davewantsmoore Posted October 28, 2014 Posted October 28, 2014 unless they have a very big amplifier. ... or sensitive speakers.
rocky500 Posted October 28, 2014 Posted October 28, 2014 @@rocky500, why the warning on tube amps? [edit: ah yes, they don't like no-load situations. Edit made] re: power ratings, maybe find a review of one's amp that includes power measurements. Reasonable amps seem to meet or exceed their rated 8 ohm power ratings when tested, in my readings of reviews over the years. I could be wrong here but I was told it could be bad for a tube amp to not have any load. (As in, not to be used unless speakers were connected)
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 There was some discussion in here regarding the ideal power response of a speaker falling gently with frequency being the ideal. As far as I understand it, I think this is subtly misrepresenting the findings of Harman. Practical and typical speakers have a declining power response - due to the shape of the drivers and cabinet. Yes - It was found that this decline should be ideally smooth (ie. that's better than ragged). However - If you could, you would have a flat power response - you would have a flat frequency response at all angles. ie. declining output with angle (but constant output with frequency) ..... not a declining output with angle and a a declining output with frequency. Eg. This. Is better than this. ... which is much better than this Thanks @@gainphile for the pics Toole: ALL of the most preferred loudspeakers are ones that exhibit the flattest, smoothest families of curves. They exhibit the fewest, and the lowest level, resonances. They have the flattest, smoothest, widest bandwidth frequency responses when measured from all angles. They have similar shapes in all of the curves – i.e. they have quite constant, or at least smoothly changing, directivity as a function of frequency. The above does not imply to me that we should seek out a smoothly declining power response .... if we have the option to have a flat power response (ie. true "constant directivity"). This is good, as it agrees with the findings of others - who find that controlling directivity as low in frequency as possible improves listener satisfaction. 1
hochopeper Posted October 30, 2014 Posted October 30, 2014 (edited) ie. declining output with angle (but constant output with frequency) ..... not a declining output with angle and a a declining output with frequency. I don't dissagree. But I think there are two subtle points missed: a] I think the consideration of behaviour in modal region should be done with a different set of criteria, especially once we're below frequencies at which human's can determine the direction of a source. b] there is a threshold for what is Good Enough. So we shouldn't set an absolute target of the first option. The threshold for human perception is likely less uniform than that. Little research has been done in establishing the lower frequency bound for which constant directivity is important, for humans. I don't think that it is 0Hz though. c] (now thinking out loud) I wonder if the falling power response preference is also a related to the freq response of the acoustic absorption in the listening room? Edited October 30, 2014 by hochopeper
Guest Posted October 30, 2014 Posted October 30, 2014 (edited) However - If you could, you would have a flat power response - you would have a flat frequency response at all angles. On of the first speaker systems I built in the early '70s was a deliberate attempt at an omnidirectional enclosure. What I found was the there was a sense of spaciousness in the room that wasn't dependent on listening position. But that was the only positive. The problem is that real acoustic sources, such as human beings and musical instruments, do not radiate high frequencies in all directions. In fact, when you measure the power response of real audio sources in an echo chamber, MOST acoustic sources have a falling power response as frequency rises that fits within generalised upper and lower limits. For a speaker to realistically mimic a real acoustic source, one aspect of design is that the speakers power response fits somewhere in the middle of these two limits. This research was first brought to my attention by John Dunlavey of Duntech fame in the mid 1980s when I was granted free access to his test facilities at Technology Park in Adelaide for the purpose of developing professional loudspeakers used in the performing arts industry. I mentioned in an earlier post an extreme, but common, example of the human voice. Humans do not generally radiate any sibilance except in the forward direction (excepting the possibility someone has a hole in the back of their neck). An omnidirectional speaker will spray sibilance all around the room creating a myriad of reflections that make the voice reproduction quite unnatural. With bipolar speakers, much of the fiddling with position and angle is to minimise the effect of their unfortunate property to do the same. Speakers that do not have a declining power response with frequency are very fussy about the room absorption and room placement. On the other hand, speakers that have an appropriately tapered, smoothly declining power response are quite tolerant of less than ideal room acoustics and much easier to live with. Edited October 30, 2014 by Guest
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 John. I'm enjoying talking to you (Hi everyone else, heh ) omnidirectional Is a form of constant directivity, yes .... however an omni radiator has the lowest ratio possible of direct to reflected sound. Increasing the ratio of direct to reflected sound enough leads to increased clarity. This is quite dependant on room size. The problem is that real acoustic sources, such as human beings and musical instruments, do not radiate high frequencies in all directions Yes.... we do not want to radiate sound in all directions (like an omni) ..... as this (a too low direct vs reflected) interferes with the way stereo works ..... we want to radiate the same sound balance (ie flat frequency response) in all directions. OTOH - to manipulate the direct to reflected balance (which is different from the power response)..... the sound radiated away from the listen should be reduced in level but it should remain flat. In fact, when you measure the power response of real audio sources in an echo chamber, MOST acoustic sources have a falling power response as frequency rises Sure... Just like speakers.... the size affects the amplitude vs frequency vs angle. For a speaker to realistically mimic a real acoustic source, one aspect of design is that the speakers power response fits somewhere in the middle of these two limits. I have never heard a case made for why a speaker should mimic a "real acoustic source". The fall in power response for the indirect sound that the original sources (may) have is already contained in the recording. I do not see a case for augmenting that by distorting the offaxis sound (non-constant directivity is distortion). This research was first brought to my attention by John Dunlavey of Duntech fame in the mid 1980s when I was granted free access to his test facilities at Technology Park in Adelaide for the purpose of developing professional loudspeakers used in the performing arts industry. Well, we can tell (by looking at his driver layouts and crossover design) .... that he is also a great proponent of looking after the offaxis sound of a speaker. However I can't escape thinking that if he is making the claim that a "falling power response" is desirable .... that it must be in a similar way to the "misrepresentation" of Toole that I referred to. If we take a typical box speaker (like one of his speakers).... You cannot escape a falling power response. The speakers "baffle step" is fixed .... and the speaker cabinet does not change it's dimensions (assuming it is a square box) to remain acoustically narrow with frequency ...... so the closest you can get to "constant directivity" is a gently falling power response. Just like Toole saw.... if we have a bunch of box speakers - then ultimately the constant directivity we can achieve is limited by their size .... the best you can do is make the power response fall gradually and smoothly. Yes, Toole has includes an example or 3 of poorly engineered dipole radiators in his data (ones that might control radiation pattern down to a low frequency .... but that have large errors in the off axis response - and poor integration with their bass sources)... When I talk about a dipole - I mean a well engineered one like the latest from Linkwitz or Kreskovsky ..... something that looks like this I mentioned in an earlier post an extreme, but common, example of the human voice. Humans do not generally radiate any sibilance except in the forward direction (excepting the possibility someone has a hole in the back of their neck). An omnidirectional speaker will spray sibilance all around the room creating a myriad of reflections that make the voice reproduction quite unnatural. With bipolar speakers, much of the fiddling with position and angle is to minimise the effect of their unfortunate property to do the same. Speakers that do not have a declining power response with frequency are very fussy about the room absorption and room placement. On the other hand, speakers that have an appropriately tapered, smoothly declining power response are quite tolerant of less than ideal room acoustics and much easier to live with. Are you confusing "falling power response" .... with differences in the direct to reflected sound ratio ?! They aren't the same thing. I would agree that a speaker with a low direct to reflected sound ratio (the extreme example is an omni) is fussy about placement. A dipole is fussy about placement due to its rear radiation. It must be placed far enough away from the rear boundary to ensure that it's interaction with the backwall is low enough in frequency (to be below the schrodinger frequency) .... however it does conversely have an advantage sideways (due to it's null). I find it striking (for the relevance of constant directivity) .... That a dipole and a horn - can have such similar sound signatures (and such startling subjective increase in quality).... They have quite a few difference - especially in direct vs reflected sound (dipole has rear radiation) .... however - due to the nulls at the side of a dipole - their direct to reflected ratios are more alike than intuition would have us believe (but still different). speakers that have an appropriately tapered, smoothly declining power response are quite tolerant of less than ideal room acoustics and much easier to live with. Sure..... I don't disagree..... The only point that I am labouring on is that we should take it further. We should aspire for a truly flat power response - truly constant directivity (not possible for a typical box speaker) ..... That they are even MORE tolerant of room acoustics. Making the reflections closer to the direct sound... and reducing the direct to reflected sound ratio (important as the room gets smaller) ---- leads to higher listener scores. As I've said before. Discounting an omni radiator .... I only know of two approaches to bring the constant directivity down to frequencies where the room takes over ---- and (importantly) consequently to increase the ratio of direct to reflected sound in this region ("typical" home speakers are all omnidirectional in the forward hemisphere at 500hz) Guide the waves to go where you want them ..... with very large cones and/or horns Destroy the waves that go where you don't want them .... by using dipole interference 2
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 (edited) Gah. Schroeder !!!! schrodinger frequency I suspect I'll never stop having trouble with this one. Schrodinger was the guy with the cat. I'm looking for Schroeder .... the kid from Snoopy. Edited October 30, 2014 by davewantsmoore
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 c] (now thinking out loud) I wonder if the falling power response preference is also a related to the freq response of the acoustic absorption in the listening room? Rooms absorb high frequencies more than lows. So do you mean that the speaker is attempting to "mimic" the frequency response distortion of a room? ..... cos if it were attempting to counteract it, then it would need to be the reverse.
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 Not missed.... but perhaps not driven home hard enough a] I think the consideration of behaviour in modal region should be done with a different set of criteria, especially once we're below frequencies at which human's can determine the direction of a source. Definitely! In the modal region, there is no directivity .... so our discussion about radiation pattern is essentially moot. In fact, in the "whole scheme of things" .... You could even say that sound in the modal region has little to do with the source (speaker) at all. b] there is a threshold for what is Good Enough. So we shouldn't set an absolute target of the first option. The threshold for human perception is likely less uniform than that. Little research has been done in establishing the lower frequency bound for which constant directivity is important, for humans. I don't think that it is 0Hz though. It cannot be zero. There is only directivity above a certain frequency. Testing this is very difficult .... as it is impractical to have a speaker which could change it radiation pattern on the fly (although that it is an extremely interesting proposition). People who do claim to have tested it (eg. Danley or Geddes) .... report that controlling the directivity as low as possible in frequency = drastic subjective improvement in sound quality. 1
hochopeper Posted October 30, 2014 Posted October 30, 2014 From Geddes' recent 'white-paper' on speaker directivity (he calls it a white paper which means the marketing subtext is that it is also a pitch for the design choices in his Summa) there is some good information: 1. Hence, somewhere between 200 Hz and 1 kHz the directivity needs to start to become narrower and 2. Figure 10. as a more realistic target polar map than the 0-20k constant directivity map.
hochopeper Posted October 30, 2014 Posted October 30, 2014 (edited) Rooms absorb high frequencies more than lows. So do you mean that the speaker is attempting to "mimic" the frequency response distortion of a room? ..... cos if it were attempting to counteract it, then it would need to be the reverse. My thought bubble here (and it's not much better than a thought bubble really) is that perhaps ears/humans are sensitive to the reflections/reverb of a few hard surfaces that may increase the reflected high freq energy and that masking those by reducing the direct energy is perceptually preferred over than maintaining flat on-axis. Edited October 30, 2014 by hochopeper
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 (edited) Depends what you call high frequencies..... High frequencies are absorbed quite readily by the air and room (even a quite live one) ... and speakers typically all narrow their dispersion pattern as you go up very high in frequency. Where I think you are on the right track, is in the midrange (and lower MF).... where typical rooms don't absorb a ton of sound - you don't even need "hard surfaces" to have trouble (but making them absorb the sound is not the solution - they sound dead) If you look at the first 2 pictures I posted (this is better than this).... The ideal (theoretical) pattern reduces the energy off axis in the 200 to 1000hz range VS the other (which is still a quite nice coverage pattern BTW) This is an extremely important range of frequencies (fundamental tones). Too much energy in this range over too wide an angle, leads to reduced clarity (as John raised with his extreme example > an omni-directional radiator) Edited October 30, 2014 by davewantsmoore
hochopeper Posted October 30, 2014 Posted October 30, 2014 (edited) Schrodinger Edited October 30, 2014 by hochopeper
Guest Posted October 30, 2014 Posted October 30, 2014 I have never heard a case made for why a speaker should mimic a "real acoustic source". The fall in power response for the indirect sound that the original sources (may) have is already contained in the recording. Sorry, but no the fall in power response for the indirect sound that the original sources (may) have is not already contained in the recording, not at all. A recording of an acoustic performance attempts to capture the direct sound field plus the room response. To accurately sample the power response would require at the very least six microphone - two diametrically opposed on each axis - around each acoustic source that was being recorded. I have never seen this done, and if it were it would sound very unnatural when played back through headphones or speakers. An speaker has to mimic the power response of the source so that when you play the recording of the source through the speaker, the speaker excites the room in a similar many to wha the source would if the source were in the room. Otherwise the speaker (in a real room) will sound unnatural.
hochopeper Posted October 30, 2014 Posted October 30, 2014 Sorry, but no the fall in power response for the indirect sound that the original sources (may) have is not already contained in the recording, not at all. A recording of an acoustic performance attempts to capture the direct sound field plus the room response. To accurately sample the power response would require at the very least six microphone - two diametrically opposed on each axis - around each acoustic source that was being recorded. I have never seen this done, and if it were it would sound very unnatural when played back through headphones or speakers. An speaker has to mimic the power response of the source so that when you play the recording of the source through the speaker, the speaker excites the room in a similar many to wha the source would if the source were in the room. Otherwise the speaker (in a real room) will sound unnatural. How does this concept extend to a multi-mic recording ? How does the room response not also capture the polar behaviour of the acoustic event being recorded? (ie the polar response of the instrument and its interaction with the room where the performance was made) I'm with Dave, the polar response used for REproduction has no relationship with the polar response of the acoustic source that is being recorded.
Guest Posted October 30, 2014 Posted October 30, 2014 How does this concept extend to a multi-mic recording ? I am not sure if you mean multi-mic'ing an instrument, like a piano, or multi-mic'ing an ensemble, like an orchestra. In the case of the former, the objective is to capture the approximate sound (including spectral balance) that an audience member would hear in the direct sound field. In the case of the latter, it is the same objective for each acoustic source/instrument, so the number of microphones required will depend on the characteristics of each individual source. The room sound would normally be captured separately. If the microphones were placed to get the same balance of direct versus room as an audience member, when the recording is played back on a stereo in another room, it will sound ridiculously "roomy" and awash with reverberation. (The exception to that is in an in-ear microphone set-up played back through headphones, which is a technique often used in quality assurance for acoustic design of auditoria.) How does the room response not also capture the polar behaviour of the acoustic event being recorded? (ie the polar response of the instrument and its interaction with the room where the performance was made) The room response is, of course, the result of the source's power response imposed on the room. That's why there is very little energy in the reverberant sound field above ~4 kHz. The problem with listening in a room is that you are adding room sound to a recording that already has room sound. As it happens, it doesn't really matter whether the objective of stereo replay is to recreate the impression that a performer is in the room, or if you are listening to something that is completely synthesised electronically. If your speaker system has a"flat" power response, the room's response will be far too bright to be "natural". Even dipoles or bipoles have a power response that falls with frequency. I would say that at best, dipoles and bipoles sit at the upper bound of natural power response of typical real acoustic sources. But physics - gets you every time.
Guest Posted October 30, 2014 Posted October 30, 2014 John. I'm enjoying talking to you I mean a well engineered one like the latest from Linkwitz or Kreskovsky ..... something that looks like this Thanks. It's good to have a lucid conversation. Do you have or have you listened to Linkwitz's Orions?
Newman Posted October 30, 2014 Posted October 30, 2014 (edited) I have (listened). BTW tackle Matho at your peril, boys. One very cluey chap has entered the room. (Welcome John!) Edited October 30, 2014 by Newman
Guest Posted October 30, 2014 Posted October 30, 2014 (edited) Thanks Newman. It's great to participate in a conversation about audio. The last 40 years has been compartmentalised into professional audio periods, when I have ignored the personal audio world, and periods when I have been involved in high fidelity pursuits. Right now, thanks to a redundancy "offer" at Christmas last year, I am out of sorts, so to speak - I decided not to seek another position as an electroacoustical consultant. The professional audio arena has gone "mad" with everything going IP based - sending audio data packets through random distribution paths is not how I would choose to construct any audio system, especially when the transport is going to be obsolete in 5 years max. I could not face going back into a retail store run by the stereotypical Hi Fi shop owner, and I have no desire to go back into business myself, as much as I enjoyed having my own hi fi shop. So here I am doing hi fi repairs and upgrades at home and learning where everyone is at on this forum. Edited October 30, 2014 by Guest
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 Sorry, but no the fall in power response for the indirect sound that the original sources (may) have is not already contained in the recording, not at all. A recording of an acoustic performance attempts to capture the direct sound field plus the room response. Sure.... and the 'power response' ... the 'dispersion pattern' of the sources is contained in the 'room response'. We can argue semantics... but I not convinced that a speaker should do what you suggest. The offaxis response of a typical speaker (on with falling power response) is high a low frequencies, and drops at high frequencies..... This is distortion. The proof is in the pudding. It is obvious when this distortion is removed
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 An speaker has to mimic the power response of the source so that when you play the recording of the source through the speaker, the speaker excites the room in a similar many to wha the source would if the source were in the room. Otherwise the speaker (in a real room) will sound unnatural. How can you know what dispersion pattern (power response) the original source had ? ..... There may be some tolerance for "real sources" ... but this isn't universal. The source will have a different pattern not only between different recordings, but within the SAME recording (different instruments, etc.) ..... there's a world full of other program material too.\ Like I said... .I don't think the reason you give is the reason why we should have a smooth (and gently falling) power response. It is very desirable ---- but one which doesn't fall at all (very difficult) is even better. If you had of said we should all have speakers with a gently falling power response because all the recording artists have the same (ie. that's what they listen to their recordings on) .... Then you might have something. As an aside, this is where I firmly believe that recordings that are not "mucked about with" too much are the best --- purely because the people doing the mucking are doing it through compromised speakers. Otherwise the speaker (in a real room) will sound unnatural. Why is it then that there is almost universal approval that speakers with even closer to constant directivity sound excellent ?!.
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 The problem with listening in a room is that you are adding room sound to a recording that already has room sound. Exactly!!! This is the reason why going even further than just a a "gently falling power response" .... to one which is flat ..... is good. You get less "room sound".... and the room sound you do get, is a closer copy to the direct sound (meaning you don't hear it as much).
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 Thanks. It's good to have a lucid conversation. Do you have or have you listened to Linkwitz's Orions? Gah... I just lost a long post. I have built a very similar speaker to the Orion in the past when I was learning dipoles.... Essentially I bought the Orion license but used different drivers - and learned the craft\ ..... but more recently I have built the next generation of the Orion from Linkwitz.... the LX521. I never took it past a prototype, as I settled instead on the speaker Linkwitz was accused of copying the Nao Note II RS (I posted a polar of it earlier) http://www.musicanddesign.com/NaO_Note_II_RS_Details.html When I say 'comparing dipole to horn' .... I'm comparing the IIRS (equivalent to next gen Orion) with a big (90x60cm) horn which covers from 250hz up.
davewantsmoore Posted October 30, 2014 Posted October 30, 2014 I have What did you think? Orion has some pretty large errors in it's off axis response which Toole would scowl at. The next generations are shaped so funny so they can remain narrow (vs frequency) and the result is much much better.
Recommended Posts