Others like the sound of AINOgradients too! Confirmed by getting the nomination for the "Best sound in show" at the Finnish DIY-audio event!
DIY-messut 2015 - Sivu 2
Pictures here (my room #6)https://www.dropbox.com/sh/fh4dswot0pa9tz2/AACY1z4k7ZDGxtTkQ3RhXRfxa?dl=0
The room had concrete walls and was quite small. I lowered bass level a bit and lifted mid-treble too. I must say that a part of my success was perhaps music selection - Dire Straits, Pink Floyd, Eva Cassidy, Phil Collins, Chris Jones, Iiro Rantala, Herbie Hancock etc. pieces with great dynamics.
DIY-messut 2015 - Sivu 2
Pictures here (my room #6)https://www.dropbox.com/sh/fh4dswot0pa9tz2/AACY1z4k7ZDGxtTkQ3RhXRfxa?dl=0
The room had concrete walls and was quite small. I lowered bass level a bit and lifted mid-treble too. I must say that a part of my success was perhaps music selection - Dire Straits, Pink Floyd, Eva Cassidy, Phil Collins, Chris Jones, Iiro Rantala, Herbie Hancock etc. pieces with great dynamics.
An externally hosted image should be here but it was not working when we last tested it.
Attachments
Last edited:
Finally my neigbour pianist Marian Petrescu had time to listen to AINOs! He has recently been in Australia making a new record with James Morrison.
Marian said that the sound is very clean and neutral, balance and timbre is very precise, just like at recording studio's mixing table! Sound image is more diffuse, like a ball in front of him - and he likes it! He could hear the breathing of horn players too. We also heard easily the difference in his piano's sound in different recordings (Tribute to OP vs. Thrivin'). I had to play louder than I am used to, minidsp's input got overdriven a couple of times but we couldn't hear it.
Here is the master with his brother Mihai playing at the school where our daughters go to. I organized this concert in 2011.
Marian said that the sound is very clean and neutral, balance and timbre is very precise, just like at recording studio's mixing table! Sound image is more diffuse, like a ball in front of him - and he likes it! He could hear the breathing of horn players too. We also heard easily the difference in his piano's sound in different recordings (Tribute to OP vs. Thrivin'). I had to play louder than I am used to, minidsp's input got overdriven a couple of times but we couldn't hear it.
Here is the master with his brother Mihai playing at the school where our daughters go to. I organized this concert in 2011.
An externally hosted image should be here but it was not working when we last tested it.
Gerrit Boers has done a rather similar dipole speaker using B&G Neo10S, a home-made planar tweeter, pair of 10" woofers and a monopole sub. His work with transducers, dsp and Ncore amplifiers is superior handcraft! Directivity measurements look like from ARTA.
http://www.diyaudio.com/forums/mult...-way-dipole-active-subwoofer.html#post4429872
http://www.diyaudio.com/forums/mult...-way-dipole-active-subwoofer.html#post4429872
12" for upper bas
Hi
I can see you use a 12" woofer from 160 Hz to 600 Hz.
I was planning on using a 10" woofer from around 150 Hz to like 800 Hz.
From like 30 Hz I am using two 12" Wavecor in a W-baffle.
For 800 Hz and up I am planning to use a Scan Speak 4" fullrange (I have that allready. But thinking of going with a BG Neo8PDR or BG Neo10PDR, and for the tweeter(s) I still do not know what to use.
Any reason you use such a big woofer. I think Linkwitz uses an 8" in his latest project, Do not remember at what frequency he is crossing.
best regards
uwe
Hi
I can see you use a 12" woofer from 160 Hz to 600 Hz.
I was planning on using a 10" woofer from around 150 Hz to like 800 Hz.
From like 30 Hz I am using two 12" Wavecor in a W-baffle.
For 800 Hz and up I am planning to use a Scan Speak 4" fullrange (I have that allready. But thinking of going with a BG Neo8PDR or BG Neo10PDR, and for the tweeter(s) I still do not know what to use.
Any reason you use such a big woofer. I think Linkwitz uses an 8" in his latest project, Do not remember at what frequency he is crossing.
best regards
uwe
Interrested by close thoughts as well
A easy answer could be SL has a higher XO between the sub and low-mids but I don't know, I surmise the 12" is better to go as low as Juhazi crossed (just at the limit where the ear can "see" the localization ! And where the mid-bass has its life (or was it jut a question of polar map and smoothing between drivers ?)
I asked to Juhazi if the same driver he uses but in its v2 Beyma has been launching stays a valid choice with the same crossover.
I thinked to something close as you planed (10") to use a Neo 8s as a full range to 8k/10k hz as I'm not sure 600 hz is a perfect match for the Neo8s & a 12"... but my undertanding is a 10" in the Aino Gradient context couldn't not be crossed as low as Juhazi did ! Don't forgett in the beginning he went as high as more than 200 hz ! But I believe if high pass can be solved, the 12" is a good choice. (I mean if it can be EQ as Juhazi made in the low end)
Oliver in the beginning of the thread wrote something very interressant about the diameter and the bafle needs where it is crossed here!
I'm looking for more a FULL RANGE above the 12" (so no Neo10... and difficult to source now) ... Neo 8s could be a choice (as Neo10 for you own needs) but 95db sensivity (2.83v) is more my goal ! xrk971 member in the full range section talked about a pro 5" from PRV Audiowhich can be seen as a FR.... Although planars has some advantages !
Juhazi, do you please know the max spl your Neo8 PDR can output without too harsch distorsions ?
A easy answer could be SL has a higher XO between the sub and low-mids but I don't know, I surmise the 12" is better to go as low as Juhazi crossed (just at the limit where the ear can "see" the localization ! And where the mid-bass has its life (or was it jut a question of polar map and smoothing between drivers ?)
I asked to Juhazi if the same driver he uses but in its v2 Beyma has been launching stays a valid choice with the same crossover.
I thinked to something close as you planed (10") to use a Neo 8s as a full range to 8k/10k hz as I'm not sure 600 hz is a perfect match for the Neo8s & a 12"... but my undertanding is a 10" in the Aino Gradient context couldn't not be crossed as low as Juhazi did ! Don't forgett in the beginning he went as high as more than 200 hz ! But I believe if high pass can be solved, the 12" is a good choice. (I mean if it can be EQ as Juhazi made in the low end)
Oliver in the beginning of the thread wrote something very interressant about the diameter and the bafle needs where it is crossed here!
I'm looking for more a FULL RANGE above the 12" (so no Neo10... and difficult to source now) ... Neo 8s could be a choice (as Neo10 for you own needs) but 95db sensivity (2.83v) is more my goal ! xrk971 member in the full range section talked about a pro 5" from PRV Audiowhich can be seen as a FR.... Although planars has some advantages !
Juhazi, do you please know the max spl your Neo8 PDR can output without too harsch distorsions ?
Last edited:
Personally, I'd take a close look at the Linkwitz LX Mini. You can scale it up or down if preferred.
Getting the woofer further from the floor (or any room boundary) seems likely to work better with room acoustics. Multiple smaller woofers in a vertical line array is likely to work better with floor and ceiling bounce. Using di-pole between 100HZ and 1kHZ creates a sense of 3-D space in the frequency range where inter-aural cancellation has largely confused any imaging cues in the recording, but only works well if the speakers are at least 3 feet from any room boundaries (psycho-acoustic effect).
Due to inter-aural crosstalk in the lower midrange, most of the stereo effect we get from speakers is from the upper midrange (1kHZ - 6kHZ generally), so that's the range that could benefit the most from controlled dispersion, if what you want is to maximize the stereo imaging cues that are in the recording. Many people actually prefer to have side wall bounce in this range for a more "spacious" feel, but that would likely be at the expense of clarity of embedded imaging cues in the recording. Plus that spaciousness would always be the same, and in some rooms could get tedious to listen to over time.
Above 6kHZ I think omni-directional might be most enjoyable. But all of this depends on the room the speaker will be used in. How a speaker interacts acoustically with the room is usually the weak link IMO. If you were to tighten the dispersion of the upper midrange, and leave the rest with wider dispersion, the off axis FR might somewhat give you a reverse Fletcher-Munson curve (Loudness comp curve), which might be nice. It would help to choose the crossover frequencies with that in mind, if you want that.
Going dipole in the lower midrange would mean cancellation at the sides, which means less side wall bounce interferance, but also means you are likely to need significant active EQ, or you will have a substantial rolloff in the "power response" (which takes into account room reflections) as you go down in frequency from roughly 600HZ down to bass frequencies.
If you are just trying to copy an old speaker you heard and liked a lot, and maybe do it even better, be careful to remember that the room it was in was a very big part of what you were hearing. It may be wiser to start from scratch, and design exactly what you want for the room you have.
Getting the woofer further from the floor (or any room boundary) seems likely to work better with room acoustics. Multiple smaller woofers in a vertical line array is likely to work better with floor and ceiling bounce. Using di-pole between 100HZ and 1kHZ creates a sense of 3-D space in the frequency range where inter-aural cancellation has largely confused any imaging cues in the recording, but only works well if the speakers are at least 3 feet from any room boundaries (psycho-acoustic effect).
Due to inter-aural crosstalk in the lower midrange, most of the stereo effect we get from speakers is from the upper midrange (1kHZ - 6kHZ generally), so that's the range that could benefit the most from controlled dispersion, if what you want is to maximize the stereo imaging cues that are in the recording. Many people actually prefer to have side wall bounce in this range for a more "spacious" feel, but that would likely be at the expense of clarity of embedded imaging cues in the recording. Plus that spaciousness would always be the same, and in some rooms could get tedious to listen to over time.
Above 6kHZ I think omni-directional might be most enjoyable. But all of this depends on the room the speaker will be used in. How a speaker interacts acoustically with the room is usually the weak link IMO. If you were to tighten the dispersion of the upper midrange, and leave the rest with wider dispersion, the off axis FR might somewhat give you a reverse Fletcher-Munson curve (Loudness comp curve), which might be nice. It would help to choose the crossover frequencies with that in mind, if you want that.
Going dipole in the lower midrange would mean cancellation at the sides, which means less side wall bounce interferance, but also means you are likely to need significant active EQ, or you will have a substantial rolloff in the "power response" (which takes into account room reflections) as you go down in frequency from roughly 600HZ down to bass frequencies.
If you are just trying to copy an old speaker you heard and liked a lot, and maybe do it even better, be careful to remember that the room it was in was a very big part of what you were hearing. It may be wiser to start from scratch, and design exactly what you want for the room you have.
Last edited:
The specific crossover points have to do more with the physical width of the drivers / polar pattern to achieve a balanced off axis response. Your 4" (off the top of my head) should be crossed around 4kHz give or take. But what the 4" can do on the low end also has to do with the baffle width eg baffle step loss and excursion limits how low it can effectively go.
Because we are adding a 6dB boost per octave lowpass to achieve a flat on axis response the choice of a larger driver for the all critical midbass/midrange we should use a larger driver as Juha has done (excellently)
heinz (et al), check out solhaga's awesome AMT build over in the exotics forum. Covers >200 and up. Very low distortion and an effeciency of 102dB. It was as high as 108dB in a previous version, but thats crazy effecient making the S/N ratio critical of all your electronics.
Because we are adding a 6dB boost per octave lowpass to achieve a flat on axis response the choice of a larger driver for the all critical midbass/midrange we should use a larger driver as Juha has done (excellently)
heinz (et al), check out solhaga's awesome AMT build over in the exotics forum. Covers >200 and up. Very low distortion and an effeciency of 102dB. It was as high as 108dB in a previous version, but thats crazy effecient making the S/N ratio critical of all your electronics.
http://prvaudio.com/wp-content/uploads/2015/05/5MR450-NDY_Datasheet.pdf
some good measurements made by krk971 on the Full range section !
good polar patern (look at Harsch XO thread) ! should be ok to cross with the Beyma Juhazi chose ?(but here now in the new v2 only avaliable ?)
B&G Neo8S could also a second choice or equal choice to cross, but higher Juhazi made : say around 8Khz/10 K hz (with what and at least how....? the answer of Juhazi is clear about what he think about tweeter today if my understanding is correct !) ! So not the Neo8S not alone in the highs for having a wider high point source dispersion (is my understanding correct?)
I just don't know looking at the datasheet of the Neo8s how could perform its curve in the highs (after 2Khz higherthan 30°) and how much the low needs EQ to match a 12" at 600hz/800hz in relation to the max Spl it can outputs without distorsion (it's a planar after all !) ! The goal is to stay without box resonnance & without loosing life and too much spl output !
At least I would have a close sup 95 db sensivity (2.83v)... The beginning is the 12" of Juhazi 's design with the XO of Juhazi to have a lively and snapy midbass (hoping the low EQ is not at the price of a lost of transcient in this lively aera... but reading SL it seems it works !) No place for Vott A7, double Onken...
Reading the post of Bob I also had the idea of a LX like design but here the 12" is not sealed : which is needed if ly understanding is correct to have a cardioid around the XO with the upper 5" (here the PRV or a Supravox 5" : but with this last I have a doubt about its max spl output and noone measured its distorsions ! but Supravox : 135 LB SUPRAVOX : subjective from many french owners (who gave up the idea of a too complicated horn in a living room) but who knows ? Bob, I need to dig more about what you wrote !
Out of Topic here ! But the main idea is to beginn from the 12" design à la Aino Gradient (raw if I undersand also what Oliver said in the beginning of this thread)... and close to the needs of the poster above too !
Problems of many planars below...600 hz : the match with a light, snappy, impactfull mid-bass with transcient ! ? No ? Not saying about max spl of a planar without distorsion ?!
I believe the mid-bass deserves its own unit in a DIY ! Not like an ESL with a XO at 300 hz ! I believe the hardest thing to acheive is an acurate mid-bass in home environment ! OK there are many choices with well documented designs, but often if not a true mid-bass horn or cabinet (CD or classic driver) ... Something is missing ! Most the time, acousticaly speaking : it looks like more Massey-Fergusson in the low with a XO on a Ferrari above : good travel ? No !
some good measurements made by krk971 on the Full range section !
good polar patern (look at Harsch XO thread) ! should be ok to cross with the Beyma Juhazi chose ?(but here now in the new v2 only avaliable ?)
B&G Neo8S could also a second choice or equal choice to cross, but higher Juhazi made : say around 8Khz/10 K hz (with what and at least how....? the answer of Juhazi is clear about what he think about tweeter today if my understanding is correct !) ! So not the Neo8S not alone in the highs for having a wider high point source dispersion (is my understanding correct?)
I just don't know looking at the datasheet of the Neo8s how could perform its curve in the highs (after 2Khz higherthan 30°) and how much the low needs EQ to match a 12" at 600hz/800hz in relation to the max Spl it can outputs without distorsion (it's a planar after all !) ! The goal is to stay without box resonnance & without loosing life and too much spl output !
At least I would have a close sup 95 db sensivity (2.83v)... The beginning is the 12" of Juhazi 's design with the XO of Juhazi to have a lively and snapy midbass (hoping the low EQ is not at the price of a lost of transcient in this lively aera... but reading SL it seems it works !) No place for Vott A7, double Onken...
Reading the post of Bob I also had the idea of a LX like design but here the 12" is not sealed : which is needed if ly understanding is correct to have a cardioid around the XO with the upper 5" (here the PRV or a Supravox 5" : but with this last I have a doubt about its max spl output and noone measured its distorsions ! but Supravox : 135 LB SUPRAVOX : subjective from many french owners (who gave up the idea of a too complicated horn in a living room) but who knows ? Bob, I need to dig more about what you wrote !
Out of Topic here ! But the main idea is to beginn from the 12" design à la Aino Gradient (raw if I undersand also what Oliver said in the beginning of this thread)... and close to the needs of the poster above too !
Problems of many planars below...600 hz : the match with a light, snappy, impactfull mid-bass with transcient ! ? No ? Not saying about max spl of a planar without distorsion ?!
I believe the mid-bass deserves its own unit in a DIY ! Not like an ESL with a XO at 300 hz ! I believe the hardest thing to acheive is an acurate mid-bass in home environment ! OK there are many choices with well documented designs, but often if not a true mid-bass horn or cabinet (CD or classic driver) ... Something is missing ! Most the time, acousticaly speaking : it looks like more Massey-Fergusson in the low with a XO on a Ferrari above : good travel ? No !
Last edited:
Hi, I was at my summer cabin for the weekend - we have a so called "indian summer" now, temperature up to 21 Celsius and lake water still 18C!
Driver width and more than that the baffle width give limits for dipole operation. It is possible (realistically) to get only 2-3 octaves true dipole radiation between xo points (roughly 2½ with Ainos). Study this with The Edge simulation.
With the 12" Beyma I have total of 20dB eq! My strategy with minidsp is to first eq the response straight and thereafter set crossovers. I get by the book slopes that way and I'm not afraid to give the dsp-chip some work.
It is possible to use 12" and 4" drivers as dipole and cross at around 6-800Hz. Kreskowsky uses two 6½" lower mids in NaO Note and a single 8" in NoteII. Check this page NaO Note II RS Details
A large Neo8 or Neo10 will start beaming a lot vertically above 4kHz. Neo3 goes dipole to 6-7kHz.
To Bob - LXMini is a completely different kind of system, AINOgradient, NaO Notes and LX521 are in same league. The monopole low/subwoofer is my personal choice in heritage of Gradient 1.x and so far I'm happy with them. I don't have plans to make even a test version of dipole subs even if SL himself has pursuaded me to do so!
I am not convinced that my choices are the best practise and every diyer must find her/his way with these! Active dsp-controlled system is very flexible and experiments with drivers and dsp settings are easy and reversible!
Driver width and more than that the baffle width give limits for dipole operation. It is possible (realistically) to get only 2-3 octaves true dipole radiation between xo points (roughly 2½ with Ainos). Study this with The Edge simulation.
With the 12" Beyma I have total of 20dB eq! My strategy with minidsp is to first eq the response straight and thereafter set crossovers. I get by the book slopes that way and I'm not afraid to give the dsp-chip some work.
It is possible to use 12" and 4" drivers as dipole and cross at around 6-800Hz. Kreskowsky uses two 6½" lower mids in NaO Note and a single 8" in NoteII. Check this page NaO Note II RS Details
A large Neo8 or Neo10 will start beaming a lot vertically above 4kHz. Neo3 goes dipole to 6-7kHz.
To Bob - LXMini is a completely different kind of system, AINOgradient, NaO Notes and LX521 are in same league. The monopole low/subwoofer is my personal choice in heritage of Gradient 1.x and so far I'm happy with them. I don't have plans to make even a test version of dipole subs even if SL himself has pursuaded me to do so!
I am not convinced that my choices are the best practise and every diyer must find her/his way with these! Active dsp-controlled system is very flexible and experiments with drivers and dsp settings are easy and reversible!
Last edited:
Hi Juha, nice tradition than your beautifull summer cabins 
I had forgotten you EQed the 12" with as high as 20 Db in the low ! So my choice of a new Beyma with a so small Qts (0.19) seems too low for the 120-150 hz aera, despite it's not planned as a sub driver and the new Beyma I linked has a very low distorsion as the discontinued you choosed ! But at 0.19 it's very far to the 0.3 of your Beyma maybe when going as low !
Hum, I have to find a web page where I can simulate the Db loss in relation to the Qts and the bafle wifth (here 12"), I will look at of Edge manage also those T/S parameters.! My plan was to have a very good electrical damping for the mid-bass aera in OB to have both articulation and micro dynamic in more or less the same aera you choosed for it (not saying than no cabinet is a dream for a non skilled enthusiast) ... but without bafle in the mid-bass perhaps it's a lost cause !
Ok, thank you very much as well for the info about the vertical beaming of a Neo 8 . Yes, on the paper the 12" XOes around 600-800 hz while the FR 5" follows to stop at around 6K/8K Hz seems easier to me on the paper... while certainly not having the resolution of a Kaladex planar !
Cheers.
I had forgotten you EQed the 12" with as high as 20 Db in the low ! So my choice of a new Beyma with a so small Qts (0.19) seems too low for the 120-150 hz aera, despite it's not planned as a sub driver and the new Beyma I linked has a very low distorsion as the discontinued you choosed ! But at 0.19 it's very far to the 0.3 of your Beyma maybe when going as low !
Hum, I have to find a web page where I can simulate the Db loss in relation to the Qts and the bafle wifth (here 12"), I will look at of Edge manage also those T/S parameters.! My plan was to have a very good electrical damping for the mid-bass aera in OB to have both articulation and micro dynamic in more or less the same aera you choosed for it (not saying than no cabinet is a dream for a non skilled enthusiast) ... but without bafle in the mid-bass perhaps it's a lost cause !
Ok, thank you very much as well for the info about the vertical beaming of a Neo 8 . Yes, on the paper the 12" XOes around 600-800 hz while the FR 5" follows to stop at around 6K/8K Hz seems easier to me on the paper... while certainly not having the resolution of a Kaladex planar !
Cheers.
Last edited:
I hope I'm not getting too off subject here, but I have a few comments about speaker design based on what I've learned.
I love low bass that's acoustically relatively flat down to the 20's. it's very hard to get that out of an open baffle system, in a small space, so I prefer monopole closed box woofers with active EQ ahead of the power amp to do the 25HZ - maybe 100HZ. It's not just to get the low bass notes, and much better sounding drums. All upper frequencies in the real world come inside envelops that have bandwidth that often goes very low in frequency. Overall sense of reality gets diminished when the bandwidth of these envelops is lost because the speaker rolls off below 40 or 50HZ (or higher).
From roughly 60HZ - 300HZ typical rooms are likely to have their biggest acoustic problems, because of the space in frequency, between the comb filter cancellations, due to various reflections, depending on many variables of room design and speaker placement. This is the main cause of boomy bass. Since inter-aural crosstalk substantially blurrs stereo imaging below about 1kHZ, it might be arguable that the main thing to focus on in this frequency range is dealing with interaction with room acoustics, rather than worrying about imaging. A vertical line array will significantly reduce floor and ceiling bounce comb filter effects, and minimally damage whatever stereo imaging does come through. A horizontal line array would minimize side wall reflection comb filter effects, but blur stereo imaging more. I chose to do a vertical line array of four 5 inch drivers on each side, floor mounted and tilted back by 10 degrees, and am pretty happy with the result. Since I occasionally switch in my re-optimized Carver Holographic Generator circuit (which cancels inter-aural crosstalk if you sit right in the middle), keeping the baffle width small horizontally helps imaging a lot. From 100HZ - 1.4kHZ I do open back di-pole, to artificially recreate a sense of imaging type space, which works well enough for me to get over the fact that it's artificial, but the speakers need to be at least 3 feet out from any walls for this to be a good thing (psycho-acoustic effect). I also set up my speakers along the long wall of the rectangular room, so side wall reflections are minimal. They can create a sense of horizontal spaciousness, but not usually in a good way.
From about 1kHZ to about 6kHZ, we get our main stereo imaging effects, since in this frequency range, we perceive image location primarily form amplitude comparison, as opposed to timing (or phase) comparison in the below 1kHZ range. So inter-aural crosstalk doesn't damage this frequency range much. So I try to avoid having a crossover point in this range, so the acoustic output is more likely to be exactly matched side to side over this freq. range. And I prefer to have a monopole in this range, so variations in room acoustics won't unbalance things and damage imaging (so I don't do di-pole in this range).
Above about 6kHZ the shape of our outer ear comes into play, helping us determine height as well. It's my opinion that above 6kHZ a dipole tweeter arrangement is a good thing. It brings into play more room reflections in a typical living room, but the comb filter effect cancellations are so close together in frequency that we don't perceive them to exist, and we get a more open and airy sound in the cymbals.
The highest output of a cymbal in a typical band is up and down, not necessarily straight from the cymbal to the listener. This energy bounces off the floor and ceiling on it's way to the listener in a real world environment. In a recording, they put a mic a foot or so above the cymbal, and then a typical speaker spits that signal straight to the listener from one place. Speakers that aim a tweeter straight up can create a better sense of reality, undoing the mic issue, but then dirt and dust accumulate in the tweeter diaphram, so going dipole is what I prefer. It gets you the more diffuse cymbal sound, and the tweeters don't accumulate nearly as much dust.
I also highly prefer using a 4th order active crossovers, especially at the lower frequency crossover points.
If simple to use very high quality digital time delay chips were available, I'd also delay the upper band drivers relative to the woofer by roughly 10mS, since any 4th order crossover causes a delay of about that on the woofer output, at 100HZ, relative to the midrange driver. Small delay differences are probably not real perceptable, but 10mS is actually significant and perceptable to some degree. But I can't seem to find a simple to use IC chip that would do this with very high quality, so I ignore this issue for now.
I'm not sure what the Aino Gradient speaker design takes into account, but there's my 2 cents.
I love low bass that's acoustically relatively flat down to the 20's. it's very hard to get that out of an open baffle system, in a small space, so I prefer monopole closed box woofers with active EQ ahead of the power amp to do the 25HZ - maybe 100HZ. It's not just to get the low bass notes, and much better sounding drums. All upper frequencies in the real world come inside envelops that have bandwidth that often goes very low in frequency. Overall sense of reality gets diminished when the bandwidth of these envelops is lost because the speaker rolls off below 40 or 50HZ (or higher).
From roughly 60HZ - 300HZ typical rooms are likely to have their biggest acoustic problems, because of the space in frequency, between the comb filter cancellations, due to various reflections, depending on many variables of room design and speaker placement. This is the main cause of boomy bass. Since inter-aural crosstalk substantially blurrs stereo imaging below about 1kHZ, it might be arguable that the main thing to focus on in this frequency range is dealing with interaction with room acoustics, rather than worrying about imaging. A vertical line array will significantly reduce floor and ceiling bounce comb filter effects, and minimally damage whatever stereo imaging does come through. A horizontal line array would minimize side wall reflection comb filter effects, but blur stereo imaging more. I chose to do a vertical line array of four 5 inch drivers on each side, floor mounted and tilted back by 10 degrees, and am pretty happy with the result. Since I occasionally switch in my re-optimized Carver Holographic Generator circuit (which cancels inter-aural crosstalk if you sit right in the middle), keeping the baffle width small horizontally helps imaging a lot. From 100HZ - 1.4kHZ I do open back di-pole, to artificially recreate a sense of imaging type space, which works well enough for me to get over the fact that it's artificial, but the speakers need to be at least 3 feet out from any walls for this to be a good thing (psycho-acoustic effect). I also set up my speakers along the long wall of the rectangular room, so side wall reflections are minimal. They can create a sense of horizontal spaciousness, but not usually in a good way.
From about 1kHZ to about 6kHZ, we get our main stereo imaging effects, since in this frequency range, we perceive image location primarily form amplitude comparison, as opposed to timing (or phase) comparison in the below 1kHZ range. So inter-aural crosstalk doesn't damage this frequency range much. So I try to avoid having a crossover point in this range, so the acoustic output is more likely to be exactly matched side to side over this freq. range. And I prefer to have a monopole in this range, so variations in room acoustics won't unbalance things and damage imaging (so I don't do di-pole in this range).
Above about 6kHZ the shape of our outer ear comes into play, helping us determine height as well. It's my opinion that above 6kHZ a dipole tweeter arrangement is a good thing. It brings into play more room reflections in a typical living room, but the comb filter effect cancellations are so close together in frequency that we don't perceive them to exist, and we get a more open and airy sound in the cymbals.
The highest output of a cymbal in a typical band is up and down, not necessarily straight from the cymbal to the listener. This energy bounces off the floor and ceiling on it's way to the listener in a real world environment. In a recording, they put a mic a foot or so above the cymbal, and then a typical speaker spits that signal straight to the listener from one place. Speakers that aim a tweeter straight up can create a better sense of reality, undoing the mic issue, but then dirt and dust accumulate in the tweeter diaphram, so going dipole is what I prefer. It gets you the more diffuse cymbal sound, and the tweeters don't accumulate nearly as much dust.
I also highly prefer using a 4th order active crossovers, especially at the lower frequency crossover points.
If simple to use very high quality digital time delay chips were available, I'd also delay the upper band drivers relative to the woofer by roughly 10mS, since any 4th order crossover causes a delay of about that on the woofer output, at 100HZ, relative to the midrange driver. Small delay differences are probably not real perceptable, but 10mS is actually significant and perceptable to some degree. But I can't seem to find a simple to use IC chip that would do this with very high quality, so I ignore this issue for now.
I'm not sure what the Aino Gradient speaker design takes into account, but there's my 2 cents.
Last edited:
You should both to read the whole thread to understand what Juha,
Wanted to acheive with the Gradient speaker in mind both about the esthtical and patterns choices.
Really worth the read....
The only question I would have talking if bass, walls, etc, to stay on topic: what could give the same bass but fireing towards the ceilling instead the floor?
Bob, great information about first-aid to choose a concept...really deserve a thread, you should copy it and start one: subscribed.
Wanted to acheive with the Gradient speaker in mind both about the esthtical and patterns choices.
Really worth the read....
The only question I would have talking if bass, walls, etc, to stay on topic: what could give the same bass but fireing towards the ceilling instead the floor?
Bob, great information about first-aid to choose a concept...really deserve a thread, you should copy it and start one: subscribed.
This thread hadn't come up in a while and there's about 500 entries, so going back and reading the whole thing would take time. Speaker design is quite an art when you consider room interaction, psycho-acoustic effects and recording process errors. It never stops being interesting.