The reason for a lower cross over point is that the effects that a cross over has carry 2-3 octaves above the cut off frequency, depending on the type of cross over (analog. For digital see below). I did say this already, but maybe the implied explanation, plainly written, is that lower cross over frequency = artifacts stay confined to lower frequencies, where the ear is not as sensitive.
This is maybe something I didn't plainly say, but to me it goes without saying that for certain applications (near field mixing), straight single driver with DSP and lack of sub frequency is better (more accurate) than a FAST or three way with woofer crossed at even <250 Hz.
We're also talking about closed box.
If the box is bass reflex it may very well be that the cross over artifacts are covered from those of the port. Hardly a "can't possibly hear it" situation, but rather a "something is screwing with the sound even more" one.
But I do want to keep a door open to possibility, because it would be great to have both accuracy and a little more max SPL, for non near field situations.
It would have to be done with brick wall (or close) digital filters, because of alignment considerations and the fact that they don't carry artifacts over for octaves above the cut off frequency. Who knows, maybe this would cut it even for near field.
Incidentally, I don't know what research you're talking about, other than anecdotal stories.
I doubt somebody ever carried out an experiment with all types of boxes built both with and without woofers crossed over at adjustable frequency to see at which frequency the cross over artifacts start to disappear.
It would need to be an immense experiment, carried out in several listening environments, with dozens of boxes and of course double-blind.
So.. no, there is no research out there. Not in the scientific sense of the word anyway.
I'm afraid both you and me are going to be confined in the opinion/anecdotal world for now.
With the caveat that one can absolutely measure the difference with and without cross over and see the impulse response change. One can only attempt fix that with FIR DSP with many taps (and close to 1/2 second processing time as a consequence).
Therefore, being that the difference is measurable, the onus is on you to point out to real scientific research that it is not audible.
This is maybe something I didn't plainly say, but to me it goes without saying that for certain applications (near field mixing), straight single driver with DSP and lack of sub frequency is better (more accurate) than a FAST or three way with woofer crossed at even <250 Hz.
We're also talking about closed box.
If the box is bass reflex it may very well be that the cross over artifacts are covered from those of the port. Hardly a "can't possibly hear it" situation, but rather a "something is screwing with the sound even more" one.
But I do want to keep a door open to possibility, because it would be great to have both accuracy and a little more max SPL, for non near field situations.
It would have to be done with brick wall (or close) digital filters, because of alignment considerations and the fact that they don't carry artifacts over for octaves above the cut off frequency. Who knows, maybe this would cut it even for near field.
Incidentally, I don't know what research you're talking about, other than anecdotal stories.
I doubt somebody ever carried out an experiment with all types of boxes built both with and without woofers crossed over at adjustable frequency to see at which frequency the cross over artifacts start to disappear.
It would need to be an immense experiment, carried out in several listening environments, with dozens of boxes and of course double-blind.
So.. no, there is no research out there. Not in the scientific sense of the word anyway.
I'm afraid both you and me are going to be confined in the opinion/anecdotal world for now.
With the caveat that one can absolutely measure the difference with and without cross over and see the impulse response change. One can only attempt fix that with FIR DSP with many taps (and close to 1/2 second processing time as a consequence).
Therefore, being that the difference is measurable, the onus is on you to point out to real scientific research that it is not audible.
Which is true as far as general concepts go, but your rather sketchily stated assumptions as to the localisation BW are yours, and as-is, do not appear to be well supported by most acoustical research. If we are pessimistic and use only 1 octave above the 40Hz you reference, we have only reached the threshold employed for quality performance with multiple subwoofers, as illustrated in Welti’s AES paper (under Harman). This is not exactly an unknown subject, aspects of which having been referenced back to Olson; Beranek and IIRC Novak also, while more recently we see elements covered in works by Keele, d’Appolito, Geddes, Toole & others. Moreover, you still haven't made any detailed statements on the implementation of the filters (order, response &c).
Nothing goes without saying if generalised statements are made that appear to contradict normal, widely understood practice, and also fail to make clear an intended context or specific design details & applications. Rather than a wideband driver crossed to a woofer at xyz frequency[ies], which is what you were talking about, you also now appear to be citing a equalised wideband drive unit run solo, which is irrelevant since the practical applications are necessarily different, and 'the' (my) question is regarding the practicalities / realities you did not originally appear to refer to or raise.
Since when are ‘we’ talking about boxes of any kind WRT bald crossover frequencies? You might be, but it's news to me that everyone else is. Since your remarks about vented boxes have precisely nothing to do with crossover frequencies in themselves however, I won't bother asking about this new generalised statement sans any context or factual data.
You appear to have omitted to mention the design context and details, together with your data for the given case / cases.
Well, since it’s fairly well built into standard engineering practice and works on acoustics, it’s a trifle surprising that you haven’t; aspects are noted by all of the above researchers / authors, and many others. The intrinsic point remains: a 40Hz – 60Hz crossover frequency, taken purely as a frequency and sans any other design details or qualifications whatsoever (as you had stated it) cannot be casually claimed as ‘less audible’ than a filter an octave or two higher in frequency, since most research as indicated does not support that view. If you wish to claim your view has general applicability, the onus is thus upon you to do so, not the other way around. Equally, if you wish to state it as applying in a narrower context, then this needs to be made clearer. If you don't do either, you shouldn't be surprised if your unsupported view largely devoid of context is questioned.
Now, may I suggest that this thread is allowed to revert back to being on topic; if this is a matter you feel strongly about, you can always create a dedicated thread in the multiway or subwoofer forum to discuss it.
Last edited:
Once the crossover point reaches a frequency where the XO frequency has a quarter wavelength that is larger than the centre-to-centre distance of the drivers being cross-over then some of the most significant issues with XOs go away.
One still hass to make sure that the drivers in question have symathetic sonic characteristics, that the box is good (so many are not), that the driver arrary brings drivers as close as possible, and that the XO slope and character fit the design.
We have been staunch believers in what FR brings to the table, and it is only recently (last 5 years of so) that we have started playing with WAW. We feel that well done, any losses due to having an XO are minimal, and having a true 25-20k-ish speaker that offers no excuses when faced with music that often challenges a singla FR.
Designing a speaker is all about compromises, and a good WAW's biggest compromise is the extra cost & complexity (either more amplifiers or a pricey passive XO (if the last is even possible)). But as Tysen V2 (with PLLXO & a decent but cheap 2nd amp for the woofers) shows this does not have to be over the top price-wize (stock drivers + PLLXO parts <$200 USD before 2nd amp).
dave
One still hass to make sure that the drivers in question have symathetic sonic characteristics, that the box is good (so many are not), that the driver arrary brings drivers as close as possible, and that the XO slope and character fit the design.
We have been staunch believers in what FR brings to the table, and it is only recently (last 5 years of so) that we have started playing with WAW. We feel that well done, any losses due to having an XO are minimal, and having a true 25-20k-ish speaker that offers no excuses when faced with music that often challenges a singla FR.
Designing a speaker is all about compromises, and a good WAW's biggest compromise is the extra cost & complexity (either more amplifiers or a pricey passive XO (if the last is even possible)). But as Tysen V2 (with PLLXO & a decent but cheap 2nd amp for the woofers) shows this does not have to be over the top price-wize (stock drivers + PLLXO parts <$200 USD before 2nd amp).
dave
This is an ongoing conversation, so I shouldn't have to specify all the underlying conditions as disclaimers in every post I write.
The closed box comes from post 490. I responded to the 400 Hz cross over suggestion on post 494 with my 40-60 Hz cross over suggestion on post 495. That same post talks about EQ and the fact that I would use the 6 inch driver, not the 3 inch one. Not much later I also specified, about the EQ, that it would be wide band to tame the rest of the bandwidth, starting at the rising low end response of the single driver.
I thought it was clear enough that I'm talking about running the single driver as much low as it can (max SPL demand being the constraining factor) , which for near field ends up being in the 40 Hz or even a little lower for a 6 inch.
The cross over at 40-60 Hz comes from the evident (to me, but now I'll write that too) intention of having the single driver take care of as much of the audio bandwidth as possible.
I'll give you that it is a generalization to say that cross-overs at 40 Hz sound better (can produce a more accurate sound) than ones at 250 Hz. It obviously depends on drivers used, cross over typology an so on. However, it was clear (again, to me, but here it is written down) that what I intended was that, all other parameters like dB/octave slope being equal, moving the cut off frequency as low as possible will prevent as much of the rest of the audio band from being affected by the bass driver. Of course, the other driver must be able to 'catch up' from the cross over point. I don't think not stating this in plain English, given the rest of the conversation, was going to be viewed as 'sketchy', but sorry if it did.
So to me the point of crossing over to a bass driver is only to gain the sub bass frequencies, not possible to being reproduced by a 6 inch without having to cut down the rest of the audio band too much (max SPL requirement).
I also specified multiple times that I'm talking about near field context, so I don't know how much clearer I should be about this.
I think I shouldn't have left any explanation unwritten now, but if so let me know and I'll try to clarify further.
Now, the point of all this debacle should be boiled down to "can a cross-over under 250 Hz be audible?", or maybe even better "can one or more cross over strategies be devised for a subwoofer to be added to a channel and it be reliably 'seamlessly' integrated to the rest of the audio band?". I gather the papers you mentioned answer with a yes?
Could you link to the specific papers, if available online?
Or maybe you could summarize what type of tests they did. Honestly, to me anything other than the test I described earlier falls short of being scientific, in the strict sense of the word, but I'll bite.
Anyway, it seems to me that if you could spend some time to describe measurements and experiment conditions of the research you base your claims on, that would be of great service for whoever will read this.
I'm definitely interested.
The closed box comes from post 490. I responded to the 400 Hz cross over suggestion on post 494 with my 40-60 Hz cross over suggestion on post 495. That same post talks about EQ and the fact that I would use the 6 inch driver, not the 3 inch one. Not much later I also specified, about the EQ, that it would be wide band to tame the rest of the bandwidth, starting at the rising low end response of the single driver.
I thought it was clear enough that I'm talking about running the single driver as much low as it can (max SPL demand being the constraining factor) , which for near field ends up being in the 40 Hz or even a little lower for a 6 inch.
The cross over at 40-60 Hz comes from the evident (to me, but now I'll write that too) intention of having the single driver take care of as much of the audio bandwidth as possible.
I'll give you that it is a generalization to say that cross-overs at 40 Hz sound better (can produce a more accurate sound) than ones at 250 Hz. It obviously depends on drivers used, cross over typology an so on. However, it was clear (again, to me, but here it is written down) that what I intended was that, all other parameters like dB/octave slope being equal, moving the cut off frequency as low as possible will prevent as much of the rest of the audio band from being affected by the bass driver. Of course, the other driver must be able to 'catch up' from the cross over point. I don't think not stating this in plain English, given the rest of the conversation, was going to be viewed as 'sketchy', but sorry if it did.
So to me the point of crossing over to a bass driver is only to gain the sub bass frequencies, not possible to being reproduced by a 6 inch without having to cut down the rest of the audio band too much (max SPL requirement).
I also specified multiple times that I'm talking about near field context, so I don't know how much clearer I should be about this.
I think I shouldn't have left any explanation unwritten now, but if so let me know and I'll try to clarify further.
Now, the point of all this debacle should be boiled down to "can a cross-over under 250 Hz be audible?", or maybe even better "can one or more cross over strategies be devised for a subwoofer to be added to a channel and it be reliably 'seamlessly' integrated to the rest of the audio band?". I gather the papers you mentioned answer with a yes?
Could you link to the specific papers, if available online?
Or maybe you could summarize what type of tests they did. Honestly, to me anything other than the test I described earlier falls short of being scientific, in the strict sense of the word, but I'll bite.
Anyway, it seems to me that if you could spend some time to describe measurements and experiment conditions of the research you base your claims on, that would be of great service for whoever will read this.
I'm definitely interested.
Last edited:
Wow! Take a little time off and ~ 100 years of real world hands on R&D is now wrong?! Wow!
Thanks Scott for filling in, at a glance I see a few things I can expand on as time permits, though as you probably know, this stuff has mostly been 'written' by mass quantities of R&D, on site mods, etc., so must study the history of the telephone and talking picture developments to see how every aspect of these designs and their spin-offs were arrived at, with the goal always being max speech intelligibility under a variety of real world situations from a hearing aid to a PA/annunciator in the extremes of world war.
GM
Thanks Scott for filling in, at a glance I see a few things I can expand on as time permits, though as you probably know, this stuff has mostly been 'written' by mass quantities of R&D, on site mods, etc., so must study the history of the telephone and talking picture developments to see how every aspect of these designs and their spin-offs were arrived at, with the goal always being max speech intelligibility under a variety of real world situations from a hearing aid to a PA/annunciator in the extremes of world war.
GM
Now, I don't have the audio experience that the previous posters have, so, I'll differ to them the discussion on particular topics regarding audio reproduction.
I do, however, have a scientific background.
When it comes to research, I can relate.
When someone comes in with a new idea that could have profound repercussion on the subject at hand, we would expect that the idea has some research and data to back it up.
Even flat-earthers take out a ruler and place it on the horizon to prove their point. The premisses are wrong, but it's at least a (misguided) try.
As the others mentioned, it is not to them to prove you wrong. It is up to you to collect the previous research, analyse it, add your own collected data to put it against the previous research and come with your conclusions. Then, others will need to reproduce your data and findings, and if it can be, then the general scientific knowledge is evolved.
That's scientific methodology, and that's how it has worked for centuries. And worked well enough.
Asking others to find the data for you, and ask them to prove your idea wrong is just... plain lazy. Sorry to be blunt.
I do, however, have a scientific background.
When it comes to research, I can relate.
When someone comes in with a new idea that could have profound repercussion on the subject at hand, we would expect that the idea has some research and data to back it up.
Even flat-earthers take out a ruler and place it on the horizon to prove their point. The premisses are wrong, but it's at least a (misguided) try.
As the others mentioned, it is not to them to prove you wrong. It is up to you to collect the previous research, analyse it, add your own collected data to put it against the previous research and come with your conclusions. Then, others will need to reproduce your data and findings, and if it can be, then the general scientific knowledge is evolved.
That's scientific methodology, and that's how it has worked for centuries. And worked well enough.
Asking others to find the data for you, and ask them to prove your idea wrong is just... plain lazy. Sorry to be blunt.
Funny, because all this stemmed from a thread about single driver designs. If there is any category of people who should be relatively immune from 'standards followed so far' I expect it to find it in that type of thread.
The hands on research I did myself, as opposed to many of the people posting here and on that original thread, shows that the impulse response is largely different whenever one adds a second or third driver, without, at the very minimum, proper considerations on the time alignment, EQ and phase flattening.
So the phenomenon in question is absolutely there, and I find the comparison with flat-earthers ridiculous.
This is about the audibility of this phenomenon. The correct analogy would be that even if the earth is round, it doesn't really matter when you are playing marbles with your friends.
As a further irony, the same people who say there is no problem crossing over at 250 Hz and cite some elusive 'scientific' research, don't have any problem telling you that a cross over at 2 kHz is audible, in the face of hands on experience from all over the world, since that design is the de-facto standard. Of course, they are right about the latter, but it's ridiculous that they dismiss the former and brush it of like that.
Luckily, I'm told there is 'scientific' research that can put this matter to rest. Great!
Since this 'research' is at the ready hand of Scottmuse, all I'm asking is that he posts the links, if they are available on line. So that I can take a look at what type of experiments they carried out.
And, FOR THE NTH TIME, I'm talking about sealed near-field, so please stay on topic. If you even ever had any experience on near-field listening/mixing, that is.
The hands on research I did myself, as opposed to many of the people posting here and on that original thread, shows that the impulse response is largely different whenever one adds a second or third driver, without, at the very minimum, proper considerations on the time alignment, EQ and phase flattening.
So the phenomenon in question is absolutely there, and I find the comparison with flat-earthers ridiculous.
This is about the audibility of this phenomenon. The correct analogy would be that even if the earth is round, it doesn't really matter when you are playing marbles with your friends.
As a further irony, the same people who say there is no problem crossing over at 250 Hz and cite some elusive 'scientific' research, don't have any problem telling you that a cross over at 2 kHz is audible, in the face of hands on experience from all over the world, since that design is the de-facto standard. Of course, they are right about the latter, but it's ridiculous that they dismiss the former and brush it of like that.
Luckily, I'm told there is 'scientific' research that can put this matter to rest. Great!
Since this 'research' is at the ready hand of Scottmuse, all I'm asking is that he posts the links, if they are available on line. So that I can take a look at what type of experiments they carried out.
And, FOR THE NTH TIME, I'm talking about sealed near-field, so please stay on topic. If you even ever had any experience on near-field listening/mixing, that is.
Last edited:
So, basically what everyone has been doing for quite a while now when dealing with XO.
And your findings and solution are to XO at 40Hz.
But, it's funny that distortion was left out of that list of data....
Also, about distortion, I'm sure you realize that any time you have a box with reflex, or even TL or any other type of orifice, that's a added sound to the direct one, of about the same power (at low frequencies). So 100% distortion at the low end.
Not to mention, incidentally, that with far field listening (which is not my field of expertise nor what this thread is about) that figure goes even higher than 100%, since reflected sound is usually more powerful than direct wave by several dBs.
So any 'critical' listening done with non sealed boxes, not in near-field, or any 'scientific' research that's not done in those conditions, is not applicable to what I'm talking about here.
Not to mention, incidentally, that with far field listening (which is not my field of expertise nor what this thread is about) that figure goes even higher than 100%, since reflected sound is usually more powerful than direct wave by several dBs.
So any 'critical' listening done with non sealed boxes, not in near-field, or any 'scientific' research that's not done in those conditions, is not applicable to what I'm talking about here.
Because everyone else keeps reverting to that. So to make a point that not near-field listening can't really be used as a non plus-ultra condition for audibility of anything, especially bass frequency artifacts.
Do you have anything constructive to add to this conversation or are you just going to nit-pick here and there?
Do you have anything constructive to add to this conversation or are you just going to nit-pick here and there?
I dont see Sax publishing a scientific paper, or asking others to write one for him. He is only suggesting that there could be sonic benefits from crossing lower than, where most people, engineers or scientists suggest.
Sometimes - not often Im glad. Senior members have a tendensy to attack an OP with a new idea like a pack of wolves.
Its strange that I with very little theoretical knowledge, perfectly understand what Sax is proposing. Yet, the Greybeards are crying for details, references and elaborations of curcomstances that arent needed to understand.
I have doubts that Sax's proposition holds water, but I see no reason to attack him so ferousiously.
I'm not only suggesting it. I'm stating it plain and clear.
That is, under sealed near-field conditions.
Anyway, I doubt that any of the R&D talked about here has been done in those conditions. Maybe, and I have my doubts, it has been done in far field.
Which to me has the same value of researching the 'smellability of farts when sitting on top of a landfil'.
That is, under sealed near-field conditions.
Anyway, I doubt that any of the R&D talked about here has been done in those conditions. Maybe, and I have my doubts, it has been done in far field.
Which to me has the same value of researching the 'smellability of farts when sitting on top of a landfil'.
Ok, that last statement was a little strong. 
There are speakers meant for living room type listening that sound great.
Still, not good for critical audibility research because of the setting being less than ideal to put a listener in conditions to discern low level details like the ones we're talking about.
There are speakers meant for living room type listening that sound great.
Still, not good for critical audibility research because of the setting being less than ideal to put a listener in conditions to discern low level details like the ones we're talking about.
Or habits, example and reference of monitoring from pro world (or at least recognised as such in the industry) which made you think only nearfield with closed box does have value for monitoring purpose?
I'm curious of your answer because in 20 years i've never encountered a full ranger used in any control room i've been in for 'critical listening' as you stated ( except the Auratone but i wouldn't qualify them as tool for critical listening even if they are useful and have a place in pro environnement).
An other question too: how do you deal with IMD in your reasonning of xoverless fullrange (or your absolute dream of monitors)?
I'm curious of your answer because in 20 years i've never encountered a full ranger used in any control room i've been in for 'critical listening' as you stated ( except the Auratone but i wouldn't qualify them as tool for critical listening even if they are useful and have a place in pro environnement).
An other question too: how do you deal with IMD in your reasonning of xoverless fullrange (or your absolute dream of monitors)?
Oh, one other thing i forgot to ask: how would you define 'panel' (dipoles) speakers which does have more than a hole in their box? 1000% distortion speakers?
I'm teasing but i'm serious as i've seen (and heard!) esl used as monitors for the style you seems to be involved in ( as well as ATC but those doesn't count in your view as they are 100% distortion box i suppose).
I'm teasing but i'm serious as i've seen (and heard!) esl used as monitors for the style you seems to be involved in ( as well as ATC but those doesn't count in your view as they are 100% distortion box i suppose).
Last edited:
The part about far field not being suitable for best audibility conditions is a direct consequence of measurements being done in real rooms and strategies of coping with it.
This thread may get you started. [ANNC] eBook on Accurate Sound Reproduction Using DSP
The book is very well written and easy to understand, so I highly recommend it. The eq software is also very good (although a little pricey), and the proprietary smoothing algorithm it uses comes from studies in psycho acoustics, which the books also talks about. Really good stuff, in my opinion.
As to the drivers and methodology, as I said, I'm afraid I'm confined to the anecdotal world.
What I want to get across though, is that so are the claims of other people here.
To get out of the anecdotal world, one would have to carry out an experiment like the one I described in post 21. Have I done it? No. NOBODY has.
So if you like a story, here it is. My claim in my case comes from a practical use in recording and mixing.
I make purist recordings with a custom binaural microphone, and the accuracy of the recordings made with such a microphone heavily rely on the EQ that needs to be applied to the raw recording in the mixing phase. The EQ's accuracy depends in turn on the speakers used in the equalizing process.
Long story short, the most accurate recordings come out from EQs that are found from sealed near-field speakers.
When in near field conditions, the audibility of adding another driver is unfortunately evident to me. The mutual distance between the two drivers is not negligible anymore and comb filtering in the cross over region is measurable (like EVERYTHING I said so far) and ALSO very audible. It's just physics. Same distance for different frequencies, reproduced by both drivers in the cross over region, means some frequencies add up and some cancel out. Not to mention the phase shift of the cross over, so that the acoustical distance is not the same as the physical distance. Comb filtering in a nutshell. Is it audible? Maybe not in far field, but in the near field it is. I have tested it with three way speakers like the Avantone mix tower, with musical content confined in the range of the mix cube when run solo (otherwise the test would be pointless). Now, that arguably depends on how they designed their cross over. Again, I can't possibly prove that, no matter what, a cross over is always audible. Not without carrying out that experiment above.
But even the opening of the bookshelf transmission line is enough to cause audible artifacts. And one can easily tell because depending on the listening distance, some details appear and disappear. This is more evident in purist type recordings, to me. To some more golden ears it might be even more evident. But to me it is enough that I can hear it in this type of setting, with that type of music (that I have the bad luck to be interested in recording ).
And not that the comb filtering doesn't happen in the far field, but it is masked by all sorts of other things, so much so it is entirely possible that crossing over at 250 Hz or less can be reliably achieved in a seamless way, in those conditions. I never said the opposite, by the way. I just don't critically listen to music or mix it in far field, so I honestly can't tell.
If you were expecting an unequivocal proof that no driver, no matter how it is crossed over, can be seamless in a near field scenario, I don't have it. As a matter of fact, I think that digital brick wall cross overs could take away enough of the comb filtering to get it under the threshold of audibility, even for near-field.
But, as planet10 said in post 23, you still need to make sure the drivers have similar characteristics.
But just to reiterate, people saying that cross over under 250 Hz can, in any listening condition, be seamlessly integrated to the rest of the audio band don't have proof of it either.
This thread may get you started. [ANNC] eBook on Accurate Sound Reproduction Using DSP
The book is very well written and easy to understand, so I highly recommend it. The eq software is also very good (although a little pricey), and the proprietary smoothing algorithm it uses comes from studies in psycho acoustics, which the books also talks about. Really good stuff, in my opinion.
As to the drivers and methodology, as I said, I'm afraid I'm confined to the anecdotal world.
What I want to get across though, is that so are the claims of other people here.
To get out of the anecdotal world, one would have to carry out an experiment like the one I described in post 21. Have I done it? No. NOBODY has.
So if you like a story, here it is. My claim in my case comes from a practical use in recording and mixing.
I make purist recordings with a custom binaural microphone, and the accuracy of the recordings made with such a microphone heavily rely on the EQ that needs to be applied to the raw recording in the mixing phase. The EQ's accuracy depends in turn on the speakers used in the equalizing process.
Long story short, the most accurate recordings come out from EQs that are found from sealed near-field speakers.
When in near field conditions, the audibility of adding another driver is unfortunately evident to me. The mutual distance between the two drivers is not negligible anymore and comb filtering in the cross over region is measurable (like EVERYTHING I said so far) and ALSO very audible. It's just physics. Same distance for different frequencies, reproduced by both drivers in the cross over region, means some frequencies add up and some cancel out. Not to mention the phase shift of the cross over, so that the acoustical distance is not the same as the physical distance. Comb filtering in a nutshell. Is it audible? Maybe not in far field, but in the near field it is. I have tested it with three way speakers like the Avantone mix tower, with musical content confined in the range of the mix cube when run solo (otherwise the test would be pointless). Now, that arguably depends on how they designed their cross over. Again, I can't possibly prove that, no matter what, a cross over is always audible. Not without carrying out that experiment above.
But even the opening of the bookshelf transmission line is enough to cause audible artifacts. And one can easily tell because depending on the listening distance, some details appear and disappear. This is more evident in purist type recordings, to me. To some more golden ears it might be even more evident. But to me it is enough that I can hear it in this type of setting, with that type of music (that I have the bad luck to be interested in recording
).And not that the comb filtering doesn't happen in the far field, but it is masked by all sorts of other things, so much so it is entirely possible that crossing over at 250 Hz or less can be reliably achieved in a seamless way, in those conditions. I never said the opposite, by the way. I just don't critically listen to music or mix it in far field, so I honestly can't tell.
If you were expecting an unequivocal proof that no driver, no matter how it is crossed over, can be seamless in a near field scenario, I don't have it. As a matter of fact, I think that digital brick wall cross overs could take away enough of the comb filtering to get it under the threshold of audibility, even for near-field.
But, as planet10 said in post 23, you still need to make sure the drivers have similar characteristics.
But just to reiterate, people saying that cross over under 250 Hz can, in any listening condition, be seamlessly integrated to the rest of the audio band don't have proof of it either.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.