Announcement

Collapse

HUG - here for all audio enthusiasts

INTRODUCTION- PLEASE READ FIRST TO UNDERSTAND THIS FORUM!

"This Harbeth User Group (HUG) is the Manufacturer's own managed forum dedicated to natural sound, realisable by controlling the confounding variables between tthe microphone and the listeners' ears.

For example, the design of and interaction between the hifi amplifier and its speaker load can and potentially will alter the sound balance of what you hear. To reproduce the sounds captured by the recording microphones, as Harbeth speakers are designed to do, you would naturally select system components (sources, electronics, cables and so on) that do not color the sound before it reaches the speakers.

Identifying components for their system neutrality should, logically, start with the interpretation and analysis of their technical, objective performance, as any and every deviation from a measurably flat frequency response at any point along the serial chain from microphone to ear is very likely to cause the total system to have an audible sonic personality. That includes the contribution of the listening room itself.

HUG specialises in making complex technical matters simple to understand, aiding the identification of audio components likely to maintain a faithful relationship between the recorded sound and the sound you hear. With our heritage of natural sound, HUG cannot be really be expected to guide in the selection, approval, endorsement or even discussion of equipment that is intend to introduce a significantly personalised sound to the audio signal chain. For that you should do your own research and above all, make the effort to visit an Authorised Dealer and listen to your music at your loudness on your loudspeakers through the various electronics offered there. There is no on-line substitute for that time investment in a dealer's showroom.

If you desire to intentionally tune your system sound to your personal taste, please consider carefully how much you should rely upon the subjective opinions of strangers. Their hearing acuity and taste will be different to yours, as will be their motives and budget, their listening distance, listening loudness and listening room treatment, not necessarily leading to appropriate equipment selection and listening satisfaction for you.

Alternatively, if faithfully reproducing the sound intended by the composer, score, conductor and musicians over your speakers is your audio dream, then understanding something of the issues likely to fulfill that objective is what this forum has been helping with since 2006. Welcome!"


Jan. 2018
See more
See less

Speaker stands for Harbeth speakers at public exhibitions

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Speaker stands for Harbeth speakers at public exhibitions

    In discussing our Bristol appearance for next month, I have brought up an issue that colleagues had never thought about: the height of the speaker stands at a show compared to the listening room at home. In my humble opinion, the material and construction of the stands can be a matter of personal taste, but the height is really important, down to the last cm. or two.

    Why?

    The speaker designer has to make a decision before he starts work on the crossover where to direct his measurement microphone, the one tool that will tell him about the technical performance of the loudspeaker - the frequency response. As most loudspeaker boxes have the drive units mounted vertically along the centre line (as all Harbeth speakers do) the position of the microphone horizontally left/right of the speaker (for a fixed vertical point off the floor) would give the same measured response. However, vertically, the frequency response that the microphone measures radically alters as the microphone is raised and lowered along that cabinet centre line, vertically.

    The sound waves from the bass/mid driver and tweeter arriving at the microphone take an path through the air from their respective voice coils. As the tweeter's voice coil is far in front of the woofer's*, and far closer to the microphone when the microphones is directed towards the upper quartile of the cabinet (approximately), it's wave leads the woofer's. Conversely, when the microphone is lowered to be directed to the lower half of the cabinet, the tweeter's sound wave lags that woofer's. That lead/lag is another way of saying that the tweeter's phase relationship with the woofer depends upon where the microphone is placed relative to the woofer and tweeter vertically. All professional speaker designers must decide for themselves where to clamp the microphone vertically relative to the drive units, and to mark that vertical point onto the prototype cabinet with a marker pen as the Reference Measurement Axis (RMA).

    Then the bun fight with the marketing department begins.

    Speaker designers generally opt for two reference measurement axis; both are equally valid and in the great scheme of things only give subtly different technical results. One designer might habitually consider his RMA to be when his measuring microphone is pointed precisely at the apex of the tweeter dome. Another designer, again following his personal preferences, might measure with a ruler the vertical separation between the centre of the tweeter and centre of the bass/midrange unit, divide that in half, and mark that point on the baffle as his RMA. That would be where he would point his measuring microphone.

    Once the designer has settled in his mind where his RMA is, he can start work on the design of the crossover. One thing is sadly inevitable: when designing, he can only adjust the system crossover/equaliser/time delayer to perfectly integrate the bass/mid and tweeter sound waves as they hit the microphone diaphragm (or listener's ears) at that one fixed, vertical RMA. Not 2cms above RMA, 13cms below RMA or at any other point in space vertically. Every multi-way can only behave at its best with the measuring microphone or listener's ears exactly at that correct vertical point and nowhere else. The RMA was fixed at the design of the speaker, and the network was conceived and adjusted to compensate for the fact that the tweeter is physically (and sonically) in front of the woofer. That compensation is not universal: it is only for one point in space above the floor, wherever the RMA was set.

    And? This is where the real world and the idealistic designer clash.

    The designer sticks to his guns that the RMA is at a certain identifiable point on the baffle: 'level with the tweeter', '3cms below the tweeter' or whatever. He knows that at that point and no other vertical position that the drive units integrate as he wishes. The marketing boys are made aware of that, and make some sketches of suitable stands, and they feed back some market survey results to the designer. They say 'sorry, but that RMA you told us about for model XYZ: it's far too low down in the cabinet. We've played around with some matching stands, and gee wizz, considering that the average listener has ears that are about 1.2m above the floor when seated, listening, we're going to have to make some really, scarily, domestically unacceptably high stands to lift those speakers up high enough for your RMA to match the listener's ears ..... can't you move your RMA upwards until it's level with the top of the cabinet, or even higher!, so that we can sell short, domestically acceptable stands..?'. The answer is no, he can't 'steer' the sound upwards. The drive units are where they are relative to each other.

    What does this all mean?

    It means that almost no domestic listening is undertaken at the correct listening height relative to the designer's RMA. Speaker stands for home use are, generally, far too short. That means that almost all domestic listening is undertaken far from the optimal (RMA) height, and the resulting sound is far from at its best, perhaps significantly so. There is likely to be a significant mis-match between the energy output from the bass/mid and the energy output from the tweeter in the crossover region, with a veiling of sound and loss of detail. The one single biggest investment that any hifi system can make, far eclipsing changing any other element in the home audio system is to raise the speakers or conversely, sit lower.

    You can see the idea here in this video I made some time ago.

    So what? Well, to showcase loudspeakers to the very maximum of their potential, regardless of all other factors of stability, cosmetics and so on, there has to be an optimal match to the loudspeaker's RMA to the ear height of a seated listener. In a small hotel bedroom, with the speakers rather close to the audience, that means the stands really must be significantly taller (50-100% or so) than consumers are accustomed to use at home. Other factors which suggest tall stands at a show are that as available seats become occupied, listeners in all but the front row are sitting in the acoustic shadow of others. In addition, with all seats taken, many/most listeners are standing, and what they hear far above RMA and what a seated listener hears a little above RMA are radically different, and very much to the disadvantage of the standing visitor.

    The problem the is not a technical one of fabricating taller stands especially for a public exhibition, it is the suggestion that it may convey that the good sound is only achievable with what may prove to be excessively tall stands for normal domestic use.


    * View the speaker cabinet from the side and you can see that the tweeter is screwed onto the baffle, the voice coil just behind the diaphragm. Conversely, the bass/mid voice coil is deep inside the cabinet, at the apex of the bass/mid cone.

  • witwald
    replied
    Importance if height fir best sound

    Originally posted by A.S. View Post
    Agreed, I missed that vital word 'acoustic'. But in reality, the situation is much less attractive than the mild disturbance you suggest around crossover frequency, even in an optimised driver/crossover combination. That says to me that the real inter-driver time-of-flight and latent crossover electrical time delays have not been modelled realistically.
    The geometric offsets and crossover phase responses have been modelled correctly. The extra disturbances that you are familiar with from your own measurements are probably due to the effects of sound diffraction from the edges of the baffle.

    But the point that you make remains entirely valid: there is only one optimum listening height for every multi-way speaker system, regardless of brand, size, shape or price. And as a general observation, but not a hard and fast rule, a) that optimised listening point (aka the reference axis) is somewhere around ear level with the tweeter or thereabouts and b) most audiophiles are listening on stands that are too short to put their ear on that plane.

    It's depressing indeed that this sort of beginner's stuff is not covered in the hifi press. Wouldn't that be a better use of ink and paper?
    Yes, that sort of basic advice does seem to be all too rarely covered in loudspeaker reviews. It makes one wonder whether the reviewers take that sort of thing into account when setting up the loudspeakers for auditioning in their listening rooms. If they can't get that aspect of the setup correct, then of what use is their "review" really going to be?

    Still, it seems that some of the recent reviews of Harbeth loudspeakers that I've read do in fact mention the importance of loudspeaker stands to get the reference axis at ear height for the (seated) listener. That's definitely a positive result. It's hard to believe that not knowing the reference axis, and then not positioning the loudspeaker's reference axis correctly, is something that is all too common, even with speakers costing multiple thousands of dollars/pounds/euros.

    Leave a comment:


  • A.S.
    replied
    Acoustic response, and general observations about stand height

    Originally posted by witwald View Post
    I'm glad it works better now.



    Not quite. Notice that I used the term "3rd-order acoustic Butterworth low-pass and high-pass responses".



    The model that I've used has assumed that the acoustic Butterworth response is the end result of the crossover network's filtering and equalisation operation on the raw responses of the woofer and tweeter drivers. I know that I assumed that this is done perfectly, but it was only meant to be a simplified model.

    The complexity of the model can be increased to incorporate the non-flat behaviour of the drive units themselves, which would then be equalised and filtered by the crossover network. To do this accurately requires a good quality model or measurement of each driver's raw complex sound pressure and impedance responses to allow correct crossover filter simulation.

    These types of effects weren't included, as my aim was to try and show what can happen when listening vertically off axis, by isolating some of the main parameters.
    Agreed, I missed that vital word 'acoustic'. But in reality, the situation is much less attractive than the mild disturbance you suggest around crossover frequency, even in an optimised driver/crossover combination. That says to me that the real inter-driver time-of-flight and latent crossover electrical time delays have not been modelled realistically.

    But the point that you make remains entirely valid: there is only one optimum listening height for every multi-way speaker system, regardless of brand, size, shape or price. And as a general observation, but not a hard and fast rule, a) that optimimised listening point (aka the reference axis) is somewhere around ear level with the tweeter or thereabouts and b) most audiophiles are listening on stands that are too short to put their ear on that plane.

    It's depressing indeed that this sort of beginner's stuff is not covered in the hifi press. Wouldn't that be a better use of ink and paper?

    Leave a comment:


  • witwald
    replied
    An "acoustic" Butterworth response was being modelled

    Originally posted by A.S. View Post
    That's much more like it!
    I'm glad it works better now.

    Now, that is the position considering only the electrical filter action of the crossover circuit.
    Not quite. Notice that I used the term "3rd-order acoustic Butterworth low-pass and high-pass responses".

    To close the gap with reality, you have to add to that model the non-flat behaviour of the drive units themselves, especially within, say, an octave of crossover frequency. That is, aside from any wiggles and squiggles in their frequency responses, the bass/mid drive behaves as a mechanical low pass filter, and the tweeter as a mechanical high pass filter. That means that when considering the acoustic picture, the summation of sound waves at the listener's ear (which is all we are ultimately concerned about) the combined acoustic + electrical filter is going to be something like 4th or 5th order in some unpredictable and irregular way. How does that model?
    The model that I've used has assumed that the acoustic Butterworth response is the end result of the crossover network's filtering and equalisation operation on the raw responses of the woofer and tweeter drivers. I know that I assumed that this is done perfectly, but it was only meant to be a simplified model.

    The complexity of the model can be increased to incorporate the non-flat behaviour of the drive units themselves, which would then be equalised and filtered by the crossover network. To do this accurately requires a good quality model or measurement of each driver's raw complex sound pressure and impedance responses to allow correct crossover filter simulation.

    These types of effects weren't included, as my aim was to try and show what can happen when listening vertically off axis, by isolating some of the main parameters.

    Leave a comment:


  • A.S.
    replied
    Now add more complexity ....

    Originally posted by witwald View Post
    I can only but agree that the model I provide was a gross simplification. It was only meant to show what might happen when listening to a loudspeaker vertically off axis when the design was created, in whatever manner, to have a flat response on the tweeter axis at the 3.0 m listening distance.

    Below are the simulations from a revised model. Here I took the acoustic centre of the woofer to be 50 mm behind that of the tweeter. The 3rd-order acoustic Butterworth low-pass and high-pass responses were left the same. The listening location was positioned 3.0 metres ahead of the tweeter on the tweeter axis. It is quite evident that there are now significant undulations in the frequency response curve of the loudspeaker. That is to be expected. A correct crossover design would of course need to address this very situation, so that the loudspeaker's designed measured response on the reference axis would be nominally flat.

    [ATTACH=CONFIG]3552[/ATTACH]
    That's much more like it! Now, that is the position considering only the electrical filter action of the crossover circuit. To close the gap with reality, you have to add to that model the non-flat behaviour of the drive units themselves, especially within, say, an octave of crossover frequency. That is, aside from any wiggles and squiggles in their frequency responses, the bass/mid drive behaves as a mechanical low pass filter, and the tweeter as a mechanical high pass filter. That means that when considering the acoustic picture, the summation of sound waves at the listener's ear (which is all we are ultimately concerned about) the combined acoustic + electrical filter is going to be something like 4th or 5th order in some unpredictable and irregular way. How does that model?

    Leave a comment:


  • witwald
    replied
    Model with real-world rearward physical offset of woofer voice coil

    Originally posted by A.S. View Post
    I strongly recommend that you introduce the real-world physical offset of the woofer coil being far behind the tweeter coil and see what a mess that makes of your otherwise entirely acceptable graphs. The problem is barely scratched by your graphs thus far! The real issue is the inter-unit physical offset, not the vertical position of the listener's ear in combination.
    I can only but agree that the model I provide was a gross simplification. It was only meant to show what might happen when listening to a loudspeaker vertically off axis when the design was created, in whatever manner, to have a flat response on the tweeter axis at the 3.0 m listening distance. As you say, the model barely scratches the complete problem.

    Below are the simulations from a revised model. Here I took the acoustic centre of the woofer to be 50 mm behind that of the tweeter. The 3rd-order acoustic Butterworth low-pass and high-pass responses were left the same. The listening location was positioned 3.0 metres ahead of the tweeter on the tweeter axis. It is quite evident that there are now significant undulations in the frequency response curve of the loudspeaker. That is to be expected. A correct crossover design would of course need to address this very situation, so that the loudspeaker's designed measured response on the reference axis would be nominally flat.

    Click image for larger version

Name:	AXO3P00.jpg
Views:	1
Size:	232.4 KB
ID:	63229

    Leave a comment:


  • A.S.
    replied
    Gross simplification

    Originally posted by witwald View Post
    I decided not to model the woofer's voice coil being well behind that of the tweeter. The model that I created is simply trying to show what will happen with vertical off axis listening versus on axis listening. For this model it effectively means that the crossover includes some time delay for the tweeter to time align its voice coil with that of the woofer at the 3.0 metre listening location on the reference measurement axis. I am also keeping in mind that the acoustic centre of the woofer may actually vary a little with frequency, so any crossover modelling needs to be backed up by accurate amplitude and phase response measurements of the drivers as they are being individually driven while mounted on the loudspeaker baffle.
    I strongly recommend that you introduce the real-world physical offset of the woofer coil being far behind the tweeter coil and see what a mess that makes of your otherwise entirely acceptable graphs. The problem is barely scratched by your graphs thus far! The real issue is the inter-unit physical offset, not the vertical position of the listener's ear in combination.

    Leave a comment:


  • witwald
    replied
    An idealised model of woofer and tweeter and filter

    Originally posted by A.S. View Post
    You need to remodel the situation. Have you forgotten to model the fact that the voice coil of the bass/mid driver is far behind that of the tweeter? If you include that fact in your model it will greatly disturb your neat graphs.
    I decided not to model the woofer's voice coil being well behind that of the tweeter. The model that I created is simply trying to show what will happen with vertical off axis listening versus on axis listening. For this model it effectively means that the crossover includes some time delay for the tweeter to time align its voice coil with that of the woofer at the 3.0 metre listening location on the reference measurement axis. I am also keeping in mind that the acoustic centre of the woofer may actually vary a little with frequency, so any crossover modelling needs to be backed up by accurate amplitude and phase response measurements of the drivers as they are being individually driven while mounted on the loudspeaker baffle.

    I recognise that the graphs greatly simplify the situation that is occurring in practice. But I believe that the vertical off axis response changes are broadly indicative of what will happen in practice. The plots I have provided not trying to model the situation that occurs with a particular set of real drivers, with their individual natural responses and roll-offs and acoustic centre positions, which add quite some complexity to the modelling situation. Neither is the off axis radiation pattern of the drivers included, which will start to affect the response in the crossover region at these sorts of vertical off axis listening conditions (at the reference axis listening location, we are a bit off axis for the woofer while being on axis for the tweeter).

    Leave a comment:


  • A.S.
    replied
    Modify your model

    Originally posted by witwald View Post
    I have performed some simulations of a woofer and tweeter to see how the summed response will vary when listening vertically off axis. The woofer and tweeter a assumed to be connected with positive polarity, and separated by a vertical distance of 150 mm. The listening position is set to be 3.0 metres away horizontally from the tweeter.

    The first set of two simulations involve the use of 3rd-order acoustic Butterworth low-pass and high-pass responses for the woofer and tweeter, respectively. The crossover frequency is chosen to be 3kHz. When listening on the tweeter axis, the response is perfectly flat. In the first plot, when the listening position is raised to be 150mm above the tweeter axis, then we see that a response peak of about 1.6dB develops, centred on about 3.3kHz. In the second plot, when the listening position is lowered to be 150mm below the tweeter axis, then we see that a broad response dip of about 2.3dB develops, also centred around 3.3kHz or so. From these two plots, it can be seen that the region of significant interaction is in the range 1.87kHz when 3rd-order Butterworth filter functions are used.

    [ATTACH=CONFIG]3547[/ATTACH] [ATTACH=CONFIG]3548[/ATTACH]

    If we reversed the polarity of the tweeter, so that it is connected out-of-phase with the woofer, then the peaks and dips would be reversed. This is purely as a result of the phase change through the crossover region.

    For comparison purposes, the second set of two simulations involve the use of 1st-order acoustic Butterworth low-pass and high-pass responses for the woofer and tweeter, respectively. Once again, the crossover frequency is chosen to be 3kHz. When listening on the tweeter axis, the response is perfectly flat. In the first plot, when the listening position is raised to be 150mm above the tweeter axis, then we see that a response dip of about 3dB develops, centred on about 7kHz or so. In the second plot, when the listening position is lowered to be 150mm below the tweeter axis, then we see that a broad response peak of about 2dB develops, which is also centred around 7kHz or so. From these two plots, it can be seen that the region of significant interaction is in the range 1.518kHz when 1st-order Butterworth filter functions are used.

    [ATTACH=CONFIG]3549[/ATTACH] [ATTACH=CONFIG]3550[/ATTACH]

    From the above plots, it is quite evident that the 1st-order Butterworth filter function produces a much greater sensitivity to vertical off axis listening location than does the 3rd-order Butterworth filter function. In addition, the response changes (in terms of peaks or dips) are reversed for a given raising or lowering of the listening point.

    I think that the above results serve to emphasise the importance of getting stands of the appropriate height to work with whatever loudspeakers that you have chosen. Anything less than that and the frequency response at the listening position where your ears are will be significantly compromised, just as Alan mentioned.
    You need to remodel the situation. Have you forgotten to model the fact that the voice coil of the bass/mid driver is the far behind that of the tweeter? If you include that fact in your model it will greatly disturb your neat graphs.

    Leave a comment:


  • witwald
    replied
    Computer simulations of vertical off axis frequency response

    Originally posted by A.S. View Post
    I'm ashamed to show you the effect on the frequency response of listening vertically off axis. It's not pretty. Huge energy suck-out.
    I have performed some simulations of a woofer and tweeter to see how the summed response will vary when listening vertically off axis. The woofer and tweeter a assumed to be connected with positive polarity, and separated by a vertical distance of 150 mm. The listening position is set to be 3.0 metres away horizontally from the tweeter.

    The first set of two simulations involve the use of 3rd-order acoustic Butterworth low-pass and high-pass responses for the woofer and tweeter, respectively. The crossover frequency is chosen to be 3kHz. When listening on the tweeter axis, the response is perfectly flat. In the first plot, when the listening position is raised to be 150mm above the tweeter axis, then we see that a response peak of about 1.6dB develops, centred on about 3.3kHz. In the second plot, when the listening position is lowered to be 150mm below the tweeter axis, then we see that a broad response dip of about 2.3dB develops, also centred around 3.3kHz or so. From these two plots, it can be seen that the region of significant interaction is in the range 1.87kHz when 3rd-order Butterworth filter functions are used.

    Click image for larger version

Name:	AX3P15.jpg
Views:	1
Size:	218.9 KB
ID:	63224 Click image for larger version

Name:	AX3M15.jpg
Views:	1
Size:	222.1 KB
ID:	63225

    If we reversed the polarity of the tweeter, so that it is connected out-of-phase with the woofer, then the peaks and dips would be reversed. This is purely as a result of the phase change through the crossover region.

    For comparison purposes, the second set of two simulations involve the use of 1st-order acoustic Butterworth low-pass and high-pass responses for the woofer and tweeter, respectively. Once again, the crossover frequency is chosen to be 3kHz. When listening on the tweeter axis, the response is perfectly flat. In the first plot, when the listening position is raised to be 150mm above the tweeter axis, then we see that a response dip of about 3dB develops, centred on about 7kHz or so. In the second plot, when the listening position is lowered to be 150mm below the tweeter axis, then we see that a broad response peak of about 2dB develops, which is also centred around 7kHz or so. From these two plots, it can be seen that the region of significant interaction is in the range 1.518kHz when 1st-order Butterworth filter functions are used.

    Click image for larger version

Name:	AX1P15.jpg
Views:	1
Size:	219.1 KB
ID:	63226 Click image for larger version

Name:	AX1M15.jpg
Views:	1
Size:	216.2 KB
ID:	63227

    From the above plots, it is quite evident that the 1st-order Butterworth filter function produces a much greater sensitivity to vertical off axis listening location than does the 3rd-order Butterworth filter function. In addition, the response changes (in terms of peaks or dips) are reversed for a given raising or lowering of the listening point.

    I think that the above results serve to emphasise the importance of getting stands of the appropriate height to work with whatever loudspeakers that you have chosen. Anything less than that and the frequency response at the listening position where your ears are will be significantly compromised, just as Alan mentioned.

    Leave a comment:


  • witwald
    replied
    Specifying the reference axis is important

    Originally posted by A.S. View Post
    If I was to give you a list of my top five recommendations for how to get the best possible sound from your system, No.1 would be to put your ear somewhere about level with the tweeter. That's not just with Harbeth speakers - it's likely to be true of most/all multi-way speakers.
    It's a pity that more loudspeaker manufacturers aren't as clear as Harbeth are in regard to specifying the reference axis. Considering how important it is for the listener to get this simple geometric relationship correct, I don't understand why this critical piece of information is usually omitted.

    Leave a comment:


  • witwald
    replied
    Already specified

    Originally posted by BOZEN View Post
    Dear Alan, could You provide RMA value in centimeters for P3ESR and M30.1?

    This could help finding the very best position in-room...
    If you take a look at the Specifications of each Harbeth loudspeaker on its product page, you will find that this is already usually given as: "Stands to bring ears level with tweeters". The height of the tweeter from the base of each speaker is also specified to help make it easier to choose a suitable stand height.

    Leave a comment:


  • BOZEN
    replied
    Reference measurement axis

    Dear Alan, could You provide RMA value in centimeters for P3ESR and M30.1?

    This could help finding the very best position in-room...

    Originally posted by A.S. View Post
    In discussing our Bristol appearance for next month, I have brought up an issue that colleagues had never thought about: the height of the speaker stands at a show compared to the listening room at home. In my humble opinion, the material and construction of the stands can be a matter of personal taste, but the height is really important, down to the last cm. or two.

    Why?

    The speaker designer has to make a decision before he starts work on the crossover where to direct his measurement microphone, the one tool that will tell him about the technical performance of the loudspeaker - the frequency response. As most loudspeaker boxes have the drive units mounted vertically along the centre line (as all Harbeth speakers do) the position of the microphone horizontally left/right of the speaker (for a fixed vertical point off the floor) would give the same measured response. However, vertically, the frequency response that the microphone measures radically alters as the microphone is raised and lowered along that cabinet centre line, vertically.

    The sound waves from the bass/mid driver and tweeter arriving at the microphone take an path through the air from their respective voice coils. As the tweeter's voice coil is far in front of the woofer's*, and far closer to the microphone when the microphones is directed towards the upper quartile of the cabinet (approximately), it's wave leads the woofer's. Conversely, when the microphone is lowered to be directed to the lower half of the cabinet, the tweeter's sound wave lags that woofer's. That lead/lag is another way of saying that the tweeter's phase relationship with the woofer depends upon where the microphone is placed relative to the woofer and tweeter vertically. All professional speaker designers must decide for themselves where to clamp the microphone vertically relative to the drive units, and to mark that vertical point onto the prototype cabinet with a marker pen as the Reference Measurement Axis (RMA).

    Then the bun fight with the marketing department begins.

    Speaker designers generally opt for two reference measurement axis; both are equally valid and in the great scheme of things only give subtly different technical results. One designer might habitually consider his RMA to be when his measuring microphone is pointed precisely at the apex of the tweeter dome. Another designer, again following his personal preferences, might measure with a ruler the vertical separation between the centre of the tweeter and centre of the bass/midrange unit, divide that in half, and mark that point on the baffle as his RMA. That would be where he would point his measuring microphone.

    Once the designer has settled in his mind where his RMA is, he can start work on the design of the crossover. One thing is sadly inevitable: when designing, he can only adjust the system crossover/equaliser/time delayer to perfectly integrate the bass/mid and tweeter sound waves as they hit the microphone diaphragm (or listener's ears) at that one fixed, vertical RMA. Not 2cms above RMA, 13cms below RMA or at any other point in space vertically. Every multi-way can only behave at its best with the measuring microphone or listener's ears exactly at that correct vertical point and nowhere else. The RMA was fixed at the design of the speaker, and the network was conceived and adjusted to compensate for the fact that the tweeter is physically (and sonically) in front of the woofer. That compensation is not universal: it is only for one point in space above the floor, wherever the RMA was set.

    And? This is where the real world and the idealistic designer clash.

    The designer sticks to his guns that the RMA is at a certain identifiable point on the baffle: 'level with the tweeter', '3cms below the tweeter' or whatever. He knows that at that point and no other vertical position that the drive units integrate as he wishes. The marketing boys are made aware of that, and make some sketches of suitable stands, and they feed back some market survey results to the designer. They say 'sorry, but that RMA you told us about for model XYZ: it's far too low down in the cabinet. We've played around with some matching stands, and gee wizz, considering that the average listener has ears that are about 1.2m above the floor when seated, listening, we're going to have to make some really, scarily, domestically unacceptably high stands to lift those speakers up high enough for your RMA to match the listener's ears ..... can't you move your RMA upwards until it's level with the top of the cabinet, or even higher!, so that we can sell short, domestically acceptable stands..?'. The answer is no, he can't 'steer' the sound upwards. The drive units are where they are relative to each other.

    What does this all mean?

    It means that almost no domestic listening is undertaken at the correct listening height relative to the designer's RMA. Speaker stands for home use are, generally, far too short. That means that almost all domestic listening is undertaken far from the optimal (RMA) height, and the resulting sound is far from at its best, perhaps significantly so. There is likely to be a significant mis-match between the energy output from the bass/mid and the energy output from the tweeter in the crossover region, with a veiling of sound and loss of detail. The one single biggest investment that any hifi system can make, far eclipsing changing any other element in the home audio system is to raise the speakers or conversely, sit lower.

    You can see the idea here in this video I made some time ago.

    So what? Well, to showcase loudspeakers to the very maximum of their potential, regardless of all other factors of stability, cosmetics and so on, there has to be an optimal match to the loudspeaker's RMA to the ear height of a seated listener. In a small hotel bedroom, with the speakers rather close to the audience, that means the stands really must be significantly taller (50-100% or so) than consumers are accustomed to use at home. Other factors which suggest tall stands at a show are that as available seats become occupied, listeners in all but the front row are sitting in the acoustic shadow of others. In addition, with all seats taken, many/most listeners are standing, and what they hear far above RMA and what a seated listener hears a little above RMA are radically different, and very much to the disadvantage of the standing visitor.

    The problem the is not a technical one of fabricating taller stands especially for a public exhibition, it is the suggestion that it may convey that the good sound is only achievable with what may prove to be excessively tall stands for normal domestic use.


    * View the speaker cabinet from the side and you can see that the tweeter is screwed onto the baffle, the voice coil just behind the diaphragm. Conversely, the bass/mid voice coil is deep inside the cabinet, at the apex of the bass/mid cone.

    Leave a comment:


  • SChat
    replied
    Motorised stands

    Originally posted by hendrik View Post
    Just ear level or has it also something to do with the distance from the floor/ceiling, position and the whole acoustical

    {Moderator's comment: The surroundings are immaterial. The only important factors are the position of the ear relative to the bass/midrange driver and the tweeter}
    I guess the last comment is for the sake of argument? Because the objective scientific reality is that soundwaves reaching the ears will NOT ONLY come from the tweeters and other drivers, there will be a fair amount of soundwaves reaching the ears after being reflected by the "surroundings" (floors, ceilings, walls).

    I have rarely seen anyone who sits still at one place like an Indian yogi while listening to music. I accept that my acquaintance with real music aficionados and audiophiles is rather limited. If alignment of ears with the RMA is so critical, what is the possible solution for a rather large population of music-lovers who suffer from a "cannot sit still at a fixed place" syndrome?

    Here is an idea for a stand designer. How about motorised mechanisms that can continuously adjust the position of the speakers along the vertical axis to keep them aligned with the listeners' ears (the technology exists in game consoles)? More advanced models can even allow alignment along other planes.

    Leave a comment:


  • Guest's Avatar
    Guest replied
    Surroundings?

    Just ear level or has it also something to do with the distance from the floor/ceiling, position and the whole acoustical surroundings ?

    Ridiculous example but if I like to listen lying on the floor should I place the speakers on the floor.. Under some circumstances I think placement of the speakers a little above ear level could make more spacious, open and natural.

    {Moderator's comment: The surroundings are immaterial. The only important factors are the position of the ear relative to the bass/midrange driver and the tweeter}

    Leave a comment:

Working...
X