here is a conceptual sketch ( click thumbnail to enlarge ):
it's a cross section view from the top. the system is 5-way with each frequency band color coded.
overall array height is about 90 inches consisting of two 45 inch tall vertical sections for efficient use of plywood.
furthermore each vertical section also breaks up into 4 parts ( front, back, left and right ) as you can see in the diagram for ease of transportation.
so we're looking at a total of 8 parts of about 100 lbs each which can be moved easily by two people.
starting from the top of the image, which is the front side:
the red orange square dot is Fountek Ribbon: 3.5 khz to 40 khz
the magneta line is Radian 10" Planar: 500 hz - 3.5 khz
the two green midbasses are 5" Faital 5PR160: 150 hz - 500 hz
the yellow woofer is 12" Eighteen Sound 12NTLW3500: 60 hz - 150 hz ( tuned to 50 hz )
the blue subwoofer is 18" Eighteen Sound 18NTLW5000: 16 hz - 60 hz ( tuned to 16 hz )
total number of drivers per speaker:
4 X subs
4 X woofers
12 X midbass
6 X planar
10 X ribbon
midbass drivers are STAGGERED so that it's the same number of drivers as there would be in a single row but they alternate sides like a checkerboard pattern to save money by not having to use more drivers. well actually they're just slightly tighter - 5" drivers every 4 inches rather than every 5 inches.
the vents for the subs will be EXTERNAL ( bolt on ) and run vertically almost the entire height of the array for minimum port compression.
basically all of the drivers used are 95 db / watt efficient and all either reviewed favorably by Vance Dickason at his "Test Bench" or had a member of the driver family favorably reviewed there.
for example Vance reviewed the 6" midbass and i am using the 5" version that has the same motor and suspension but smaller cone for less beaming.
i did the math and the drivers would cost about $25,000 ( for a pair of these speakers ) which i am not happy about, but i don't see a way to do this any cheaper.
do you ?
thoughts ?
it's a cross section view from the top. the system is 5-way with each frequency band color coded.
overall array height is about 90 inches consisting of two 45 inch tall vertical sections for efficient use of plywood.
furthermore each vertical section also breaks up into 4 parts ( front, back, left and right ) as you can see in the diagram for ease of transportation.
so we're looking at a total of 8 parts of about 100 lbs each which can be moved easily by two people.
starting from the top of the image, which is the front side:
the red orange square dot is Fountek Ribbon: 3.5 khz to 40 khz
the magneta line is Radian 10" Planar: 500 hz - 3.5 khz
the two green midbasses are 5" Faital 5PR160: 150 hz - 500 hz
the yellow woofer is 12" Eighteen Sound 12NTLW3500: 60 hz - 150 hz ( tuned to 50 hz )
the blue subwoofer is 18" Eighteen Sound 18NTLW5000: 16 hz - 60 hz ( tuned to 16 hz )
total number of drivers per speaker:
4 X subs
4 X woofers
12 X midbass
6 X planar
10 X ribbon
midbass drivers are STAGGERED so that it's the same number of drivers as there would be in a single row but they alternate sides like a checkerboard pattern to save money by not having to use more drivers. well actually they're just slightly tighter - 5" drivers every 4 inches rather than every 5 inches.
the vents for the subs will be EXTERNAL ( bolt on ) and run vertically almost the entire height of the array for minimum port compression.
basically all of the drivers used are 95 db / watt efficient and all either reviewed favorably by Vance Dickason at his "Test Bench" or had a member of the driver family favorably reviewed there.
for example Vance reviewed the 6" midbass and i am using the 5" version that has the same motor and suspension but smaller cone for less beaming.
i did the math and the drivers would cost about $25,000 ( for a pair of these speakers ) which i am not happy about, but i don't see a way to do this any cheaper.
do you ?
thoughts ?
i could sketch front / side / rear views but i think that can just be explained in words.
each side ( left and right ) will have two subs - one in the top section and one in the bottom.
the rear will have four woofers - two in the top section and two in the bottom.
and the front will be like a normal array except midbass drivers will be in a zig-zag / checkerboard arrangement to keep the number of drivers lower as they're $130 each.
the fountek ribbon will just be hanging out there similar to this design from JBL:
https://adn.harmanpro.com/product_a...728928178/pd764i-front-nogrill_z_x_large.webp
except it won't be a horn loaded compression driver but a ribbon. the ribbons of course will form a continuous array, just like the planars.
the goal is to have a do-it-all system that has audiophile SQ, rock concert SPL and Home Theater bass extension all while having a small enough foot print to fit in literally any room.
the other goal was to have integrated subwoofers with 60 hz crossover frequency so the system can be run as self-contained in smaller rooms or with additional subs in larger rooms.
when i say end-game i mean it. if i build this it will be the last speaker i build. this is why it must do everything and work in every space.
and because proper subwoofers are so expensive and take up so much space i wanted to make sure i wasn't wasting any bass by building something that goes to say 30 hz only to then cut the bottom octave off to cross to subs at 60 hz and have to find additional space for those subs. instead, whether the system is run with or without additional subs 100% of the bass capability of the system is used 100% of the time with no waste.
16 hz tuning is chosen because 16 hz is the lowest note on a pipe organ.
each side ( left and right ) will have two subs - one in the top section and one in the bottom.
the rear will have four woofers - two in the top section and two in the bottom.
and the front will be like a normal array except midbass drivers will be in a zig-zag / checkerboard arrangement to keep the number of drivers lower as they're $130 each.
the fountek ribbon will just be hanging out there similar to this design from JBL:
https://adn.harmanpro.com/product_a...728928178/pd764i-front-nogrill_z_x_large.webp
except it won't be a horn loaded compression driver but a ribbon. the ribbons of course will form a continuous array, just like the planars.
the goal is to have a do-it-all system that has audiophile SQ, rock concert SPL and Home Theater bass extension all while having a small enough foot print to fit in literally any room.
the other goal was to have integrated subwoofers with 60 hz crossover frequency so the system can be run as self-contained in smaller rooms or with additional subs in larger rooms.
when i say end-game i mean it. if i build this it will be the last speaker i build. this is why it must do everything and work in every space.
and because proper subwoofers are so expensive and take up so much space i wanted to make sure i wasn't wasting any bass by building something that goes to say 30 hz only to then cut the bottom octave off to cross to subs at 60 hz and have to find additional space for those subs. instead, whether the system is run with or without additional subs 100% of the bass capability of the system is used 100% of the time with no waste.
16 hz tuning is chosen because 16 hz is the lowest note on a pipe organ.
i should also note that with exception of 3.5 khz crossover to fountek ribbon ( which should be brick wall FIR ) none of the other crossovers will be crossovers per se ...
for example there could be an octave or so of overlap between midbass and planar because they are co-axial and as long as phase aligned they can operate on the same frequencies ... in fact by using gentle sloping shelves and digital filters without phase shift this can be used to control directivity precisely ...
in other words instead of having a sharp step in directivity when the system goes from wider midbasses to narrower planar this transition can be spread over more than an octave for a consistent bandwidth ... i think this is the trick that JBL is using in their VTX arrays that have RULER FLAT horizontal directivity at precisely 100 degrees at all frequencies ... it is only possible with digital filter trickery that gradually mixes in signals from different drivers rather than having an abrupt crossover ... it is also only possibly with HORIZONTALLY SYMMETRICAL ARRAY.
so yes the system was inspired by this:
https://jblpro.com/en-US/products/vtx-v25-ii
you can see the seemingly impossible performance here:
https://jblpro.com/en-US/site_elements/jbl-professional-vtx-v25-ii-spec-sheet
but it just comes down to geometry plus DSP.
my system is intended to be a home audiophile version of that JBL VTX. it will be able to go an octave higher in the highs, an octave lower in the lows, have an order of magnitude lower midrange distortion and cost 5% as much.
magic ? not really. home audio is simply a less demanding application than tour sound. in tour sound they have to maximize SPL per pound of system weight because they literally hang the system from the ceiling. i don't need to worry about this. i can build a heavier system with less output and in the process have it sound better and cost less.
but i will still use the same science because math is universal - it applies equally at home as it does in prosound.
for example there could be an octave or so of overlap between midbass and planar because they are co-axial and as long as phase aligned they can operate on the same frequencies ... in fact by using gentle sloping shelves and digital filters without phase shift this can be used to control directivity precisely ...
in other words instead of having a sharp step in directivity when the system goes from wider midbasses to narrower planar this transition can be spread over more than an octave for a consistent bandwidth ... i think this is the trick that JBL is using in their VTX arrays that have RULER FLAT horizontal directivity at precisely 100 degrees at all frequencies ... it is only possible with digital filter trickery that gradually mixes in signals from different drivers rather than having an abrupt crossover ... it is also only possibly with HORIZONTALLY SYMMETRICAL ARRAY.
so yes the system was inspired by this:
https://jblpro.com/en-US/products/vtx-v25-ii
you can see the seemingly impossible performance here:
https://jblpro.com/en-US/site_elements/jbl-professional-vtx-v25-ii-spec-sheet
but it just comes down to geometry plus DSP.
my system is intended to be a home audiophile version of that JBL VTX. it will be able to go an octave higher in the highs, an octave lower in the lows, have an order of magnitude lower midrange distortion and cost 5% as much.
magic ? not really. home audio is simply a less demanding application than tour sound. in tour sound they have to maximize SPL per pound of system weight because they literally hang the system from the ceiling. i don't need to worry about this. i can build a heavier system with less output and in the process have it sound better and cost less.
but i will still use the same science because math is universal - it applies equally at home as it does in prosound.
as for why i have drivers on all 4 sides whereas JBL only has them on the front it's like this ...
firstly prosound speakers never have anything on sides or back because they must be stackable in any way desired and by having all drivers and ports only on the front that ensures that ...
secondly prosound arrays must be curve-able to match directivity to the venue. you can only curve an array of elements that are shallow and wide, which again dictates that drivers must spread to the sides rather than to the back.
thirdly when an array is suspended over a crowd it can't take advantage of REAR WALL that all the rooms have. the reason i put woofer in the back is so it can be shoved against the wall and use 6db gain from that wall reflection instead of having a dip in response.
and subs on the sides is simply to keep the system narrow so it can fit in a small room and still allow space to put a big TV between two speakers.
firstly prosound speakers never have anything on sides or back because they must be stackable in any way desired and by having all drivers and ports only on the front that ensures that ...
secondly prosound arrays must be curve-able to match directivity to the venue. you can only curve an array of elements that are shallow and wide, which again dictates that drivers must spread to the sides rather than to the back.
thirdly when an array is suspended over a crowd it can't take advantage of REAR WALL that all the rooms have. the reason i put woofer in the back is so it can be shoved against the wall and use 6db gain from that wall reflection instead of having a dip in response.
and subs on the sides is simply to keep the system narrow so it can fit in a small room and still allow space to put a big TV between two speakers.
going to replace ( in the design ) four 18" subs per speaker with two 21" subs ( from the same NTLW5000 series ) ... this will cut the cost of sub channel in half and bring it in line with the cost of other frequency bands.
for all the other frequency bands the cost of drivers was about $3,500 or so but for the sub channel it was almost $10,000 ... now it is half that.
of course now the speaker won't have enough sub bass to match the output of other frequency band but this is better handled with subs elsewhere in the room because because of both standing wave mode issue and also the fact that cramming four subs into one speaker would limit the enclosure volume to where it would be very difficult to make deep bass anyway.
it was tempting to have a fully self contained system for a clean look but the thought of spending $10,000 + amps on the subs and then losing all the output to some room mode was too frightening.
i need the freedom to place at least half of the subs in locations other than where the speakers are to optimize the room modes.
i also need more enclosure volume for the subs and 21" drivers are a better value than 18" ones for deep bass and there just wasn't enough volume for them in the original design.
for all the other frequency bands the cost of drivers was about $3,500 or so but for the sub channel it was almost $10,000 ... now it is half that.
of course now the speaker won't have enough sub bass to match the output of other frequency band but this is better handled with subs elsewhere in the room because because of both standing wave mode issue and also the fact that cramming four subs into one speaker would limit the enclosure volume to where it would be very difficult to make deep bass anyway.
it was tempting to have a fully self contained system for a clean look but the thought of spending $10,000 + amps on the subs and then losing all the output to some room mode was too frightening.
i need the freedom to place at least half of the subs in locations other than where the speakers are to optimize the room modes.
i also need more enclosure volume for the subs and 21" drivers are a better value than 18" ones for deep bass and there just wasn't enough volume for them in the original design.
That's a lot of words, but it seems you haven't put an SPL target or listening distance in there anywhere. Vague references like "rock concert SPL" or "home theater bass extension" don't give enough information to judge whether what you are proposing is overkill or reasonable. Calculating the displacement limited output of the array elements would be a good place to start. This will mostly apply to midrange and below, since a large array of tweeters or upper mid drivers is likely to deafen you before it runs out of steam in a home audio scenario.
Something of this complexity would optimally be modeled. Your descriptions lead me to believe you have potential spacing issues in the array. Without physical measurements of all of the elements and the crossover slopes, it's hard to judge how much of a problem those will be. The Danley Unity and Synergy designs are worth looking at for some ideas about cross points and driver spacing. While DSP can be used to do interesting things and correct nearly anything for a single listening position, it cannot fix poor driver layout across multiple angles (barring very large and complex arrays that are far beyond what's being discussed here). It's generally considered good form to do as much as you can with physical design before resorting to DSP. Assuming DSP is going to save your bacon can back you into an unpleasant corner. For good imaging and soundstaging, it's typical to shoot for smooth off-axis behavior. Similarly, good power response tends to produce speakers that sound good in multiple types and sizes of rooms.
Something of this complexity would optimally be modeled. Your descriptions lead me to believe you have potential spacing issues in the array. Without physical measurements of all of the elements and the crossover slopes, it's hard to judge how much of a problem those will be. The Danley Unity and Synergy designs are worth looking at for some ideas about cross points and driver spacing. While DSP can be used to do interesting things and correct nearly anything for a single listening position, it cannot fix poor driver layout across multiple angles (barring very large and complex arrays that are far beyond what's being discussed here). It's generally considered good form to do as much as you can with physical design before resorting to DSP. Assuming DSP is going to save your bacon can back you into an unpleasant corner. For good imaging and soundstaging, it's typical to shoot for smooth off-axis behavior. Similarly, good power response tends to produce speakers that sound good in multiple types and sizes of rooms.
Last edited:
yes and sadly i have again redesigned the bass section while i was in bed last night so more words to come ...
so i have been modeling systems like this in HornResp for years and have a general idea about output levels from the woofer sections that can be expected but you are correct i haven't done the math for the ribbons for two reasons:
1 - it is a more complicated math because not all elements will add in phase.
2 - i have always tried to avoid ribbons because i felt they were overpriced until i did the math and realized most of system cost will be in subwoofers and their amps anyway, even in a ribbon array.
the last time i posted an "end game" design here ( in my previous internet life ) years ago it was based on a BMS compression midrange as opposed to Radian Planar array. i have evolved that design as well:
in the image above the speaker stops at 12" woofers ( orange ). the subs are external so not shown, though they would be the same as in current design ( 18 Sound NTLW 5000 series ). total driver cost would be less and output would still be in the 130+ decibel range but it would be an awkward system to build due to all the odd angles and the sweet spot in the room from where it could be appreciated would be small compared to an array, because the low frequency section points up and the midrange section points down so you could only be in the sweet spot at one distance from the speaker and not closer or further.
ultimately i realized that the cost saving of using a single $1,000 compression midrange versus $2,000 array of 6 planar mids isn't enough to justify the complexity of building that horn system nor all the compromises with sound quality of compression drivers or every frequency band having different directivity etc. the array simply makes more sense but you're right - i haven't crunched the numbers on the mid and high sections. i am not worried about having too much SPL though - i am more worried about not having enough.
so the low frequency parts of the system simply migrated from my horn based design to the array based design and the ribbons then simply fill out the space to reach the necessary height of the array for optimum coverage whether sitting down or standing up.
the cone midrange in the array system was this:
https://www.eighteensound.it/en/products/lf-driver/8-0/8/8m400f
chosen for its ultra high output with 100 db efficiency and 300W power handling. by contrast in the array system the cone midrange is chosen for its small size ( 5" in array versus 8" in horn system ). so the two systems diverge from cone midrange and up in frequency but the same down low.
haven't done any math on the array system but i did do all the math for the horn system, of which this is an evolution.
this is why this system is slightly half baked and keeps evolving because it's really an adaptation of the old horn system to ribbon arrays and i'm still shaking out some of the details. like last night i realized i don't need 12" woofers and looking at the design of the old horn system i see where the 12" woofers originally came from - they were needed in the original system but not in the current one, so i will eliminate them.
Last edited:
actually "rock concert SPL" is anything but vague. it is precisely defined as 100 db A-Weighted with 1 second average. this is a legal definition.
this is why everybody is using flown J-Arrays at concerts now - because that's the only way to avoid exceeding that legal limit. with traditional setups of simply having speakers at the edge of the stage the front rows would blow way by 100 db limit, which is now illegal. so the speakers had to be moved to a safe distance from the crowd by suspending them in the air and also each box has its own DSP so the ones firing down into the crowd are run at lower output level than ones firing into the distance to reach the end of the crowd.
you may know that OSHA limit is 85 dba for 8-hour per day exposure and official theory is that hearing damage is cumulative based on total energy which means that higher dba is allowable for shorter periods of time. because nobody is at a rock concert for 8 hours a day, 5 days a week they allow 100 dba at concerts ( rock or otherwise ).
this is at 1 khz ( A-Weighting ) and 1-second average, which means that with about 10 db crest factor you can have 110 db peaks at 1 khz and with music response rising in the bass plus the house curve and still fitting into the A-Weighting curve
you can have 140 db in subwoofer range and it would still be legal and in fact Bass Nectar concert did register 130+ decibel bass and that's also what my system would put out as well.
i am not really flexible on number of ribbons used because they must fill the vertical space but i am flexible on number of subs which is why i had to do the actual math to see how many i need. i used HornResp simulation for that modeling both thermal ( watt ) and mechanical ( xmax + xmech ) limits.
well you're right it is somewhat vague, but i have done research on this as well.
it appears that music energy peaks at 50 hz:
and while i do not have any charts showing AVERAGE energy distribution in movies i have looked at charts for INDIVIDUAL movies at it seems to me the peak in movies is at around 30 hz ...
this is why JBL cinema subs are tuned to 24 hz while their PA subs are tuned closer to 35 hz ...
however the ELITE level JBL gear ( like the VTX array that inspired this ) are still tuned to 24 hz even though they are designed for music ... and Genelec Main monitors seem to be tuned to about 20 hz ...
so 24 hz tuning would be absolutely the highest i would consider if i want to call my system elite.
on other hand guys on AVS tune their subs to like 10 hz but it's a bit of a ******* contest over there for whose subs can go lower and they also use low-efficiency drivers, whereas the NTLW5000 drivers i propose here are basically clones of JBL 2269 which is what is used in the VTX and in fact although 2269 aren't officially sold they are available as replacement parts and cost EXACTLY the same as NTLW5000 and pretty much also have the same specs as well so i can see how JBL used these drivers in both cinema and prosound ( and they have used them in both ) and in all instances JBL tuned them to 24 hz ( both in prosound and cinema ) when using full-size enclosures ( they are tuned a bit higher in compact cabinets ).
however 2269 is 18" only while NTLW5000 is available both as 18" and 21" and it makes sense then that 21" can be tuned a bit lower ... how much lower ?
well obviously that is a bit of a subjective choice but i picked 16 hz like i said because this is the lowest pipe organ note, while still not too far from the 24 hz that JBL uses.
the other reason why i can tune lower than JBL's 24 hz is because their systems are for large venues ( or altogether outdoor use ) while i would benefit from room gain. this is how AVS guys can get away with 10 hz tuning, but like i said they use low-SPL high-excursion subs while i am using balanced all-rounder drivers that want to see higher tuning. so 16 hz seems like a good compromise between 10 hz used at AVS for low efficiency drivers in small rooms and 24 hz used by JBL for high efficiency drivers in open air.
i dunno i mean most of the summation would be out of phase and we aren't talking about prosound drivers here.
even though Radian LARPs as prosound it really isn't. i don't know anybody using those Planars in prosound. they are instead a high-end alternative to GRS planars used at home.
so basically i am using prosound drivers for the low frequencies, then switching over to sort of semi-pro for midrange ( radian ) and consumer ( fountek ) at the top. kind of an odd way to do things but from my experience i never complained about bass being too loud in the clubs nor about compression drivers not being fatigueing enough LOL so i think this actually may make sense.
strong bass is the strength of prosound drivers that would still apply at home because you want to feel the bass in your chest at home as much as you do in the club. on other hand compression drivers are meant to "cut through" the noise of the crowd and this isn't a problem you have at home. by contrast the detail of ribbons and planars is what you want.
i used to poo-poo everything Danley but i must admit that i may need to understand the science that goes into those systems because as far as spacing and so on the same would apply to my system.
also i always thought Unity and Synergy was the same thing - clearly i need more education in that area.
do you have links that explain the science behind it with actual formulas ? i have a vague concept of how it's supposed to work but don't know the math i would need to design my own.
agree that DSP can't fix everything and that you want good power response etc ... but ... if DSP will be available then you MUST design with DSP in mind.
NOT as a crutch, but rather to make sure that you don't miss any opportunities that DSP presents.
you don't want to design around artificial limitations that you imposed on yourself.
JBL SRX 835 is a good example of how NOT to do it. they designed a passive speaker and offered it as both passive and active to cover both markets. but as a result the active version only sounds as good as the passive version. all the speakers that were designed from the ground up as active DSP are better, because they weren't designed around physical limitations of having to horn load the mid to match the efficiency of woofer and tweeter for example, or having to make the two horns ( mid and hf ) of equal depths to time align the speakers physically.
DSP gives you more options to work with. And especially with symmetrical arrays it allows you to control directivity in ways otherwise impossible and that opportunity must not be passed up.
So can we boil most of that text down to: "At a 15-foot listening distance in a home I want 140 dB at 20 Hz and then a gradual decrease to 110 dB at 1 kHz and above"?
I don't get the merry-go-round of lengthy explanations about what concert level means from a legal standpoint today, then countering most of it anyway because it's A-weighted and bass limitations are likely given your descriptions. You may just be a person that likes a winding story. I realize that some think and design that way. But if that's not your bag, I think it's worth mentioning that circuitous discussions bog many people down and make it hard to get a feel for what you want and what you know. If you can say it in 25 words, spending a page trying to describe it obliquely is a lot to chew on. People here genuinely want to help others, and a lot of them have real-world industry experience to share if you make it easy for them. For many, their egos are mostly out of the equation as well, so there's no need to prove yourself with more words in case that's any part of it. Just talk economically, precisely, and directly about what you want and people will help you if they can. I hope none of that offended you, and I apologize if it did. I'm honestly trying to increase the chances that you'll get better feedback going forward.
If you're already accurately modeling the output limitations and overall design, that would be good to know also. It saves people from talking about things you've already addressed. Seeing some of those models for off-axis behavior would be useful. While your sketch gets the basic point across, there's just not enough there to say what's problematic. And if you have a good model that proves a concern is irrelevant, it circumvents that distraction.
One of the Danley patents:
US6411718B1, Sound reproduction employing unity summation aperture loudspeakers
https://patents.google.com/patent/US6411718B1/en
Thread discussing them:
www.diyaudio.com/community/threads/the-mysterious-danley-crossover.364891/
I don't get the merry-go-round of lengthy explanations about what concert level means from a legal standpoint today, then countering most of it anyway because it's A-weighted and bass limitations are likely given your descriptions. You may just be a person that likes a winding story. I realize that some think and design that way. But if that's not your bag, I think it's worth mentioning that circuitous discussions bog many people down and make it hard to get a feel for what you want and what you know. If you can say it in 25 words, spending a page trying to describe it obliquely is a lot to chew on. People here genuinely want to help others, and a lot of them have real-world industry experience to share if you make it easy for them. For many, their egos are mostly out of the equation as well, so there's no need to prove yourself with more words in case that's any part of it. Just talk economically, precisely, and directly about what you want and people will help you if they can. I hope none of that offended you, and I apologize if it did. I'm honestly trying to increase the chances that you'll get better feedback going forward.
If you're already accurately modeling the output limitations and overall design, that would be good to know also. It saves people from talking about things you've already addressed. Seeing some of those models for off-axis behavior would be useful. While your sketch gets the basic point across, there's just not enough there to say what's problematic. And if you have a good model that proves a concern is irrelevant, it circumvents that distraction.
One of the Danley patents:
US6411718B1, Sound reproduction employing unity summation aperture loudspeakers
https://patents.google.com/patent/US6411718B1/en
Thread discussing them:
www.diyaudio.com/community/threads/the-mysterious-danley-crossover.364891/
didn't see any formulas in that thread ...
not everything i post is a direct response to somebody's question ...
pretty close actually. but we need to consider peak and average levels separately because peak is mechanical displacement limited while average is thermally limited.
crest factor ( ratio of peak to average ) increases with frequency. it is around 6 db at the lowest for subwoofers playing DubStep and may be 20 db or more for the tweeters depending on type of music. A sinewave would have a crest factor of 0 db of course.
so DubStep bass is close to a sinewave - it's just distorted and shaped into various ups and downs but the crest factor is low because of that.
on other hand in treble the situation is reversed - we can have high but short peaks. it was ironically Ivan Beaver from Danely who told me over at ProSoundWeb that a nail clipper hits 120 decibels for like a millisecond - of course i immediately attacked him and said it was BS because in a real recording that would be compressed down and my threads were removed ... but now that i think of it he was right.
what happens is Compact Disc has 96 db of dynamic range and most of that is used by bass that is compressed the living **** out of, but for music to sound rich bass has to be higher level than mids or highs so referring to the chart i previously posted:
https://www.diyaudio.com/community/attachments/1740615617984-png.1427921/
in the treble there is about 15 db MORE room left for peaks than in bass, so crest factor in treble can be about 20db.
now the 100db A-Weighting applies to AVERAGE ( 1 second average to be exact ) and NOT peaks. so in practice we should design for peaks of over 110 db all the way up to 10 khz ! although we only need about 90 db average at 10 khz ( and 100 db @ 1 khz ).
as for bass sadly reaching desired SPL at 20 hz is only achievable in car audio due to relatively air tight and small cabin. the way most subwoofer drivers are made they begin to lose peak output capability below 50 hz or so, even when vented and that is why dance music energy content peaks at around 50 hz ( otherwise it would just keep increasing as we go down in frequency ).
so realistically we try to track the Equal Loudness curves ( specifically the 100 phon curve ):
as far as we can and then we accept that we are going to lose output below that frequency.
referring to above chart that would be 120 db average @ 30 hz or about 126 db peak at @ 30 hz and then roll off below that with SOME useful output at 16 hz but nowhere near full scale.
with full scale output at 30 hz you will still get full impact of movies because they put maximum energy around 30 hz. and you will also get full benefit of the sickest electronic music as well as typically even when content at lower frequencies is present it is mixed at a lower level, so as long as you have full scale output at 30 hz and SOME output at 16 hz ( you're not below port tuning ) you're pretty much covered.
also make that distance 12 feet. i am assuming 20 X 20 foot room which is the size of average 2 car garage. most homes i been touring that's about the maximum size of space that can be used for the speaker setup. and even if more space was available i would still be as close to the speakers as possible provided i can still fit a TV screen between them for music videos.
EDIT: also keep in mind that the SPLs described above should be reached at THD levels below audibility threshold which is about 1% from 1 khz and up but higher in the bass and practically irrelevant in subwoofers.
in practice this means the system should be capable of HIGHER output than target SPL and that buffer between max output capability ( at which point THD would be unacceptably high ) and actual playback levels is what keeps THD down to target levels. we probably want about 10 db or more buffer, but this depends on the type of driver ( how prone it is to distortion )
what's great about Radian planars is they are very low distortion which means we can push them almost to the point of melting and still have acceptable distortion. with 95 db / watt and 100W power handling that's 115 db @ 1 m but with arrays the SPL doesn't fall off that much over distance especially when considering room reflections so we basically hit the target there and i think we should also hit it everywhere else because arrays sort of naturally have a way of balancing output with frequency - this is something i wrote about on other forums.
basically Unity Horns are choked by the compression driver which must energize the entire room with a single voice coil and that may involve compromises such as using a larger VC that in turn results in less detailed highs etc. arrays avoid this pitfall which is why prosound is 100% arrays now.
there is a natural slope to the way energy falls off with frequency in music and arrays also have that natural roll off which is why i don't stress it too much - i know as long as it's an array of properly engineered efficient drivers i am already close to optimal and all the drivers in my design are 95 db / watt efficient or better.
Last edited:
correct. i have already redesigned it. i will post new design as separate thread.
