Hi,
Isn't it amazing what you can get when you overcome the room effects. Few people realise how much high directivity improve the sound in a room especially below about 1kHz.
I'm using high directivity dipole line arrays below 1kHz and I can share the same observations.
- Elias
Was this by any chance the much maligned "Sweet 16" ?
This was originally published in 1961 in Popular Electronics and is absolutely notorious with the local Hifi industry here..
Sweet 16
Last edited:
Was this by any chance the much maligned "Sweet 16" ?
This was originally published in 1961 in Popular Electronics and is absolutely notorious with the local Hifi industry here..
Sweet 16
Nice link. The bottom part of that page should provide some great motivation to build a line array with these drivers.
well, presumably long line arrays (of 3m and around) have some problems, and exactly because of those problems people are trying various forms of curving and/or tapering
10x10 (close to around 0.6x0.6m?) is relatively free from those problems
not to mention that full range dipole long line array would need to be really huge
best,
graaf
10x10 (close to around 0.6x0.6m?) is relatively free from those problems
not to mention that full range dipole long line array would need to be really huge
best,
graaf
it could be misleading to evaluate and compare stereo speakers in mono because they could sound significantly different in stereo than in mono
for comparison purposes it would better to build two 7x7 square arrays and two 50-driver line arrays and comapre them in pairs and with stereo program
best,
graaf
well, I know
therefore my suggestion is - give up potentially meaningless/misleading comparison and just build another 10x10 to try how they sound in stereobest regards,
graaf
I know little about this stuff, but thought I'd throw some ideas out there anyway. It seems my idea is not very wacky at all when looking at some of the others either suggested or actually built!
So anyway, what about this:
Take a piece of schedule 80 PVC, maybe 6" pipe size (That would be about 7" OD I think; maybe you want smaller?) and mount your line array in the side of that. Now the baffle does not have any corners (except for the speaker flanges- which is why you may want smaller pipe).
Now cap the top of the pipe.
Have the bottom of the pipe open into a box. In the box, we place our bass drivers. Maybe 2 10's? Then we port (or not) the box, tuning the port as needed. Perhaps a push-pull of 2 12's (one front, one rear, or one left, one right?)
For a dash of flavor, add a rear-firing tweeter with an attenuator...
Is that crazy enough for you?
PS you can heat the pipe to bend it if you feel the need...
So anyway, what about this:
Take a piece of schedule 80 PVC, maybe 6" pipe size (That would be about 7" OD I think; maybe you want smaller?) and mount your line array in the side of that. Now the baffle does not have any corners (except for the speaker flanges- which is why you may want smaller pipe).
Now cap the top of the pipe.
Have the bottom of the pipe open into a box. In the box, we place our bass drivers. Maybe 2 10's? Then we port (or not) the box, tuning the port as needed. Perhaps a push-pull of 2 12's (one front, one rear, or one left, one right?)
For a dash of flavor, add a rear-firing tweeter with an attenuator...
Is that crazy enough for you?
PS you can heat the pipe to bend it if you feel the need...
graf, I don't understand why you like the 10 x 10 idea so much... have you looked at the polar response WRT frequency? Do you like what you see??
I don't.
I prefer the 1 x 100 in all regards, especially stereo.
Have you compared them in the real world?
Keep in mind that a vertical array with no baffle surface area will not sound the same as a 10 x 10 with a large baffle surface area and dimension. The baffle counts.
_-_-bear
Jim, you do not want the vertical array to vent into a box with bass drivers. Also, the smaller pipe will not be as good sounding as a large diameter pipe. Of course the small diameter pipe will look thinner...
I don't.
I prefer the 1 x 100 in all regards, especially stereo.
Have you compared them in the real world?
Keep in mind that a vertical array with no baffle surface area will not sound the same as a 10 x 10 with a large baffle surface area and dimension. The baffle counts.
_-_-bear
Jim, you do not want the vertical array to vent into a box with bass drivers. Also, the smaller pipe will not be as good sounding as a large diameter pipe. Of course the small diameter pipe will look thinner...
why do I like the idea? generally for the same reasons all proponents of multiple drivers arrays like it - much lower distortions of any kind, better sensitivity, better power handling/less compresion etc.
why I prefer this square array idea to the idea of long line array?
because of the time domain, there is much better integration of individual driver's output in time domain than in case of long line array, much less delays resulting in much less smear in time response
to some degree it is probably a matter of subjective preferences but it remains an undisputed fact that many listeners speak unfavourably about phasiness and/or bloated imaging of long line arrays
I am concerned with what I can read for example in Dr Griffin's paper:
http://www.systemsolutions.co.za/Tech Zone/Line Array White Paper.pdf
I think that indeed this significant smear in time response can be responsibe for this alleged phasiness and as to bloated imaging it can be the result of rather unnatural acoustical excitation of the listening room due to multiple delayed reflections off the floor and ceiling
we have to remember that such an array is not a real line source, it can simulate some characteristics of a line source but unlike in the case of a true line source there are those multiple singular reflections off the floor and ceiling
just my hypotheses of course
me neither, and I not only looked at - I even posted some measurements of 4x4 square array above in this thread - and yes, off axis response is a nightmare but they are generally very directional and probably all this horribly coloured reflected sound is below the level of direct sound enough to be audibly harmless as such
all I know is that the people that have tried it - like it
perhaps it looks bad but it doesn't sound bad?
I like the idea of 10x10 more, tried to explain above why
but You say You prefer 1x100 in all regards, especially stereo
Am I to understand that You have made such real world comparisons of 10x10and 1x100?
best regards,
graaf
Last edited:
graf,
The idea that the image is "bloated" in a large line array I do not think is correct.
I have not experienced this at all... perhaps the converse is true, the image in a "normal" system is highly compressed?? Something to consider.
A 4x4 array is not the same as a 10 x 10 array by any means. The 4 x 4 is perhaps within a range of somewhat acceptable in terms of polar response.
IF you are going to make the case for the 10 x 10 array being "good on axis", then the same thing can be said for the vertical array - get on the center line axis!
Anyone listening to a large vertical ESL can tell you that the idea that sound is perceived as "smeared" because of the long vertical axis being equally radiated is simply untrue. It just doesn't sound that way at all.
Your array if I remember the picture is supplemented by an HF horn - so your array is mostly operating in the region where it tends to more or less look like an omnidirectional (ok, half space) source... the long line array can operate satisfactorily up to a much higher frequency.
So, let's do a thought experiment for a minute, ok?
What would you say about a 2 x 50 array?
Now, how about a 4 x 25? (that's just stacked 4 x 4s...)
Next, how about a 6 x 16?
etc...
Yes I have made comparisons of multiple driver arrays in various configurations...
People are doing all sorts of things, and the level of one's satisfaction for any given way of doing things depends on a variety of factors and individual preferences. I can't deny anyone's preferences, I can only advocate for what I have found to be superior, and try to say why...
Any line source made of individual drivers is by definition going to have some compromises, the only question is which of the various collateral compromises does one find acceptable.
_-_-bear
The idea that the image is "bloated" in a large line array I do not think is correct.
I have not experienced this at all... perhaps the converse is true, the image in a "normal" system is highly compressed?? Something to consider.
A 4x4 array is not the same as a 10 x 10 array by any means. The 4 x 4 is perhaps within a range of somewhat acceptable in terms of polar response.
IF you are going to make the case for the 10 x 10 array being "good on axis", then the same thing can be said for the vertical array - get on the center line axis!
Anyone listening to a large vertical ESL can tell you that the idea that sound is perceived as "smeared" because of the long vertical axis being equally radiated is simply untrue. It just doesn't sound that way at all.
Your array if I remember the picture is supplemented by an HF horn - so your array is mostly operating in the region where it tends to more or less look like an omnidirectional (ok, half space) source... the long line array can operate satisfactorily up to a much higher frequency.
So, let's do a thought experiment for a minute, ok?
What would you say about a 2 x 50 array?
Now, how about a 4 x 25? (that's just stacked 4 x 4s...)
Next, how about a 6 x 16?
etc...
Yes I have made comparisons of multiple driver arrays in various configurations...
People are doing all sorts of things, and the level of one's satisfaction for any given way of doing things depends on a variety of factors and individual preferences. I can't deny anyone's preferences, I can only advocate for what I have found to be superior, and try to say why...
Any line source made of individual drivers is by definition going to have some compromises, the only question is which of the various collateral compromises does one find acceptable.
_-_-bear
I am not talking about an idea but about reported observations of both sceptics (for example Linkwitz) and supporters (for example Dr Griffin writing about "Sound Bloom Caused by Unequal Vertical Sound Path Lengths Between Drivers and Listener in Near Field for a Flat Baffle" and the resulting need of line tapering)
I can understand that, exactly therefore I say "to some degree it is probably a matter of subjective preferences" but at the same time "it remains an undisputed fact that many listeners speak unfavourably...", doesn't it?
not really, not that one I have experience with, picture and measurement above in this thread
but on the other hand polar response is not a goal in itself, it is the final sound and it probably depends ultimately on combination of many factors, and it could be probably in this case of 10x10 - perhaps high overall directivity makes uneven offaxis response harmless as such
IF you are going to make the case for the 10 x 10 array being "good on axis", then the same thing can be said for the vertical array - get on the center line axis!
of course, but the smear in time domain is greater in case of 1x100 and perhaps this is the source of problems reported by some listeners
that is my point
let me repeat what I have already said - large vertical ESL is true line source, long array is not and it makes the crucial difference from the time domain perspective
I would say that more probably there could be audible polar response problems because overall horizontal directivity could not be high enough to cover nastiness of the off-axis sound
therefore I think that those would be worse alternatives to 10x10 and also 1x100
for me it is a choice between 1x100 and 10x10, if one is happy with 1x100 that's fine, if not then one can try 10x10 (aside purely practical problems - I think that 10x10 even on a big baffle, say 80x130 cm, is still more practical than loudspeaker that is 300 cm tall)
including 10x10 vs 1x100?
yes, I agree
and that is also exactly what I am doing...
best regards,
graaf
Last edited:
graf,
You did not address the key point I made about your 10 x 10 which is that you apparently roll it off and switch to a horn of some sort.
I understand what others have said about the line array and time smear. It may become evident when the HF output of a given driver is flat out above some frequency starting at 5kHz and going up... dunno. Otherwise I have not experienced the same subjective auditory impression myself - so far.
No, I have not built a 10 x 10.
others, not a 10 x 10 in particular.
Would you please state what size drivers you are using and what frequency you are rolling them off, and at what measured acoustic slope that works out to be? I am curious.
I did a quick read of Dr. Griffin's paper. In the main it merely recapitulated things that came before. I'm not sure that it breaks any new ground. A good overview for sure!
As far as what other people have said in various published places, their opinions may or may not be correct. In the past the "orthodoxy" of the time has ended up being rather wrong...
There are ESLs that are made up of an array multiple smaller cells - like some Sound Labs products... I never ever heard any "time smear" from them... thinking back. So, while it is possibly an issue, I remain skeptical.
As far as the issues surrounding the alleged lack of horizontal energy from the 10 x 10, again the question remains what frequencies you are running at?? Low enough and the shape of the array becomes moot and meaningless... only when you start getting higher does it come into play in a serious way.
And, imo, the "sound bloom" issue is actually a non-issue. Standard "point source" approximating speakers can create an unnaturally small soundfield in many instances. Although properly set up the soundfield of a stereo pair should not be "small" or "short" especially at moderately loud listening levels and above - imo, of course.
Also, from the time domain perspective the line array and the single source vs. long ribbon or ESL have no difference. If you have a radiating source some distance from the center point, you have it period. It can be an edge diffraction for that matter. There is a difference in terms of comb filter effects due to the physical difference in spacing between radiating sources at frequencies HIGH enough for that to take place. Dr. Griffin mentions this clearly.
_-_-bear
PS. for clarity - i think a 10 x 10 array of the 2" drivers we are talking about running full range would not be a good solution. One could
discuss what frequency one might cut them off, and then discuss the effect of having a supplemental HF driver at some distance from the acoustic center of the array?
You did not address the key point I made about your 10 x 10 which is that you apparently roll it off and switch to a horn of some sort.
I understand what others have said about the line array and time smear. It may become evident when the HF output of a given driver is flat out above some frequency starting at 5kHz and going up... dunno. Otherwise I have not experienced the same subjective auditory impression myself - so far.
No, I have not built a 10 x 10.
others, not a 10 x 10 in particular.
Would you please state what size drivers you are using and what frequency you are rolling them off, and at what measured acoustic slope that works out to be? I am curious.
I did a quick read of Dr. Griffin's paper. In the main it merely recapitulated things that came before. I'm not sure that it breaks any new ground. A good overview for sure!
As far as what other people have said in various published places, their opinions may or may not be correct. In the past the "orthodoxy" of the time has ended up being rather wrong...
There are ESLs that are made up of an array multiple smaller cells - like some Sound Labs products... I never ever heard any "time smear" from them... thinking back. So, while it is possibly an issue, I remain skeptical.
As far as the issues surrounding the alleged lack of horizontal energy from the 10 x 10, again the question remains what frequencies you are running at?? Low enough and the shape of the array becomes moot and meaningless... only when you start getting higher does it come into play in a serious way.
And, imo, the "sound bloom" issue is actually a non-issue. Standard "point source" approximating speakers can create an unnaturally small soundfield in many instances. Although properly set up the soundfield of a stereo pair should not be "small" or "short" especially at moderately loud listening levels and above - imo, of course.
Also, from the time domain perspective the line array and the single source vs. long ribbon or ESL have no difference. If you have a radiating source some distance from the center point, you have it period. It can be an edge diffraction for that matter. There is a difference in terms of comb filter effects due to the physical difference in spacing between radiating sources at frequencies HIGH enough for that to take place. Dr. Griffin mentions this clearly.
_-_-bear
PS. for clarity - i think a 10 x 10 array of the 2" drivers we are talking about running full range would not be a good solution. One could
discuss what frequency one might cut them off, and then discuss the effect of having a supplemental HF driver at some distance from the acoustic center of the array?
Last edited:
Yeah, so you could make 5-channels worth!
The low frequencies...there's a goodly chunk of total driver area, but such tiny drivers probably have a resonance frequency of 150-200 Hz, so you simply will not get low frequency output, you'll need a subwoofer. I suppose you could boost the heck out of them...
My inclination would be if only two speakers to build an angled array, i.e. each speaker has 2 columns of the 2" but they are on baffles angled away from each other somewhat. That way you could spread the high frequencies instead of just interference. But you would need to make polar response measurements and calculate the best angle.
Also keep in mind that since you will probably wire some in series, and they each have a different resonance, they will not share the power/excursion evenly at those frequencies, so at each frequency certain of the drivers will be driven harder. Be careful of damage.
time smear can be a problem regardless of frequency, it is a question of transient and waveform shape
the array was built by another guy inspired by my posts on a local Polish audioforum audiostereo.pl, that particular tweeter was His idea, I think it could be better without it
therefore I have referred to it. Dr Griffin is a diyaudio user BTW
it is not alleged, it is pure physics, and measured, take a look on the measurement above
no need telling me that - I am an omni/reflecting guy, and I use extreme reflecting setup
this 10x10 is just a thought experiment
hmmm... Are You sure?
I can accept that comb filter is in most cases an inaudible (from frequency domain perspctive) measurement artifact as Dr Toole suggests but here we have separate sound sources, separate wavefronts, the wavefronts of sound from separate sources in an array are not adding to form one wavefront
in time domain totally different situation than a true line source where the wavefront is as one
well, I don't know, what I know is that some people who tried it liked it
and that "If at first, the idea is not absurd, then there is no hope for it"
well, anyway, to each his own
best
graaf
Last edited:
Hi Thornspawn,
You better consider to buy a head clamping devise too in order to maintain a precise listening position because the central phantoms at the sweet spot will be very extreme. b
Ps: I have to my disposal 370 of those 2-inch drivers for upcoming projects but non will be used more than paired horizontally for SQ reasons.
Attachments
Hi Thornspawn,
b
Ps: I have to my disposal 370 of those 2-inch drivers for upcoming projects but non will be used more than paired horizontally for SQ reasons.
for SQ reasons or rather for misinterpreted simulated results reasons?
anyway, thanks for those graphs!
what I can see in them is that the directivity is very high from around 500 Hz and up, just as expected
and at normal listening distance of around 300 cm the sweet spot would for single listener but not that tiny - around 50 cm wide, enough for single listener's enjoyment
a bit of surprise for me are butterfly patterns above 5 kHz, it may be that te sound of such an array when driven full range could benefit from putting some absorption for those highest frequencies at first reflections points, no problem at all though, it is easy to absorb such high frequencies
best regards!
graaf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.