I am debating weather using an extended mid-range driver or a bigger low-mid crossed over to a high mid driver at 1000-2000. Using a bigger (10-12") low-mid driver has obvious advantages (power handling, compression, dynamics) at the price of the crossover.
Maybe someone has done the following experiment:
How about comparing the sound of 1) a full range 2) two of the same fullrangers with a cross over between them arond 1KHz. I`m sure someone with a digital xover has done it before. I would be happy to hear your comments.
regards,
seb
Maybe someone has done the following experiment:
How about comparing the sound of 1) a full range 2) two of the same fullrangers with a cross over between them arond 1KHz. I`m sure someone with a digital xover has done it before. I would be happy to hear your comments.
regards,
seb
Crossovers often are like your headline sentence; your meaning came across to the reader, just rather poorly.
(I don't mean this insultingly, I just thought it was a good analogy)
(I don't mean this insultingly, I just thought it was a good analogy)
I have not done exactly that test, but I have tried crossing two fullrange driver, one smaller and one larger using a 1:st order filter. Due to the large overlap of the drivers the x-over hade true 1:st order acoustic slope. Not common with ordinary drivers. I also time aligned the drivers. The result was quite pleasing.
The sum of a true 1:st order filter is phase linear (if drivers are time aligned). The only difference against a single driver is loobing issues and of cause some passive components in the signal path.
/Jesper
The sum of a true 1:st order filter is phase linear (if drivers are time aligned). The only difference against a single driver is loobing issues and of cause some passive components in the signal path.
/Jesper
swak said:
never use 800Hz - 2KHz region for xover point;
this is where your ears-brain combo is most sensitive;
Hi choky. The Linkwitz Orion is xo'd @ 1.44k. In the middle of that range you mentioned. Perhaps you can describe its faults?
http://www.linkwitzlab.com/TAC-review.htm
http://www.linkwitzlab.com/Orion-TSS-review.pdf
Thanks.
cheers,
AJ
Right, many very well regarded speakers cross over in the midrange. That`s why I posted the question on the first place. I thougt the 2 fullrange test would be quite conclusive on that point. Jespers experiment is very interestign, altough I would consider it better if one took the same FR driver for the highs and the lows. Very interesting anyway.
AJinFLA said:
to each his own
anyway-OP mentioned two fullranges ;
using fullranges crossing it in most sensitive area of human hearing is waste of potentials
at least for me
In my opinion, a well done fullrange is better than a mediocre crossover and a well done crossover is better than a mediocre full range system.
Perhaps full range would take more money and cabinet skills, and crossover would take more time.
Perhaps full range would take more money and cabinet skills, and crossover would take more time.
I see. Thanks for your answer.
Swak,
I guess the better question would be, if you are going to cross at 1-2k, why use a "fullrange" driver at all? That's tweeter territory.
Either use a small (3-5" fullrange with decent treble) driver and cross to a dedicated bass unit used below 150-300hz.
Or use a larger (8-10" extended range with decent bass) and cross high to a tweeter above 5k, etc. to avoid some of the inevitable breakup peaks/sound. Unless that's what you crave. I think that's why you sometimes see "helper" tweeters or "helper" woofers used with fullranges (I guess words like "crossover" and "2-way" are strictly taboo😉 ).
cheers,
AJ
This is a subject which I think have not been debated enough.
I’ve tried to list some parameters and compared between design variants. Feel free to comment and add/remove parameters.
What can we hear?
Uneven frequency response
Uneven phase response
Bad transient response (and low level ringing)
Tonal characters of different drivers
Passive components in signal path
Uneven reverbant field
Breakup modes
Limited power handling
How is most 2-ways (with x-over in the critical region) done?
+ Even frequency response
- 2:nd order (or higher) acoustic filter slopes -> uneven phase response
- Bad-ish transient response due to the above
- Large cone driver crossed to small dome or ribbon driver -> Tonal character difference
- Passive components in the signal path
- Uneven reverbant field
+ No or few breakup modes
+ Good power handling
How can a 2-way of fullrange drivers be done?
+ Even frequency response (if good drivers are used)
+ 1:nd order acoustic filter slopes -> even phase response
+ Good transient response due to above
+ Larger cone driver crossed to a smaller cone driver of preferable same material and structure -> Low tonal character difference
- Passive components in the signal path
- Uneven reverbant field
+ Few breakup modes
+ Good power handling
What about a single driver?
+ Even frequency response (good driver)
+ Even phase response (good driver)
+ Good transient response
+ No tonal differences
+ No (or very few) passive components
+ Even revarbant field (in the critical region)
- More breakup modes
- Lesser power handling
If the x-over is done right I would say we have a package with almost the qualities of a fullrage driver but with some other advantages. A matter of compromise of cause, but worth considering IMO.
/Jesper
I’ve tried to list some parameters and compared between design variants. Feel free to comment and add/remove parameters.
What can we hear?
Uneven frequency response
Uneven phase response
Bad transient response (and low level ringing)
Tonal characters of different drivers
Passive components in signal path
Uneven reverbant field
Breakup modes
Limited power handling
How is most 2-ways (with x-over in the critical region) done?
+ Even frequency response
- 2:nd order (or higher) acoustic filter slopes -> uneven phase response
- Bad-ish transient response due to the above
- Large cone driver crossed to small dome or ribbon driver -> Tonal character difference
- Passive components in the signal path
- Uneven reverbant field
+ No or few breakup modes
+ Good power handling
How can a 2-way of fullrange drivers be done?
+ Even frequency response (if good drivers are used)
+ 1:nd order acoustic filter slopes -> even phase response
+ Good transient response due to above
+ Larger cone driver crossed to a smaller cone driver of preferable same material and structure -> Low tonal character difference
- Passive components in the signal path
- Uneven reverbant field
+ Few breakup modes
+ Good power handling
What about a single driver?
+ Even frequency response (good driver)
+ Even phase response (good driver)
+ Good transient response
+ No tonal differences
+ No (or very few) passive components
+ Even revarbant field (in the critical region)
- More breakup modes
- Lesser power handling
If the x-over is done right I would say we have a package with almost the qualities of a fullrage driver but with some other advantages. A matter of compromise of cause, but worth considering IMO.
/Jesper
It is thought to be an experiment
I think I was not clear enough that the sole purpose of crossing over two *identical* (same model) FR drivers is to experiment in what degree a crossover is detrimental to percieved sound quality. A real setup would naturally use deivers appropriate for each of their frequency ranges. The point was to have an apples to apples comparison. The 1-2kHz piont was suggested because that is where we should be able to percive de negative effects most easily.
The *real* system I am thinking of is a 2 or 3 way with a compression driver in a controlled directivity waveguide above 1-1.5kHz or a 3 way with a fullrange from 300 to 6-12kHz dipole augmented below and horn augmented above.
I think I was not clear enough that the sole purpose of crossing over two *identical* (same model) FR drivers is to experiment in what degree a crossover is detrimental to percieved sound quality. A real setup would naturally use deivers appropriate for each of their frequency ranges. The point was to have an apples to apples comparison. The 1-2kHz piont was suggested because that is where we should be able to percive de negative effects most easily.
The *real* system I am thinking of is a 2 or 3 way with a compression driver in a controlled directivity waveguide above 1-1.5kHz or a 3 way with a fullrange from 300 to 6-12kHz dipole augmented below and horn augmented above.
I've tried both ...
I've a pair of JBLCS3115 speakers, each of which has a 15" woofer in BR arrangement, an 8" mid driver with a front horn covering about 350-1600Hz, and a horn-loaded compression driver for tweeter. I've also built various OB speakers using fullrange/widerange drivers. Some of the OB systems use woofers in big BR boxes, and some use dipolar woofers on the same baffles with the fullrange/widerange drivers, and some of them simply use a subwoofer for bass. For instance, I currently have OB speakers in my home theater system as front speakers. Each of them uses a BR box with two 15" woofers, an Audax PR170MO on OB for 300Hz to perhahps 7KHz (no lowpass), and a cheap Fostex horn tweeter.
The JBL speakers sound quite nice, but somehow I am more attracted to the sound of OB systems. To my ears, the Audax-based OB system rivals the JBL speakers in sound quality and perhaps surpass them in terms listening enjoyment. My experience is of course only one data point, and is not conclusive because I likely have not have extracted the best performance out of either system. Nevertheless, I just want to say that I think a 3-way system using a fullranger in a dipolar configuraiton can be quite good.
Cheers,
Kurt
I've a pair of JBLCS3115 speakers, each of which has a 15" woofer in BR arrangement, an 8" mid driver with a front horn covering about 350-1600Hz, and a horn-loaded compression driver for tweeter. I've also built various OB speakers using fullrange/widerange drivers. Some of the OB systems use woofers in big BR boxes, and some use dipolar woofers on the same baffles with the fullrange/widerange drivers, and some of them simply use a subwoofer for bass. For instance, I currently have OB speakers in my home theater system as front speakers. Each of them uses a BR box with two 15" woofers, an Audax PR170MO on OB for 300Hz to perhahps 7KHz (no lowpass), and a cheap Fostex horn tweeter.
The JBL speakers sound quite nice, but somehow I am more attracted to the sound of OB systems. To my ears, the Audax-based OB system rivals the JBL speakers in sound quality and perhaps surpass them in terms listening enjoyment. My experience is of course only one data point, and is not conclusive because I likely have not have extracted the best performance out of either system. Nevertheless, I just want to say that I think a 3-way system using a fullranger in a dipolar configuraiton can be quite good.
Cheers,
Kurt
Unfortunately, two identical drivers (fullrange or not), crossed at 1k, would have a completely different polar response than a single driver. So it wouldn't be an apples to apples comparison. Actually, someone has already done exactly what you described, using 2 small midbass units. I can't remember all the details, or I'd give you a link.
Maybe if you lowered the xo to 200hz or so, but even then still would be an imperfect comparison.
If you were interested in the audibility of crossover phase distortion, there are tests, like what I thought maybe choky had taken, such as SL's: http://www.linkwitzlab.com/x-phs-dist.htm
A little info you might also be interested in: http://www.linkwitzlab.com/frontiers.htm#F
cheers,
AJ
AJ,
Polar response would be different, but isn't that part of the point? Crossovers do awful things to polar patterns - even an LR4 is far from perfect in this regard (and in general it gets worse as you go down in order).
Linky himself addresses this (polar patterns) - not discussing his own designs, but discussing planars - and suggests off axis/reflected contribution must 'fill in the gaps' in the polar pattern. I think of it as a type of 'averaging' that goes on - and I can definitely tell the difference between my own dipole design and the full range design I put together. Someone on one of these sites suggested that when we move away from a single point source (to a multi-way) design the 'brain has to work harder.' I agree with this one hundred percent. We divide a signal electrically - through a very inexact process - then expect it to sum acoustically? No. I think we're making the ear/brain do most of the work.
I'm not a die hard full range person, but I do think there are significant benefits to keeping the mid band (lower/mid hundreds to 5 or 6 k) free of crossovers. The ear is significantly less sensitive outside of this midband.
Polar response would be different, but isn't that part of the point? Crossovers do awful things to polar patterns - even an LR4 is far from perfect in this regard (and in general it gets worse as you go down in order).
Linky himself addresses this (polar patterns) - not discussing his own designs, but discussing planars - and suggests off axis/reflected contribution must 'fill in the gaps' in the polar pattern. I think of it as a type of 'averaging' that goes on - and I can definitely tell the difference between my own dipole design and the full range design I put together. Someone on one of these sites suggested that when we move away from a single point source (to a multi-way) design the 'brain has to work harder.' I agree with this one hundred percent. We divide a signal electrically - through a very inexact process - then expect it to sum acoustically? No. I think we're making the ear/brain do most of the work.
I'm not a die hard full range person, but I do think there are significant benefits to keeping the mid band (lower/mid hundreds to 5 or 6 k) free of crossovers. The ear is significantly less sensitive outside of this midband.
Pick your poison. Single driver systems suffer from either a lack of and/or compromised bass or treble. Crossed systems have crossover probs (to some) but have the potential to have both resolved and balanced bass and treble.
Try using two identical "full range" drivers, say two 4", but instead of crossing @1200Hz, just low pass one of the drivers and let the other run full range. Then compare to a single 8" or a 4" to see which sounds better.
Try using two identical "full range" drivers, say two 4", but instead of crossing @1200Hz, just low pass one of the drivers and let the other run full range. Then compare to a single 8" or a 4" to see which sounds better.
hi ultrakaz,
but then aren't you just hearing the driver with its high end rolled off? crossovers enable driver integration as well as bandwidth limiting. and i think integration is where they have most or all of the problems. maybe i misunderstand what you're suggesting.
i agree that full range drivers are usually too much of a compromise by themselves (i hate 'polite bass' - i want to feel bass as well as hear it, even if that means some loss of accuracy). but if you augment the top and/or bottom octaves - you can design them to put crossover(s) out of the most critical band. with some care, you can even leave the crossover off the mid/wide range driver, where it seems to matter the most.
it's just like the different ways to cabinet load - BR works pretty well (despite being a resonant system) for the bottom two octaves or so, but I wouldn't want to do it somewhere in the upper mids - the ear is decidedly more forgiving outside of the critical (voice) band.
i don't want to be an evangelist - but in my own experience, the difference between a traditional two or three way and using a wide range in the middle band - the difference is not subtle. I love the spaciousness of a three way dipole system - and love a lot of traditional 3 ways in general - but a wide range driver system is absolutely the most involving of anything i've listened to. YMMV
😉
but then aren't you just hearing the driver with its high end rolled off? crossovers enable driver integration as well as bandwidth limiting. and i think integration is where they have most or all of the problems. maybe i misunderstand what you're suggesting.
i agree that full range drivers are usually too much of a compromise by themselves (i hate 'polite bass' - i want to feel bass as well as hear it, even if that means some loss of accuracy). but if you augment the top and/or bottom octaves - you can design them to put crossover(s) out of the most critical band. with some care, you can even leave the crossover off the mid/wide range driver, where it seems to matter the most.
it's just like the different ways to cabinet load - BR works pretty well (despite being a resonant system) for the bottom two octaves or so, but I wouldn't want to do it somewhere in the upper mids - the ear is decidedly more forgiving outside of the critical (voice) band.
i don't want to be an evangelist - but in my own experience, the difference between a traditional two or three way and using a wide range in the middle band - the difference is not subtle. I love the spaciousness of a three way dipole system - and love a lot of traditional 3 ways in general - but a wide range driver system is absolutely the most involving of anything i've listened to. YMMV
😉
Yes, that is a very important part of the point, since crossovers with one single fullrange are net very common 😉
Swak, I was simply pointing out why your comparison using 2 fullranges would be flawed. I think both you and m@ are mixing up the issues. The polar response is different (worse) not because of the crossover, but because the two (fullrange) drivers are non coincident. They can't be.
If you believe crossovers cause poor polar response, then you have never seen the polar response of a good coaxial, or better yet coincident driver, where there is greater alignment in the z-axis. They certainly use crossovers, don't they?
The polar responses will look very much like a fullrange, except most likely (much) better in the treble region.
cheers,
AJ
If you believe crossovers cause poor polar response, then you have never seen the polar response of a good coaxial, or better yet coincident driver, where there is greater alignment in the z-axis. They certainly use crossovers, don't they?
The polar responses will look very much like a fullrange, except most likely (much) better in the treble region.
cheers,
AJ
AJinFLA, you make a point.
Yes, that is right, the worse polar response problems are not due to the crossover, but to the driver separation. I have heard mixed things about coaxials and have not heard any. Some peaplo that know their stuff say (look over at AA) they introduce worse problems than the ones they solve.
I should have specified I was more interested in a comparison between 1 driver, no crossover (over a range) against two non coaxial/coincidental with XO.
Yes, that is right, the worse polar response problems are not due to the crossover, but to the driver separation. I have heard mixed things about coaxials and have not heard any. Some peaplo that know their stuff say (look over at AA) they introduce worse problems than the ones they solve.
I should have specified I was more interested in a comparison between 1 driver, no crossover (over a range) against two non coaxial/coincidental with XO.
Kurt (KCHANG)
I had overlooked your post. That is very interesting. May I ask if you measured the directivity of the audax @ 7k? That was one on my list, too bad the paper cone version is no longer available. Magnetars setup with 2 of these has olways catched my attention. There are similar drivers anyway (b&c and phl for example, ther was someone who compared the audax to a Pr65-neo, and found the latter better http://db.audioasylum.com/cgi/t.mpl?f=hug&m=101904). The visation B200 up to 12k is another candidate, although I have no idea about their dispertion pattern that way up. One advantage of this setup is certainly the simplicity of the system.
Did you prefered the BR over the OB subs?
PS: I read you were willing to audition the Summas, did you listen to them?
I had overlooked your post. That is very interesting. May I ask if you measured the directivity of the audax @ 7k? That was one on my list, too bad the paper cone version is no longer available. Magnetars setup with 2 of these has olways catched my attention. There are similar drivers anyway (b&c and phl for example, ther was someone who compared the audax to a Pr65-neo, and found the latter better http://db.audioasylum.com/cgi/t.mpl?f=hug&m=101904). The visation B200 up to 12k is another candidate, although I have no idea about their dispertion pattern that way up. One advantage of this setup is certainly the simplicity of the system.
Did you prefered the BR over the OB subs?
PS: I read you were willing to audition the Summas, did you listen to them?
- Status
- Not open for further replies.