My dedicated audio room build - QuadDiffusor's Big Dig
- Thread starter QuadDiffuser
- Start date
Hi SBNX, yes the Sonus Faber Aida does have rear-projection drivers, but there is zero information on the design goals, and technical details on how the crossovers/rear drivers complement the front firing drivers. I've auditioned them in various locations/set ups, mainly at AXPONA 2022 and AXPONA 2023, but regret that I have not been moved by their design nor sound.
Last edited:
Luckily, I'm going to be able to widen my dedicated basement listening room's width.
Here are the latest dimensions of the structural slabs/walls:
4.20m(H) or 13.8ft = 40.9Hz 1st order resonance frequency
7.00m(W) or 23.0ft = 24.6Hz 1st order resonance frequency
10.45m(L) or 34.3ft = 17.2Hz 1st order resonance frequency
Surface area (floor): 73.2 square meters, or 787.4 square feet
Volume: 307.2 cubic meters, or 10,850 cubic feet
Shroeder frequency: 88Hz
Critical distance: 1.29m
The dimensional ratios of 1.00 x 1.67 x 2.49 will outside of the (theoretically) optimal Bolt-Area graph, but I'm not worried, as there will be PLENTY of bass mode suppression, including a room-in-a-room construction using damped sheetrock, and dozens of Hemholtz resonance plates mounted directly on the four structural walls, including the ceiling.
Here are the latest dimensions of the structural slabs/walls:
4.20m(H) or 13.8ft = 40.9Hz 1st order resonance frequency
7.00m(W) or 23.0ft = 24.6Hz 1st order resonance frequency
10.45m(L) or 34.3ft = 17.2Hz 1st order resonance frequency
Surface area (floor): 73.2 square meters, or 787.4 square feet
Volume: 307.2 cubic meters, or 10,850 cubic feet
Shroeder frequency: 88Hz
Critical distance: 1.29m
The dimensional ratios of 1.00 x 1.67 x 2.49 will outside of the (theoretically) optimal Bolt-Area graph, but I'm not worried, as there will be PLENTY of bass mode suppression, including a room-in-a-room construction using damped sheetrock, and dozens of Hemholtz resonance plates mounted directly on the four structural walls, including the ceiling.
Some preliminary plans for my newly-positioned basement listening room.
[note that my target RT60 will be ~0.5sec, not 0.8sec]
1) positioned into the eastern corner, directly under the main entrance to the house
2) three structural walls with tons of earth/soil behind them, ensuring zero environmental and EMI noise
3) slightly lower (and more manageable) floor-to-ceiling slab height of 4.2m, compared to the previous crazy height of 6.4m
4) width of 7.0m is identical, and length of 10.6m is slightly longer than the previous dimensions
5) axial room nodes calculated to be ~32Hz, ~43Hz, and ~49Hz (but needs to be recalculated, as they do not match those in post #44)
[note that my target RT60 will be ~0.5sec, not 0.8sec]
1) positioned into the eastern corner, directly under the main entrance to the house
2) three structural walls with tons of earth/soil behind them, ensuring zero environmental and EMI noise
3) slightly lower (and more manageable) floor-to-ceiling slab height of 4.2m, compared to the previous crazy height of 6.4m
4) width of 7.0m is identical, and length of 10.6m is slightly longer than the previous dimensions
5) axial room nodes calculated to be ~32Hz, ~43Hz, and ~49Hz (but needs to be recalculated, as they do not match those in post #44)
Last edited:
Room node reduction will be achieved via wall-mounted Helmholtz resonator metal plates tuned to different fundamental frequencies.
[note that the bottom drawing should show 1.5mm (green) plates on the ceiling, and 1mm (pink) plates on the sidewalls - as drawn now, they've been incorrectly reversed]
[note that the bottom drawing should show 1.5mm (green) plates on the ceiling, and 1mm (pink) plates on the sidewalls - as drawn now, they've been incorrectly reversed]
Last edited:
Current proposal, with outstanding questions.
Please chime in with your experiences and suggestions for those prefixed with a "Q" !
Four walls:
1) first layer: dampened gypsum boards (air gap filled with dense Rockwool)
2) second layer: metal-plate Helmholtz bass resonators (tuned to the ~32Hz and ~43Hz axial room nodes)
3) third layer: P23 quadratic diffusors with the deepest bass diffusion bandwidth, for the front wall
4) shallower P17 quadratic diffusers for the sidewalls, but believe it's best not to have any closer than 2.0m from the listening position, due to undesirable comb filtering artifacts (which become inaudible at greater distances)
Q1) should the metal-plate Helmholtz bass resonators be mounted directly onto the structural walls, instead on the sheetrock?
Q2) should the sheetrock be mounted over the Helmholtz bass absorbers, and the larger air gap filled with dense Rockwool?
Q3) can the room be over-damped with too many Helmholtz bass absorbers, and if so, what might be the optimal % fill factor?
Q4) position/quantity of the sidewall P17 quadratic diffusers - one pair, two pairs, three pairs, or keep all four pairs?
Q5) quadratic 1D diffusers, BAD half-cylinders, or absorption (via dense open-cell foam) at the first reflection zones?
Q6) with so many metal-plate Helmholtz resonators, are corner bass traps needed or can they be disregarded to save space?
Ceiling:
1) first layer: metal plate Helmholtz bass resonators (tuned to the ~49Hz axial room node)
2) second layer: canopy, either P13 quadratic diffusers in a 2D configuration, or /lighter shallower 2D quadratic diffusers
Q1) ~3.5m high false ceilings imply 7.0m roundtrip; as acoustic energy returning from the ceiling are relatively weak and potentially inaudible, are ceiling diffusers a must-have?
Q2) deep, heavy, large P13 1D quadratic diffusers arranged into 2D configuration (90* orientation, alternating)?
Q3) alternatively, "2D diffusers" with 10x10 or 12x12 small squares?
Floor:
1) base finish: self-leveling epoxy + paint finish
Q1) option 1: vinyl tile flooring (covering the base finish)
Q2) option 2: polyester/nylon carpet "squares" (covering the base finish)
Q3) plywood over concrete, then engineered wood parquet
Carpet/rug:
Q1) option 1: thin rug (Persian-style?) placed on the rectangle between the loudspeakers and listening position
Q2) option 2: decorative rug made of open-weave fabric/yarn, to discourage over-absorption
Q3) option 3: low-pile carpet (IKEA type, typically available in 2m x 3m size), to encourage absorption of mids/highs
Please chime in with your experiences and suggestions for those prefixed with a "Q" !
Four walls:
1) first layer: dampened gypsum boards (air gap filled with dense Rockwool)
2) second layer: metal-plate Helmholtz bass resonators (tuned to the ~32Hz and ~43Hz axial room nodes)
3) third layer: P23 quadratic diffusors with the deepest bass diffusion bandwidth, for the front wall
4) shallower P17 quadratic diffusers for the sidewalls, but believe it's best not to have any closer than 2.0m from the listening position, due to undesirable comb filtering artifacts (which become inaudible at greater distances)
Q1) should the metal-plate Helmholtz bass resonators be mounted directly onto the structural walls, instead on the sheetrock?
Q2) should the sheetrock be mounted over the Helmholtz bass absorbers, and the larger air gap filled with dense Rockwool?
Q3) can the room be over-damped with too many Helmholtz bass absorbers, and if so, what might be the optimal % fill factor?
Q4) position/quantity of the sidewall P17 quadratic diffusers - one pair, two pairs, three pairs, or keep all four pairs?
Q5) quadratic 1D diffusers, BAD half-cylinders, or absorption (via dense open-cell foam) at the first reflection zones?
Q6) with so many metal-plate Helmholtz resonators, are corner bass traps needed or can they be disregarded to save space?
Ceiling:
1) first layer: metal plate Helmholtz bass resonators (tuned to the ~49Hz axial room node)
2) second layer: canopy, either P13 quadratic diffusers in a 2D configuration, or /lighter shallower 2D quadratic diffusers
Q1) ~3.5m high false ceilings imply 7.0m roundtrip; as acoustic energy returning from the ceiling are relatively weak and potentially inaudible, are ceiling diffusers a must-have?
Q2) deep, heavy, large P13 1D quadratic diffusers arranged into 2D configuration (90* orientation, alternating)?
Q3) alternatively, "2D diffusers" with 10x10 or 12x12 small squares?
Floor:
1) base finish: self-leveling epoxy + paint finish
Q1) option 1: vinyl tile flooring (covering the base finish)
Q2) option 2: polyester/nylon carpet "squares" (covering the base finish)
Q3) plywood over concrete, then engineered wood parquet
Carpet/rug:
Q1) option 1: thin rug (Persian-style?) placed on the rectangle between the loudspeakers and listening position
Q2) option 2: decorative rug made of open-weave fabric/yarn, to discourage over-absorption
Q3) option 3: low-pile carpet (IKEA type, typically available in 2m x 3m size), to encourage absorption of mids/highs
Last edited:
Hi Quad,
Congratulations on your amazing room development.
I am curious what crossover network configurations do you like? Do you like smooth sloping baffles?
As stated in my posts #28, #38 and #43 - my loudspeaker preferences are NOT for 1st order [correction, of post #38's mention of 2nd order] crossover networks, and baffles with sharp corners/edges resembling Lego-bricks.
Last edited:
Pretty comprehensive system and well documented.
You and I probably dont agree on qrd so I will leave that out
WALLS
I think the sheetrock layer introduces a membrane absorber of unpredictable effect .. Is it for attenuating structure born sound .. if not I would delete it and hard plaster ( or stain ) concrete walls. You can then add and subtract the plate absorbers to get the desired result .. flexibility is important with bass and there is nothing worse than overdamping. Additionaly you may need some broadband absorbtion just to calm upper frequency that you can install on non critical locations as needed. You can have untreated concrete and all could be covered with a stetch fabric to visually calm the room when finalised ( I think @Bobvin did this)
BTW I would use bass absorbers in all 4 corners and all four wall/ceiling junctions for bass absorbtion .. its where energy is greatest
CEILINGS
I have a 3.4m ceiling and added ceiling baffles to stop reflections and it was a big improvement so I would say yes .. it calmed the room in a nice way
FLOOR
This is potentially the worst first refection .. there is an argument that are we used to this and our brain filters it easily but folks have reported improvements .. I would favour a persian rug over and underlay in the zone between speakers and chair and the rest reflective to keep that nice reverbarent tail while calming high fr reflections.
Anything else will not make much difference.
You can always add more rugs if too bright
You could always build a pit and grid floor over at reflection point
Regardless its bound to be a brilliant room.. looking forward to its progress
Cheers
Phil
Last edited:
Yes, but I am wondering what your preferences are for - just curious.
OK I re-read post 28, there is some more info there. I think you meant 1st order along with baffle alignment.
Last edited:
Dcathro, you're correct! Thanks for alerting me. Back in post #38, I meant 1st order (6dB/octave), and not 2nd order (12dB/octave). I've made a note of it in post #49. I prefer not to belabor this as it's simply my personal preference, but I believe that essentially, the trade-off is that 1st order designs limit/narrow the optimal listening position, as augmentations/cancellations (between drivers overlapping/reproducing a broader range of frequencies) are magnified. I've experienced this through Thiel 2 loudspeakers (see Stereophile review link below) which I owned for 3+ years in the 1980's, as well as more recently, brand "V" loudspeakers at trade shows. I prefer steeper-slope crossovers, despite the purported non-linearities in the time domain.
Thiel CS2 loudspeaker
The latest edition of Audio's annual equipment directory lists 238 speaker manufacturers. At best I can claim to have heard one product from 10–15% of the manufacturers on this list, and the top of the current product line from a far smaller percentage.
www.stereophile.com
Last edited:
It's always good to know what you like and want! I wish you good luck on your journey! Will be looking out for the progress on your room.
My belief is that the room is by far the most important component!
My belief is that the room is by far the most important component!
Hi Quad, I didn't mean to put you on the defensive. No crossover topology is perfect, they all have their strengths and weaknesses and it depends on what you are sensitive to. I agree that 1st order crossovers are very difficult to implement and have some big limitations:
- inability to play loud;
- difficulty masking any driver resonances or frequency anomalies;
- much more difficult to get a flat response;
- need drivers that play way beyond their passband
- less than optimal polar response due to drivers being audible beyond their passband.
FRONT WALL (and front sidewall) ACOUSTIC TREATMENT: QRDs
The objectives of enhancing and preserving soundstage articulation and expression in two-channel stereo reproduction are fundamental goals for audiophiles. Decorrelating the spectral and temporal characteristics of the primary signal reflections is the key to reaching this objective. Essentially, the challenge is to create an ambient sound field which is highly uncorrelated to the primary signal, uniformly scrambled in directional, polar, spectral and temporal response, such that no direct "mirror image" hot spot reflections are discernible, delivering the aural illusion that the wall boundary has extended and retreated into the distance beyond its actual physical location.
Rather than aggressively absorbing precious mid/high frequency energy through foam, building insulation, thick carpets, upholstery, etc. on the front hemispherical wave-launch strike boundary strikes which undesirably "suck the life out" of the critical energy band necessary for ambient cues and soundstage envelopment, QRDs are engineered to create a decorrelated, incoherent cloud, while retaining most of the "lively" energy.
As my WBF moniker implies, I'm a huge fan of quadratic diffusers ("QRD"s). Through comb filtering, they redirect and scramble acoustic energy (laterally in 1D QRDs, and hemispherically in 2D QRDs) so the ear/brain combo can more readily distinguish and resolve the primary wave launch from the secondary and tertiary reflections. QRDs will selectively absorb certain frequencies pertaining to the quarter wavelength rule of the depth and volume of the zig-zagging wells, but by using a prime-number sequence (the larger the better!) of non-repeating, non-periodic sequences, the potential for hearing coherent "mirror image" direct in-phase reflections (the bane of binary-amplitude and cylindrical/spherical diffusers) is mitigated.
QRDs can be custom-built, sized with different well widths (defines upper frequency cutoff frequency), and most importantly, prime number periodicity and well depth (determines lower frequency cutoff frequency). Understanding which QRD (P13, P17, P23?) is most optimal for different locations within the listening room is important.
Theoretically, the lower the bandwidth (and the broader/more uniform the dispersion/scattering the better), but there are some real-world constraints and limitations such as viscous losses in cavity resonances (lumpy attenuation corresponding to the well resonant frequencies), expense/size, material flex, placement geometry/location affecting angle of incidences, and most importantly, adequately provisioning for the minimum listening distance (two times, but ideally, three times the wavelength of the lowest frequency diffused by the QRD).
The two first-reflection points on the two sidewalls deserves special attention, as they have the potential to completely destroy the carefully crafted soundstage, by being:
- too loud (not attenuated enough)
- too correlated (not scrambled enough)
Together with perfect L/R symmetry, I’d like the sidewall reflections to meet four of the following psychoacoustical criteria:
1) placed outside of the minimum prescribed distance (as per the 2x-3x lowest frequency wavelength rule)
2) Haas effect: that the sidewall signals arrive inside of 12-20ms relative to the primary signal
3) precedence effect: ensure that the sidewall signals are attenuated by at least 6dB (in the presence region) relative to the primary signal
4) that the reflected energy is delivered from a direction (relative to the midline to the center of the front wall) 60* to the L, and 60* from the R, in a 120* spread (see my post #30 on the thread below). The L and R speakers will be placed 30* to the L, and 30* to the R, in a 60* spread, assuming an equilateral triangle placement.
www.whatsbestforum.com
I suspect that raytracing simulations will recommend not having any P23 QRDs in close proximity to the listening position. Perhaps I will have P23 QRDs on the sidewalls corners (with its width occupying 1.26m of sidewall space), followed by P17s (its width occupying an additional 0.905m of sidewall space), in a one-step gradient? Together, the two different QRDs will cover 2.165m of horizontal/linear sidewall space on both sidewalls, creating a "C" shape of wall modules (when viewed from above). As I plan to distance my speakers ~2.5m from the front wall slab, the sidewall QRDs will not protrude into the space directly to the L and R space of the speakers.
AFMG Reflex is the software which most appropriately models and quantifies the characteristics of various QRDs which I'll be using for my future dedicated basement listening room.
www.mediterraneanacoustics.com/afmg-reflex---for-the-design-and-application-of-diffusers.html
"AFMG Reflex is a two-dimensional acoustics simulation software to model the reflection, diffusion, and scattering of a sound wave incident upon a defined geometrical structure.
Within AFMG Reflex the shape of the reflective surface is entered as a two-dimensional cross-section as would be seen from overhead looking down at the shape. Reflex assumes that this cross-sectional geometry extends infinitely in the third dimension, i.e. into and out of the computer monitor screen. The surface defined by this geometry is also assumed to be perfectly rigid. This means the surface is 100% reflective and does not in any way absorb sound or allow sound to be transmitted through it.
These assumptions are required for the mathematical approach that Reflex is using. It is based on the Boundary Element Method (BEM) to calculate the reflection, diffusion, and scattering properties of the surface. The reflective properties are displayed as a polar response graph for any angle of incidence and frequency of a sound wave. The scattering and diffusion coefficients are displayed as frequency response graphs.
The calculation of the scattering coefficients is based on the work of Mommertz, Vorländer, and ISO 17497-1. The calculation of the diffusion coefficients is according to Cox and D’Antonio and to the ISO 17497-2 standard which is currently in preparation."
The objectives of enhancing and preserving soundstage articulation and expression in two-channel stereo reproduction are fundamental goals for audiophiles. Decorrelating the spectral and temporal characteristics of the primary signal reflections is the key to reaching this objective. Essentially, the challenge is to create an ambient sound field which is highly uncorrelated to the primary signal, uniformly scrambled in directional, polar, spectral and temporal response, such that no direct "mirror image" hot spot reflections are discernible, delivering the aural illusion that the wall boundary has extended and retreated into the distance beyond its actual physical location.
Rather than aggressively absorbing precious mid/high frequency energy through foam, building insulation, thick carpets, upholstery, etc. on the front hemispherical wave-launch strike boundary strikes which undesirably "suck the life out" of the critical energy band necessary for ambient cues and soundstage envelopment, QRDs are engineered to create a decorrelated, incoherent cloud, while retaining most of the "lively" energy.
As my WBF moniker implies, I'm a huge fan of quadratic diffusers ("QRD"s). Through comb filtering, they redirect and scramble acoustic energy (laterally in 1D QRDs, and hemispherically in 2D QRDs) so the ear/brain combo can more readily distinguish and resolve the primary wave launch from the secondary and tertiary reflections. QRDs will selectively absorb certain frequencies pertaining to the quarter wavelength rule of the depth and volume of the zig-zagging wells, but by using a prime-number sequence (the larger the better!) of non-repeating, non-periodic sequences, the potential for hearing coherent "mirror image" direct in-phase reflections (the bane of binary-amplitude and cylindrical/spherical diffusers) is mitigated.
QRDs can be custom-built, sized with different well widths (defines upper frequency cutoff frequency), and most importantly, prime number periodicity and well depth (determines lower frequency cutoff frequency). Understanding which QRD (P13, P17, P23?) is most optimal for different locations within the listening room is important.
Theoretically, the lower the bandwidth (and the broader/more uniform the dispersion/scattering the better), but there are some real-world constraints and limitations such as viscous losses in cavity resonances (lumpy attenuation corresponding to the well resonant frequencies), expense/size, material flex, placement geometry/location affecting angle of incidences, and most importantly, adequately provisioning for the minimum listening distance (two times, but ideally, three times the wavelength of the lowest frequency diffused by the QRD).
The two first-reflection points on the two sidewalls deserves special attention, as they have the potential to completely destroy the carefully crafted soundstage, by being:
- too loud (not attenuated enough)
- too correlated (not scrambled enough)
Together with perfect L/R symmetry, I’d like the sidewall reflections to meet four of the following psychoacoustical criteria:
1) placed outside of the minimum prescribed distance (as per the 2x-3x lowest frequency wavelength rule)
2) Haas effect: that the sidewall signals arrive inside of 12-20ms relative to the primary signal
3) precedence effect: ensure that the sidewall signals are attenuated by at least 6dB (in the presence region) relative to the primary signal
4) that the reflected energy is delivered from a direction (relative to the midline to the center of the front wall) 60* to the L, and 60* from the R, in a 120* spread (see my post #30 on the thread below). The L and R speakers will be placed 30* to the L, and 30* to the R, in a 60* spread, assuming an equilateral triangle placement.
My dedicated audio room build - QuadDiffusor's Big Dig
There is of course another room acoustical technology that provides a third approach to dealing with low frequencies. Developed for and in use by well known film/recording/mixing studios (links to examples below), it is based on very high numbers per sf of non-parallel surfaces (DHDI ZR...
www.whatsbestforum.com
I suspect that raytracing simulations will recommend not having any P23 QRDs in close proximity to the listening position. Perhaps I will have P23 QRDs on the sidewalls corners (with its width occupying 1.26m of sidewall space), followed by P17s (its width occupying an additional 0.905m of sidewall space), in a one-step gradient? Together, the two different QRDs will cover 2.165m of horizontal/linear sidewall space on both sidewalls, creating a "C" shape of wall modules (when viewed from above). As I plan to distance my speakers ~2.5m from the front wall slab, the sidewall QRDs will not protrude into the space directly to the L and R space of the speakers.
AFMG Reflex is the software which most appropriately models and quantifies the characteristics of various QRDs which I'll be using for my future dedicated basement listening room.
www.mediterraneanacoustics.com/afmg-reflex---for-the-design-and-application-of-diffusers.html
"AFMG Reflex is a two-dimensional acoustics simulation software to model the reflection, diffusion, and scattering of a sound wave incident upon a defined geometrical structure.
Within AFMG Reflex the shape of the reflective surface is entered as a two-dimensional cross-section as would be seen from overhead looking down at the shape. Reflex assumes that this cross-sectional geometry extends infinitely in the third dimension, i.e. into and out of the computer monitor screen. The surface defined by this geometry is also assumed to be perfectly rigid. This means the surface is 100% reflective and does not in any way absorb sound or allow sound to be transmitted through it.
These assumptions are required for the mathematical approach that Reflex is using. It is based on the Boundary Element Method (BEM) to calculate the reflection, diffusion, and scattering properties of the surface. The reflective properties are displayed as a polar response graph for any angle of incidence and frequency of a sound wave. The scattering and diffusion coefficients are displayed as frequency response graphs.
The calculation of the scattering coefficients is based on the work of Mommertz, Vorländer, and ISO 17497-1. The calculation of the diffusion coefficients is according to Cox and D’Antonio and to the ISO 17497-2 standard which is currently in preparation."
Last edited:
Very interesting.. have you tried it out.
Nice if pretty easy to get your head around ... quite expensive but may well be worth it.
At a simpler level I see the amroc guys are developing a more " pro version " for sale .. will also be interesting to see how deep it dives
Do you see yourself doing different diffusers as the location changes
Phil
Sharing below simulation results for P13, P17, and P23 QRDs, provided by my acoustic consultant which I hired. The AFMG Reflex program has tremendous flexibility and accepts multiple parameters to model a variety of geometric diffusion modules, but for this example I'm keeping things simple.
Inputs:
- primary signal directed from a 45 degree angle (as opposed to 0 degrees, or head-on) which better simulates reality
Outputs:
- spatial response curves of 500Hz, 1kHz, and 2kHz (overlaid on a hemispheric polar graph)
- broader 100Hz to 4kHz range for the calculation of scattering, normalized diffusion, and diffusion coefficients
Inputs:
- primary signal directed from a 45 degree angle (as opposed to 0 degrees, or head-on) which better simulates reality
Outputs:
- spatial response curves of 500Hz, 1kHz, and 2kHz (overlaid on a hemispheric polar graph)
- broader 100Hz to 4kHz range for the calculation of scattering, normalized diffusion, and diffusion coefficients
Last edited:
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Co-Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |