just a glimpse until he gets his website back up: http://www.avsforum.com/t/1311851/new-waveguides-for-the-octagon
here is the text from his site:
Paul's AV System
A personal journey…
My parents provided early and pleasant exposure to live big-band and classical music. Even through the eyes of a kid, I’ll never forget that brunette singer at a Les Elgart concert! However, the late 1960’s rock bands built the truly lasting foundation for my love of music and audio. Those bands possessed seeming endless creativity and energy consistent with the times. Cream, Jimi Hendrix, Pink Floyd, Jefferson Airplane, Buffalo Springfield…the list goes on and on. I would never lose touch with jazz and classical -in fact I enjoy both now more than ever- but those early rock bands were great! These were my college days…I had little money, but nearly always managed to have sound. My first speaker was a cheap driver with a whizzer cone mounted in a heavy shipping carton, the first turntable an old Garrard, but I had sound!
The 70’s bands were a continuation of the late 60’s for me…and the bands continued to build my interest in audio. A real paycheck gave me access to some of the better drivers of the day, including KEF, Polydax, Dynaudio, etc…I even managed to acquire a pair of 24” Hartleys. However, other than understanding the basic functions of inductors, capacitors, and resistors, I lacked crossover design skills. My measurement equipment consisted of an ADC SPL meter and test tones on a couple of LPs, so I was nearly flying blind. I enjoyed building those speakers but, by modern standards, none performed particularly well.
In the early 1980’s, results improved when I began building ribbon planars and large ESL panels. These were all two-way, so my crossover skills were not such a limiting factor as they were with the three and four ways I built in the 70’s. During the 80’s I used dipole and sealed bass modules and eventually built the big Hartleys into the fireplace…my first IB sub before I’d heard the term associated with subwoofers.
The music of the late 80’s and 90’s was not compelling and, by that time, life had become way too busy. All of my hobbies including sailing, skiing, racing, etc, suffered from an increasingly demanding work life. All of the equipment was sold or put in storage and serious audio was nearly forgotten for about fifteen years.
The second time around…
Ironically, DVD video rekindled my interest in audio. In 2000, my name-brand home theater speakers did not sound that good and, considering the money required for truly upscale commercial speakers, I thought I might be better off building my own…deja-vu.
I had not built a speaker since the early 80’s so I read several newer books on the subject, re-read my old books, armed myself with LspCAD, and restarted my long-lost hobby. The synergy of simulation and optimization software interacting with the books proved to be an extremely powerful learning tool…far better than either alone. Additionally, internet sites like those of Seigfried Linkwitz and John Kreskovsky have been incredibly helpful. Backed by these resources, subsequent measurements and simulations produced fairly quick success building a new generation of speakers.
With improved speakers and reasonable video quality through RPTV came the realization that I really enjoyed concert videos. Recording a live performance is a dicey proposition on several levels, but including the visual part of the show -and a good bottle of wine- makes overlooking a few audio warts very worthwhile. Better yet, a strong performance with HD video and well recorded lossless audio can be downright euphoric, even without the wine! A large screen, clean picture, and clear sound with great dynamics would be the long term package for me.
Unfortunately, with no time for serious audio or video in the 90’s, AV had not been a significant consideration in home design in 2000. While the multi-purpose great room available was very large (a good thing) the acoustics were too live and the light color scheme was not video friendly. However, considering room limitations, the system evolved nicely with a drop-down screen in front of the fireplace, HD front projector, large IB subwoofers and a series of good speakers. Weekend nights were really enjoyable…life is good!
2009, the new system…
Building another home ignited an undeniable urge to include a purpose-specific AV room. The opportunity to combine a room and speakers into a tightly integrated system was too good to let pass…plus it would present worthwhile challenges along the way. The new room would have reasonable volume, but potentially problematic dimensions of 10’ x 21’ x 30’ mean that room modes must be addressed very carefully. Two solid concrete walls and a concrete floor would ensure a complete lack of boredom!
Waiting for home construction provided more than enough time to design the system. With a career in commercial product planning, it was a true joy to plan a “product” without a particular P&L constraint; with my own obsession the only customer need. In fact, the build plan evolved into this web page…so I hope you enjoy the read!
The new room…
The basic objective would be removing the room from the listening experience…to deliver program material unfettered by the environment. To accomplish this, three front channels would illuminate a broad soundstage with minimal high-level early reflections while room acoustics permit surround channels to create the subtle illusion of a much larger, more diffuse, acoustic space. Rather than fight to tame a typical lively room, the strategy was to create a somewhat dry outdoor-like environment and then add surrounds and modular diffusers to create acoustic life at appropriate level, direction, and timing.
First on the agenda was a bit of sound isolation for the neighbors. The exterior framed wall was doubled with 2x6 framing on 8” centers, the 2x6’s tied together in a “trellis” pattern with horizontal 2x2’s and, finally, a second layer of 2x6’s on 16” centers…all to make the long outside wall as stiff as practical.
The reinforced exterior wall and a false wall on the opposite side of the room, enclose surround speakers and at least 12” of fiberglass insulation. The deep-wall fiberglass provides broad-band absorption as recommended by Floyd Toole…“absorb all frequencies or absorb nothing”. For adjustment flexibility, absorption is exposed where wanted and covered by curved panel “scattering devices” at appropriate locations, in particular the walls to either side of listeners. The curved panels promote diffuse late reflections while preventing slap-echo.
With a goal of zero moisture penetration, the concrete floor was first sealed with sodium silicate, then an asphalt sealer, and covered with roofing felt. Next, 8”x3/8” fiber cement sleepers were laid with a 24”x40” open pattern and covered by a second layer of felt. Next, a “floating” floor of overlapping ½” and ¾” inch OSB is screwed together only at the seams, the two layers separated by a third layer of felt. Finally, a half-inch 8-pound pad, and carpet complete the floor. Though it might sound overly complex, the floor system was actually quite easy to build. The reward is a strong, well insulated, well damped, resilient floor capable of conveying a sense of tactile bass…and, it is extremely comfortable underfoot.
At higher frequencies, the floor is relatively reflective so the ceiling was made near “dead” with black fabric covering 12” of fiberglass between open trusses…no diffusion. This substantially reduces high frequency ceiling bounce and creates nearly opposite acoustic qualities in the relatively close parallel surfaces…the floor and ceiling are by far the largest surfaces in the room. To reduce floor bounce, large floor pillows are placed partway between the speakers and listeners.
From an acoustic standpoint, the basic design thrust for the rear wall was to provide high frequency diffusion and low frequency damping…sometimes conflicting objectives. Angled false walls define the rear corners of the the AV room and separate it from the lobby, recreation area, workshop, and storage areas. These false walls house “limp-mass membranes” of layered roofing felt suspended from ceiling to floor. These membranes damp and vent very low frequencies while reflecting higher frequencies back into the room. The angled walls also incorporate membrane-backed media shelves and additional surround speakers. An eight foot desk/open equipment console spans the center-rear wall. An eight foot closet behind the console provides easy access to equipment cabling and “crossover central”. The rear wall of the closet also includes limp-mass membranes. Full louver doors to the lobby complete the lossy/diffusive rear wall.
Diminishing a potential recording studio look, the surround speakers, diffusers, and membranes are all concealed behind dark cloth panels to create smooth finished walls. When desired, curtains can be drawn to conceal the console.
Having previously lived with leather furniture (and heard and measured the interference from reflections) only cloth upholstered seating is used for the new room. Mid-back chairs (no headrests to interfere) make great critical listening seats while a favorite sofa and extremely comfortable chairs provide seating for movie nights.
Moving from 7 to 11 channels…
For this room, surround channel program material should provide a virtual reality; the ambience of a far larger space replacing short reflections of the acoustically small room. This system would depend less on room reflections -which are too early to adequately portray a large space- and depend more on actively managed surround channels.
With subtle, music oriented, spatial envelopment a high priority, I sought a even greater level of apparent diffusion and space than presented by 7.1. Floyd Toole provided the solution in his excellent book: “Sound Reproduction, Loudspeakers and Rooms”. Floyd describes a 3/6 system (3 front channels, 6 surround) where the additional channels are derived by means of electronic delay and spectral contouring. My system is a simple extension of this process where four additional channels are derived from the original surround channels through digital delay and very slight attenuation of high frequencies to emulate propagation loss due to the additional delay (virtual distance).
The resulting eleven channels are arranged at 0, +/-30, +/-60, +-90, +/-120, and +/-150 degrees. For now, the +/-60 degree speakers are the normal left/right surrounds, +/-90 is the delayed version of those channels, +/-120 is the delayed version of the rear surrounds, +/-150 are the normal rear surrounds.
The surround speakers…
Dipole surrounds have horrendous polar response and bipoles would disrupt the smooth look of the walls…so in-wall monopole surrounds were specifically designed for very wide dispersion to create a fully immersive soundfield. Robust four-way surrounds make wide dispersion possible at high output and low distortion. Increasing the number of surround channels from four to eight not only decreases localization, but further reduces driver excursion (and distortion) for any given sound level. These surrounds may seem like overkill but, particularly with action movies, they must “keep up” with extremely dynamic LCR speakers.
The drivers for each surround channel are arranged in vertical lines, moving up from the woofer at ear level. This places the soundfield at the intended height and approximates a slanted baffle on a floor-standing speaker, providing closer alignment of driver acoustic centers with respect to the listening area. Crossover networks are loosely based on 3rd order acoustic filters at 260Hz, 900 Hz, and 3.3k and optimized for a listening sweet spot at 3 meters. To provide greater compensation flexibility, the lower crossover point was determined by the geometry of woofer and mid-bass floorbounce. Mid and upper crossover points were selected based on driver directional and distortion characteristics.
Surround tweeters are the 19mm Seas 22TAF/G magnesium/aluminum dome, upper midrange is the wide dispersion 2” Dayton RS-52 aluminum dome, lower midrange is the very capable 7” Dayton RS-180, while the woofers are 12” Peerless SLS 830669. Driver output is fairly well balanced to yield high overall system sensitivity.
Driver backwaves are directed into the 12” fiberglass side walls for quasi-aperiodic absorption. Relieved of the need for baffle step compensation, the in-wall solution yields extra headroom which is consistent with one of the basic system objectives.
Subwoofers and Front Channel speakers…
All of the speakers I’ve built in recent years have included the objectives of high output with low distortion. Why the focus? Reducing HD and IMD provides a background free of spurious signals not present in the original program material. This “black” low-noise background not only allows greater detail retrieval, but lower distortion can also reduce corruption of musical overtones so the natural timber of vocals and instruments is reproduced with greater accuracy.
Audiophile speakers are generally measured at levels around 88db output at a distance of one meter…and a well executed design will measure quite favorably under those conditions. However, that same audiophile design, at a realistic listening distance and level will generate increased distortion. Particularly on peaks, waveform compression further reduces the difference between the desired signal and the distortion products. No wonder so many speakers become muddy when the volume control is cracked a bit!
One way to reduce distortion was demonstrated by the “Gemini” IB project. In that case, merely average performing 15” subwoofers were teamed in two arrays of nine drivers each. At normal listening levels, excursion was exceptionally low and, in contrast to typical single or double driver subwoofer systems, there was no audible distortion. Headroom is your friend!
Striving for high SPL capability resulting in low distortion, the new speakers would extend the concept of high swept area and low excursion to the full audio spectrum. The signal to individual drivers would be relatively small…the drivers would be merely “tickled” at normal listening levels.
Enclosures and Baffles…“designed for a soundstage and imaging junkie”
Speaker projects over the past few years, and more than a dozen recent experimental baffles, provided several other important lessons including three more that would substantially influence the new Left, Center, and Right channel speakers (LCRs).
#1: “Boxless” speakers are fantastic for their lack of coloration. Cardioid/quasi-aperiodic (acoustic flow resistance), dipole, and infinite baffle all fall in the “fantastic” category because a high pressure backwave is not compressed in a small box to emerge back through the driver cone or excite lingering panel resonances. For me, those attributes provide the real magic in “boxless” speakers. Dipoles work well in some rooms but, unless they are located far from the front wall, the reflected backwave can degrade imaging and confuse the soundstage. With a large projection screen, this room is not sized to position LCR speakers six to ten feet from the front wall, so true dipoles were eliminated from consideration.
#2: Waveguides and other forms of loudspeaker directivity control are valuable tools in the quest for both technical accuracy and an acoustically pleasing environment. Managed polar response can reduce high level early reflections that add little to envelopment, thus revealing finer program details from the speakers. With reduced early reflections from the LCRs, the same room can promote late reflections from all speakers, supporting spatial envelopment while providing a normal conversational environment. No need for an unpleasant anechoic chamber!
#3: For presentation clarity and image stability, the lower edge diffraction, the better. Diffraction can disrupt otherwise smooth response, so better to cure the acoustic source of diffraction than try to hide it with EQ or other directionally sensitive band-aids. If there must be an edge or other imperfect surface, make the distance as far from the driver as practical so the interference event occurs at lower SPL. Low-angle surface transitions and large radius roundovers are a great start in the quest for low diffraction.
So, the new LCRs would be designed for a convincing soundstage through precise imaging by carefully managing radiation patterns, including the difficult-to-control midrange. High radiating area would not only support high output with low distortion, but also aid in directivity control. Large low-diffraction baffles and waveguides with large roundovers would help reduce early reflections while one of the boxless solutions would suppress backwaves. Physical construction would focus on low energy storage.
However, the new room is not big enough for three large stand-alone baffles plus IB subs. IB…humm…the concept quickly progressed to oversized “IB style” sealed enclosures for the LCRs. Clearly, the enclosures could not be truly infinite in size, but they could be large enough to deliver very low system Q. For minimal coloration, backwaves would be diffused and absorbed with a goal of zero reflection to the rear of the drivers.
A little further evolution produced a front wall integrating the subwoofers and LCRs into a 20’ wall-to-wall main baffle with individual sub-baffles splayed in a broad arc across the front of the room. The L&R speakers are placed at a music oriented +-30 degrees, with axis intersecting at the main listening position 4 meters distant. This mild toe-in provides full listening coverage through the depth of the room.
Though the main baffle may resemble the surfaces of an F-117 stealth fighter, front wall reflections, baffle step, and diffraction are all addressed with this single solution. Placing the subwoofers in roughly 1/4 space increases operating headroom and forms the foundation for the “low excursion = low distortion” theme implemented throughout the system. With no need for efficiency robbing baffle step compensation, the LCRs continue the theme.
Large enclosures are a double-edged sword. Pressure modulation is low, so they reflect less energy back through the cones; however, larger panels tend to vibrate vigorously at lower frequencies. To develop a construction strategy, an accelerometer was used to test samples of laminated material and adhesives. This resulted in building the baffles as stiff as practical to move mechanical resonances above the operating range. “Screw & glue” construction was used for the baffles and throughout the room.
Each LCR midbass displaces about 30 cubic feet, with six-layer OSB walls. All three enclosures are trellis braced with 2x2’s, inside and out. The concrete front wall and floor of the room serve as the floor and back wall of the enclosures. To begin, the centers of the enclosures are partly filled (not stuffed) with fiberglass batts…though heavy freestanding internal baffles can rest on the floor to enable adjustment of reflections and standing waves without performing surgery on the basic boxes. OC 703 fiberglass panels spaced atop the internal 2x2 lattice further minimize internal reflections.
High driver density reduces the four-layer LCR baffles to “Swiss cheese”, plus they are heavily relieved to reduce driver shrouding…however the baffles are carefully matrix-braced with 5/4” Spruce to be even stiffer than the main baffle panels. The central vertical partition also serves as a “ski” to ease installation of the 100+ pound driver-heavy baffles.
At about 600 cubic feet, the remaining irregular space behind the main baffle forms the subwoofer enclosure. Four-layer subwoofer baffle panels are laminated from OSB under an outer skin of MDF. These panels are installed over a closely spaced 2”x6” skeleton with 4” deck screws. The center of each upper skeleton member is braced directly to the concrete walls behind the baffle; the lower third of the baffle is anchored to the floor and walls by the exceptionally rigid LCR enclosures. Since fiberglass is relatively ineffective at very low frequencies, limp-mass membranes are suspended from the ceiling of the subwoofer enclosure to damp potential internal acoustic resonances.
Horns & Waveguides…
After reading for more than two years, I realized one could spend a lifetime studying horns and waveguides, for it is truly a humbling topic. The flexibility to experiment with Constant Directivity, as well as flares that narrow directivity with increasing frequency, like the Spherical and Tractrix, is very intriguing but it introduced a challenge with regard to baffle construction. One difference between CD and Variable Directivity flares is that VD flares tend to “beam” higher order distortion products directly on-axis…perhaps this is the source of occasional subjective comments that Tractrix horns can “hot spot” on-axis.
Should the main baffle be built so the waveguide would be on axis with listeners, or, fire straight into the room with the axis above the listeners? Changing the angle of the waveguide baffle after-the-fact would produce offsets at the top and/or bottom of the baffle introducing small, but unnecessary, diffraction generators.
In the end, the potential of focusing high-order distortion products directly at listeners seemed a really bad gamble so the upper main baffle was built with a three degree down-angle, firing just above listeners’ ears. The three degree angle is fully compatible with CD flares and a very reasonable compromise for the beamy VD flares.
The relatively abrupt frequency “shelving” generated by constant directivity flares and the beamy response of the VD flares both have strong proponents…but both extremes seem compromised. For home systems, a better answer may lie between the two, so I plan to eventually experiment with smooth, slowly increasing directivity from as low in the midrange as practical through the top octave. Controlled directivity, not constant directivity.
However, ya gotta start somewhere…and I wanted to listen! While not religiously attached to any flare, I decided to begin with constant directivity flares as they appear to strike a good balance of compromise for home application and would provide a baseline to compare performance of later experiments. Geddes uses 90 degree coverage while Toole indicates horizontal coverage as narrow as 60 degrees is appropriate for LCR speakers…I elected to start with 80 degree waveguides. Eighty degrees would significantly reduce early reflections from nearside walls while promoting later reflections from walls opposite the waveguides.
The waveguides designed for this system are “compound spheroid”. For low diffraction in the throat area, a 10 degree oblate throat entry matches the HF driver exit angle. The oblate throat then flares tangent to a large prolate spheroid mouth. The low initial angle of the PS flare tends to hold directivity as low in frequency as practical. More rapid expansion toward the mouth prevents the radiation pattern from abruptly narrowing toward cutoff and provides a low diffraction transition to the main baffle. An overall size of 30” diameter (~six wavelength perimeter) extends driver loading below crossover as well as reducing LF energy storage and group delay through the crossover region.
After exploring several different alternatives, the physical waveguides were built of well braced fiberglass for a rigid platform of reasonable weight. The waveguide base and mounting panels simply slide into place atop the midbass enclosures and are then screwed into the main baffle.
At 9X the radiating area of a 1” dome tweeter, compression drivers for the three LCR waveguides are the Eighteen Sound ND1460A. These short-throat 1.4” exit drivers are SOTA with ultra light 3” aluminum diaphragms centered by elliptical polyester suspensions driving a three-slot circumferential phase plug. The motor utilizes a radial neodymium magnet structure, copper shorting ring on the pole piece for reduced distortion and extended HF response, edge-wound aluminum voice coil on a Nomex former, and active cooling for reduced power compression.
Midbass drivers for the LCRs are the Dayton RS-180. The motor implements two copper shorting rings for reduced non-linear distortion while the aluminum cone remains pistonic well above crossover. The clarity and articulation of fine detail provided by the small metal cone is further enhanced by an excursion/distortion reducing twelve driver array. The array also provides the power handling and directivity of a much larger driver.
Eighteen Dayton 15” DVC subwoofers support the bottom octaves (these are the same drivers originally used in the “Gemini” twin IB project). This relatively low-Q, low inductance driver uses a Kevlar impregnated cone and implements dual voice coils on a Kapton former. As with the compression drivers and midbass arrays, high swept area reduces excursion, resulting in low distortion with exceptional dynamic range.
Configured as a “Planar Bass Array”, the subwoofers and room operate as a closely coupled system. The PBA minimizes room height and width modes and launches a planar wave at low frequencies. The room depth, or length, mode is damped by the limp-mass membrane just outside the back wall. This arrangement is somewhat analogous to the acoustically infinite “plane-wave tube” used to measure compression driver response…except this tube is for subwoofers, with listeners inside the tube rather than a microphone. Overall, the PBA with membrane provides very even bass throughout the room by minimizing roller-coaster frequency response and lingering boom of room resonances.
Developed using ArrayShow, ARPE, and LspCAD, the “+” shaped LCR midbass driver arrays provide vertical and horizontal pattern control to reduce interaction with surrounding surfaces. For improved front-wave coherence and reduced floor bounce, the midbass panels are tilted upward six degrees toward listeners. While a small offset has trivial impact on directivity, impulse response is improved by delaying the center four drivers by 11mm. Additional polar management is possible by changing the driver layout and/or changing the frequency and phase response of individual drivers within the array. Radiating into an approximate quarter space environment, the twelve driver midbass arrays are capable of prodigious output…again, the unusually high output capability increases headroom and lowers distortion.
Both the upper and lower LCR crossover points were determined in light of driver distortion measurements at high, medium, and low SPL levels. These measurements resulted in a fourth-order crossover at 70 Hz for the midbass/subwoofer crossover. Midrange cone breakup in subwoofers is often overlooked because it is assumed crossover slope is steep enough to adequately attenuate the breakup…in high definition systems, this assumption is not necessarily accurate. This system includes two midrange notch filters to tame subwoofer breakup. Subwoofer EQ is limited to “hi-cut” which minimizes group delay and effectively extends LF output with a very slow rolloff toward DC.
For each LCR, in the lower octaves where excursion is highest, twelve midbass drivers share the load equally. As frequency increases and excursion becomes insignificant, a capacitive shunt gradually diverts energy from the outer eight drivers to the inner four drivers; effectively altering array size with frequency. This smooth change from the full array to the central drivers prevents in-band array lobing while providing excellent lower and upper midrange directional control. (800Hz wavelength is 17”, outer driver CTC is 25”, inner driver CTC is less than 8”.) Additional benefits of the array/crossover include a 6 db increase in sensitivity while maintaining an amplifier-friendly 8 ohm minimum load and a flat acoustic transfer function. Adjusting the twelve-to-four driver transition provides a means of fine tuning the midrange radiation pattern to ensure a smooth directivity handoff to the HF waveguide.
A steep active crossover at 800Hz provides extremely narrow overlap for minimal lobing between the midbass array and the compression driver/waveguide. Appropriate diaphragm breakup notch filters are included in all filter branches.
Including the surrounds, all eighty-nine drivers (whew, that was a lot of cutting!) are wired to crossover patch panels mounted on the walls in “crossover central”…a small room behind the equipment console. The patch panels contain all of the passive crossover components; allowing quick and easy access for circuit changes. Easy access encourages continuing improvement!
First, the neighbors are happy! All along, I was concerned that the long, straight exterior wall would shake at LF. Thankfully, the triple 2x6 wall and trellis bracing did the job. Sound outside the wall is about the same level as the cicadas in the trees!
The bass, ahhh the bass…the Planar Bass Array sounds like live bass in a larger room…clean, deep, and powerfully easy. In even better harmony with the room, the PBA is a clear improvement over my previous IB installation in that it provides truly seamless integration with the LCRs. Guests have asked if I have subs on because they “can’t hear subs”…this even when there is a house curve lift at 20Hz!
If you’ve read this far you know this system will play silly loud. However, the goal here is not silly loud, rather it is the highest practical sound quality at reasonable levels. High amplifier power coupled with low driver excursion for a given SPL translates to low compression and low non-linear distortion. Listeners are rewarded with a very articulate presentation of the finest recorded details and, seemingly, unlimited dynamics. The clarity of lossless audio is simply outstanding.
The two-channel-plus-center soundstage extends through the LCR spread, with depth generally dependent on program material. With early reflections suppressed, imaging is extremely precise…individual instruments and vocals “pop” across the soundstage.
Activating the surrounds provides an astonishing contribution to stereo source material. In particular, the surround channels toward the front of the room provide substantial ambient envelopment, pushing apparent source width and depth well beyond the front walls. The channels toward the rear complete the illusion of an acoustically large space. “Airy” ambient effects are about the same, whether from a 2-channel or multi-channel source.
I had been concerned that following Toole’s lead in placing surrounds toward the front of the room would disrupt the presentation of discrete multi-channel music and movies. Instead, movie sound effects and music with instruments in the surround channels are presented in a “panoramic” form much more to my liking than ping-pong surround. Since soundstage width is pushed beyond the side walls, it exceeds any possible screen width. With movies, the panoramic soundstage seems entirely natural. Some concerts have acoustic images well off screen so, in those cases, the visual/auditory disconnect is easily noticed…but far more realistic than musicians beside or behind! Just close the eyes and enjoy the “front and center” concert seat!
Behind cloth walls, all of the speakers are “visibly invisible” so it is quite easy to “get lost” in the program material whether it be concert or movie. Power with finesse.
The speaker and room package was developed using ArrayShow, LspCAD, Hornresp, SoundEasy, RRC, SPLmax, and several other software tools. I’d rather run a thousand simulations than rebuild any significant portion of this system. In the end however, software simulations cannot substitute for overall system measurements and extended listening. Due to uncertain absorption coefficients for example, the complexities of “real room” conditions will trump the simulations. Therefore, ongoing evaluation and improvement is an important part of the plan. The process will be a lengthy journey, involving hundreds of measurements and hundreds of listening hours.
First under the microscope will be the room itself, including the limp-mass membrane rear wall (there are many alternatives). Energy/time/frequency characteristics of the room will be carefully measured and evaluated and changes made as appropriate. Basic building materials are used rather than expensive commercial sound treatments so there is no “investment incentive” to stay with a compromised result. For example, if listening indicates more diffusion is appropriate, diffusion is easily added behind the fabric walls and ceiling.
The first room modification was actually made during development of the surround crossover networks. Undulations in the curved panel scattering devices were causing diffraction ripple in the surround channel midrange and treble, so the recesses directly to the sides of the drivers were covered with hardboard. This reduced peak/peak ripple from 5db to 1db…not bad for a $10 room modification!
Since cutting new baffles is far easier than building new waveguide molds, midbass driver configuration and radiation pattern optimization are next on the list. The three LCR baffles are easily replaced and each one can support up to sixteen drivers to study pattern interaction with a “real room” and real listeners. After the midbass is finalized, the need for any new waveguides will be determined…and built as appropriate.
At that point the active LCR crossovers may be converted to steep passive crossovers to eliminate an unnecessary analog-digital-analog conversion. Additional headroom is always a good thing, so passive crossovers will also facilitate running the Halo A21 main amplifiers as balanced monoblocks, effectively doubling rail voltage and increasing power output to about 800 wpc.
Also on the list is a critical evaluation of the surround speakers. There will be extensive subjective experimentation with timing and frequency response of the derived channels as well as experimentation with difference signals; the difference between Right Surround and Right Rear for example. There are four unused surround speaker locations already framed into the room, and two extra channels of amplification. There is Dolby Pro Logic IIz and Audyssey DSX with height channels. Audyssey conveniently locates the DSX width channels at the same 60 degree positions I used…sooo many interesting possibilities to explore!
Come to think of it, this process will be a very lengthy journey!
What about video? (This is an AV page, right?)
Because of an outright obsession with audio, video may seem to have lower priority in my system. For me, large-screen video began with an Advent front projector “back in the day”. However, as with audio, the large screen gave way to my work and shrank to direct-view CRT until the late 90’s. At that point the screen became a 55” RPTV followed by a 65” HDTV through about 2005. None of these really transported me to the concerts…though the 65” Toshiba Cinema Series came close enough to whet my appetite.
By 2006 I could no longer resist the urge for a front projector and finally bought an Optoma H78-DC3 DLP projector and teamed it with a 119” Da-Lite Video Spectra drop-down screen over the fireplace…I was definitely hooked. The 1.5 gain screen helped keep reflected light off the white great-room walls and ceiling, but the room was clearly a limiting factor.
Another issue with the room was the lack of a center channel speaker with the projection screen in front of the fireplace. Wide dispersion, low diffraction left and right speakers produced a broad sweet-spot with a solid phantom center, but the “spot” was still substantially “sweeter” than the seats to either side. Since it was impossible to build a high quality center channel into the fireplace, we lived with compromise.
The new room solves both of those problems…no white walls and no fireplace to interfere with a center speaker firing through an acoustically transparent screen.
The projector has been upgraded to to 1080p with more light output to compensate for the lower gain AT screen. Brightness and black levels are personal “hot buttons” so the Epson 6500UB won the day with an excellent price/performance ratio.
Choosing a great acoustically transparent screen is more difficult than I first realized. I finally settled on a 120” Seymour XD screen with 1.2 gain.
The “gear” is housed in an equipment console centered on the back wall. A generic AMD/Compaq PC with all of the speaker measurement and simulation software is connected directly to the AV system. When I hear something I want to improve, I can easily switch to diagnostic-simulation-correction mode. The same PC also performs as a video and digital music server, as well as HDTV tuner.
There are a few more bits on the shelves, but here is a list of the equipment actively supporting the system:
Oppo 970HD digital transport for SACD, DVD-A, CD, DVD-V
Panasonic DMP-BD35 Blu-Ray transport
Integra 9.8 AV processor
Epson 6500UB projector
Three Parasound Halo A21 power amplifiers driving the LCR channels
Behringer DCX2496 LCR crossover
Emotiva XPA-2 subwoofer power amplifier
Behringer DEQ2496 subwoofer notch filter and equalizer
Behringer DCX2496 surround channel processor
Two Emotiva XPA-5 surround channel power amplifiers
Sennheiser HD650 headphone reference
Behringer SRC2496 sample rate converter/DAC/ADC/headphone amplifier
Measurement Specialties ACH 01 accelerometer
Behringer ECM-8000 calibrated measurement microphone
Speaker measurement and simulation software including LspCAD Pro, SoundEasy, BDS, ARPE, SPLmax, UniBox, etc, etc, etc.
Some Assorted System Stats…
Total amplifier power: 5,000 watts RMS.
Total number of drivers: 89
Speaker cable: more than 1,000 feet
OSB: 90 sheets.
Wood glue: 24 gallons.
Construction adhesive: 18 gallons.
Wood screws: 75 pounds.
Nails: less than 1 pound.
Construction time: 1 year.
Loosing 24 pounds during construction: Priceless!
This Web page was written during home construction as a build plan for the AV room. Writing about the design before building the system proved very beneficial in evaluating ideas…if an idea didn’t pass the “sniff test” when documented, something needed improvement. As the system was actually built, details like crossover points etc were updated, and “Results” reported for the Web page.
We all have different objectives and priorities so it is highly unlikely anyone will agree with all of the decisions and tradeoffs made for this system. However, I do hope at least one idea has triggered your own imagination!
The project name, “Octagon”? It comes from the general shape of the room. Not very imaginative, but I have always thought the word sounds cool!