and because the air molecules bouncing around in their Brownian fashion do not all head straight out down the horn length in lock step like a molecular laser. The application of a transforming phenomena -- exchanging one kind of energy for another -- introduces a source of entropy beyond the random thermal noise of a simple object. This is thermodynamics. Now, if the horn is not perfectly round, the different distances that reverberations can travel inside the horn will produce the dreaded effect, and if there are bends in the path, all the more so. And then comes the horn mouth and the diffraction caused by the sudden expansion of the conducted wavefront into open air all of the sudden.... Talk about a comb filter nightmare ... well, go ahead and talk about it. It doesn't care, and you can't hear it. If you can actually measure it, great. If you don't measure it at every source of every different kind, you're picking your parameters. Nature doesn't do that.

                            The Sweet 16 front with 4 woofers behind, two towers with 2x6 arrays of the same drivers are the 16, and a pair of Fender combo amps with 12" drivers and both have expansion cabs with identical drivers, and all spread around the stage... how come it is that 16 identical drivers packed parallel into a panel less than 1000 square inches is a nightmare but three times the driver surface area spread over 200 square feet of stage is a concert?

                            And then there's this obviously horrible sounding junk pile:
                            http://makinmusic.com.au/estore/imag...emSpkrWall.jpg
                            Floyd thought it was pretty cool, but Pink kept looking for the knob to turn it up to 11.

                            Just because you can make some numbers dance doesn't mean you can use them to prove something. Take it from a scientist, you have to use ALL the numbers, including those the world around you forces on you. Otherwise it's all exercises in applied make believe.

                            Comment


                            • #15
                              Re: Another Cube Cab

                              Originally posted by drmcclainphd View Post
                              how come it is that 16 identical drivers packed parallel into a panel less than 1000 square inches is a nightmare but three times the driver surface area spread over 200 square feet of stage is a concert?
                              Your making the assumption that the concert you described is not full of problems.

                              We should all strive to understand the difference between perfect, ideal, acceptable, tolerable and doable. Frankly what you have done here is not altogether different then some of the original (pre-production) Bose designs, and in some ways reminiscent of a turned around 901.

                              I find the experimentation aspect interesting, in spite of the pessimism of others. There is a value in the experiment, regardless of the outcome.

                              Hope your wife is doing well.

                              Comment

                              Working...
                              X