Speaking as somebody who designs both multiway and single driver speakers, I think you'll find most people who use the latter couldn't give two hoots.
@ Bill Coltrane, I see what you're saying, and I've been following along. It may have been difficult getting the words out, but I can see that you and planet10 were covering different angles to the problem...
I'm not sure exactly why you're highlighting this point of view, certainly it isn't discussed often.. but it's a fundamental way of looking at drivers and I think an important point can come out of it. That is simply that the response from a driver can be arbitrary. This means you can mount it with much care so as not to introduce acoustic issues or cabinet related resonances.. and while sometimes the result is good on its own, sometimes it may not be flat and sometimes the result may not be optimum without a little EQ.
I'm not sure exactly why you're highlighting this point of view, certainly it isn't discussed often.. but it's a fundamental way of looking at drivers and I think an important point can come out of it. That is simply that the response from a driver can be arbitrary. This means you can mount it with much care so as not to introduce acoustic issues or cabinet related resonances.. and while sometimes the result is good on its own, sometimes it may not be flat and sometimes the result may not be optimum without a little EQ.
Sorry for the late reply.
And how is that given characteristic being put into reality?
Magic does not happen.
In this case, the speaker casing creates an acoustic "phenomenon" (to simplify the semantics as the math in all these cases is exactly the same).
And it can be compensated for by an electrical "phenomenon". (only in 1 place in space, cause you can't adjust a 3d event with a 1d event.)
And how is that given characteristic being put into reality?
Magic does not happen.
In this case, the speaker casing creates an acoustic "phenomenon" (to simplify the semantics as the math in all these cases is exactly the same).
And it can be compensated for by an electrical "phenomenon". (only in 1 place in space, cause you can't adjust a 3d event with a 1d event.)
My point is that what you hear is the result of lots of "phenonenons" (filter, resonanse, peak, dip, tilt etc...name it what you want.)
And to think that a full range speaker has no "phenomenons" is just rediculous.
Which 'they'? If that is a blanket statement it's wrong, because I for one use them. QED. However, many people do not use double-blind listening tests for a variety of reasons, one of which is the same as the above: they couldn't give two hoots. As noted, I do not share that view myself, but they are perfectly within their rights to have different priorities.
Who said wideband drive units have no response deviations? I think you are reading something into what other people have written which isn't actually there. Of course they have response deviations from a nominal flat response.
What I and a couple of others queried was simply your particular use of the word 'filter', which you were applying in a way it is not normally used. That does not mean 'we' are ignorant of the electromechanical behaviour of moving coil drive units, or are unaware of the existence of response deviations from a given baseline: it simply means we do not typically describe them as 'filters', because that is not the terminology usually employed. In terms of general usage, as noted, most people use that word to describe additions, either electrical or mechanical, for the specific purpose of altering the response of a given drive unit
Last edited:
You use double blind listening tests on loudspeakers?
Can 99% of all loudspeaker manufacterers borrow your speaker shuffler?
Who said wideband drive units have no response deviations? I think you are reading something into what other people have written which isn't actually there. Of course they have response deviations from a nominal flat response.
What I and a couple of others queried was simply your particular use of the word 'filter', which you were applying in a way it is not normally used. That does not mean 'we' are ignorant of the electromechanical behaviour of moving coil drive units, or are unaware of the existence of response deviations from a given baseline: it simply means we do not typically describe them as 'filters', because that is not the terminology usually employed. In terms of general usage, as noted, most people use that word to describe additions, either electrical or mechanical, for the specific purpose of altering the response of a given drive unitfrom a baseline curve. As repeatedly pointed out, Dave was specifically referring to the use of electrical filter circuits -given that he has spent decades mechanically modifying drive units to help linearise their responses, I think it's fairly safe to say he is aware wideband drive units have a wide variety of deviations in their response characteristics, most of which are caused by forms of resonance in the cones, suspensions &c.
Semantics aside, the math for all these "phenomenons" is the same. Its the same in every field of engineering, audio is not an exeption.
And their influence on audibility is the same for all these "phenomenons".
A mechanical highpass "filter" or "phenomenon" sounds exactly the same as its electrical counterpart.
A speaker with no "filters" (or "phenomenons" or what ever you call them) does not exist.
Of course, as / when relevant. However, I also strongly believe this is a hobby, not a dictatorship, so if other people cannot, or prefer not to, fair enough.
I really don't know how many more times I can say this: we know there is no such thing as a loudspeaker with a mathematically perfect response. You are preaching to the preacher / choir. Nobody who has contributed to this thread believes there is any such thing. I hope that is now clear. However, stating that a mechanical high pass filter sounds the same as an electrical high pass, without any kind of context or qualification, is nonsense I'm afraid.
Unfortunately I have no measurements at all, I generally make speakers on a " suck it and see " basis, and tune by ear. I believe that German Physiks have made a few speakers on similar lines, and Rountree Acoustics made the OmniMon 1, both have appeared and been reviewed ( with measurements ) in hifi world magazine.
I haven't noticed any diffraction issues with the basket or magnet, I presume that I would be able to here such, as I experimented with a diffuser in front of a
6 inch full range driver, and the negative effect of it in place was quite audible. I have moved my head around whilst listening to them, and they do sound better with my ears at 90 deg to the axis of the drivers, this could be a diffraction issue, however I feel that it is more to do with the dispersion. I intend on making some more, but with the driver axis vertical. I have used both paper and woven glass fibre coned drivers, both paper cones didn't sound good but the woven glass fibre does, although this could be a coincidence.
Simple. It might. It might not. But our friend provided zero context as to how either are to be implemented, ergo, taken as written i.e. as a generalised propostion, it's nonsense. To give one example, you might create a mechanical low pass filter with, say, some form of damped slot loading over the front of the drive unit. But this will also alter the polar response in a manner consistent with the implementation, which is not the case for an electrical low-pass tracking the same acoustical response characteristic. As we all know, they don't (can't) work like that.
As an alternate example, you may be able to use a passive electrical filter to achieve a given response characteristic, but said filter may introduce, say, some series R into the circuit, which changes the box alignment (which is why we should always account for relevant series R for wire, connections, components etc. when designing an enclosure alignment), which a mechanical method of achieving the same basic response trend is unlikely to do.
To be clear, I am not criticising or promoting either electrical or mechanical manipulation (although when possible I favour mechanical solutions to mechanical problems). Both have their place. I was simply pointing out that the statement made was devoid of context, so is incorrect in a generalised sense.
Last edited:
The sirst is true, The second very misleading because the 2 filters will not be the same.
dave
Well, if we're going to go deep into semantics then they're all acoustic ultimately, since in the final analysis we listen to air movement and none of it would make a whole lot of difference in a vacuum.
Be that as it may, if you've got sufficient data, you can model any and all of it. The 'sufficient data' being the elephant in the room part.
Be that as it may, if you've got sufficient data, you can model any and all of it. The 'sufficient data' being the elephant in the room part.
You appear to have taken this in the intended context, and your assessment is reasonable...
When an ideal theoretical definition like this is brought up in this way, the default context should be of an equivalence.. nothing but the element in question has changed.
Here's the off-axis polar map for two very different size drivers.
Scanspeak 10F/8424 3" Driver
Off-axis characteristic:
100Hz-5kHz 120 coverage using a -6dB downpoint.
Above 5kHz coverage quickly narrows to 40 degree coverage
Seas FA22RCZ 8" Driver
Off-axis characteristic:
100Hz-2kHz 120 degree coverage
Above 2kHz coverage gradually narrows to 60 degrees by 5kHz
Above 5kHz coverage gradually narrows to 30 degrees by 10kHz
Conclusion
The 8" Seas driver has slightly less coverage in the upper treble compared to the 3" Scanspeak (30 degree versus 40 degree)
The 3" Scanspeak has much wider coverage in the 2kHz-5kHz region.
The 8" Seas has a more even power response (Gradual narrowing in the 2kHz-5kHz region)
The 3"Scanspeak would have a strong room reflection presence in the 2kHz-5kHz region making it difficult to optimize in terms of EQ.
I hope this helps.
Scanspeak 10F/8424 3" Driver
Off-axis characteristic:
100Hz-5kHz 120 coverage using a -6dB downpoint.
Above 5kHz coverage quickly narrows to 40 degree coverage
Seas FA22RCZ 8" Driver
Off-axis characteristic:
100Hz-2kHz 120 degree coverage
Above 2kHz coverage gradually narrows to 60 degrees by 5kHz
Above 5kHz coverage gradually narrows to 30 degrees by 10kHz
Conclusion
The 8" Seas driver has slightly less coverage in the upper treble compared to the 3" Scanspeak (30 degree versus 40 degree)
The 3" Scanspeak has much wider coverage in the 2kHz-5kHz region.
The 8" Seas has a more even power response (Gradual narrowing in the 2kHz-5kHz region)
The 3"Scanspeak would have a strong room reflection presence in the 2kHz-5kHz region making it difficult to optimize in terms of EQ.
I hope this helps.
Attachments
It 'should', but we all know about assumptions
so unless it's stated that's what is intended, I for one ain't going to put words into anybody else's mouth. To my mind, one of the problems with having unspoken default assumptions is that it itself assumes a basic knowledge base which may not be the case for some less experienced readers. You only need to look at a typical thread on wire to see how some do not even accept something as fundamental as equivalence. (I'm not saying or implying that is the case here, it's an example only)