Heya,
There are many different opinions in these circles on what the ideal directivity profile in a home loudspeaker should be (and the extent to which it is even important in the first place).
Irreconcilable theories abound, including:
- a speaker's directivity profile is only marginally important
- constant directivity (at least above a certain frequency) is an essential part of any good loudspeaker (e.g. Gedddes)
- omni-directivity is best (e.g. MBL)
- controlled directivity with a downward-sloping off-axis response is ideal (e.g. le Cleac'h)
Although I don't even hope to begin a thread that settles this discussion, it did get me thinking about the polar responses of real acoustic events (e.g. instruments, the voice) and their relevance (if any) to the polar response of the speakers on which they are reproduced.
It seems to me that there must be something relevant to be learnt here.
So in this vein, I'd like to hear people's thoughts on the following:
Can and should the question of a speaker's ideal directivity be informed by understandings of the directivity profiles of the real acoustic events it (often) seeks to reproduce?
And have good measurements ever been posted/published of polar responses for typical musical instruments (including the human voice)?
Without having taken any measurements of musical instruments, voices, etc. myself, I can only speculate on what the polar responses might be, and I'd expect that the range of different polar responses would be very large: think of the difference between a snare being hit - sound shooting off almost uniformly in every direction (other than downwards) - and compare that to what I imagine would be the very controlled directivity sound coming out of a trumpet (at least above a certain frequency determined by the width of its bell).
Other instruments may be even more complex. Indeed, it would be very unlikely that the polar response profile of a speaker could ever resemble that of all the sounds on most recordings.
But is there nevertheless anything we can learn as speaker designers from thinking about these questions?
Thanks,
Andreas
There are many different opinions in these circles on what the ideal directivity profile in a home loudspeaker should be (and the extent to which it is even important in the first place).
Irreconcilable theories abound, including:
- a speaker's directivity profile is only marginally important
- constant directivity (at least above a certain frequency) is an essential part of any good loudspeaker (e.g. Gedddes)
- omni-directivity is best (e.g. MBL)
- controlled directivity with a downward-sloping off-axis response is ideal (e.g. le Cleac'h)
Although I don't even hope to begin a thread that settles this discussion, it did get me thinking about the polar responses of real acoustic events (e.g. instruments, the voice) and their relevance (if any) to the polar response of the speakers on which they are reproduced.
It seems to me that there must be something relevant to be learnt here.
So in this vein, I'd like to hear people's thoughts on the following:
Can and should the question of a speaker's ideal directivity be informed by understandings of the directivity profiles of the real acoustic events it (often) seeks to reproduce?
And have good measurements ever been posted/published of polar responses for typical musical instruments (including the human voice)?
Without having taken any measurements of musical instruments, voices, etc. myself, I can only speculate on what the polar responses might be, and I'd expect that the range of different polar responses would be very large: think of the difference between a snare being hit - sound shooting off almost uniformly in every direction (other than downwards) - and compare that to what I imagine would be the very controlled directivity sound coming out of a trumpet (at least above a certain frequency determined by the width of its bell).
Other instruments may be even more complex. Indeed, it would be very unlikely that the polar response profile of a speaker could ever resemble that of all the sounds on most recordings.
But is there nevertheless anything we can learn as speaker designers from thinking about these questions?
Thanks,
Andreas
I don't think there is much to be learned from instruments radiation patterns because all that is being 'screwed up' by the recording process in between.
You would have to take into account the pick up pattern of the microphones used, their distance from the sound source, the recording space with its acoustic properties and possible use of artificial reverb.
You would have to take into account the pick up pattern of the microphones used, their distance from the sound source, the recording space with its acoustic properties and possible use of artificial reverb.
This question is at the crux of the directivity battles and in my opinion, the reason there is no agreement. Different instruments exhibit different polar patterns and will be more accurately reproduced by speakers that exhibit similar patterns.
Since no speaker can match every instrument's polar response, you end up having to choose - monopole, dipole, omni, high directivity - each of which sound great when their pattern matches that of the instrument it is reproducing. Apart from a polar pattern match, there is a sonic signature to each of the above patterns that some may prefer over others.
I have noticed recently that small dome tweeters (19-22mm) sound different to me than larger domes (27-29mm). There is more air and spaciousness, which I assume is at least partially due to the wider dispersion of the smaller dome. I prefer this on most tracks, but on some tracks it is distracting and would probably prefer a more controlled directivity. On some tracks, in some rooms, I prefer the sound of my waveguide loaded domes, or sometimes the dipoles. At the end of the day, I think most people prefer their seaker to simulate the radiation pattern of the instrument being reproduced. When that can't happen, it's just a matter of which pattern you prefer the sound of.
And that's the beauty of DIY. You can make speakers that, more often than other speakers, cater to your room, music collection and listening preferences.
Dan
Since no speaker can match every instrument's polar response, you end up having to choose - monopole, dipole, omni, high directivity - each of which sound great when their pattern matches that of the instrument it is reproducing. Apart from a polar pattern match, there is a sonic signature to each of the above patterns that some may prefer over others.
I have noticed recently that small dome tweeters (19-22mm) sound different to me than larger domes (27-29mm). There is more air and spaciousness, which I assume is at least partially due to the wider dispersion of the smaller dome. I prefer this on most tracks, but on some tracks it is distracting and would probably prefer a more controlled directivity. On some tracks, in some rooms, I prefer the sound of my waveguide loaded domes, or sometimes the dipoles. At the end of the day, I think most people prefer their seaker to simulate the radiation pattern of the instrument being reproduced. When that can't happen, it's just a matter of which pattern you prefer the sound of.
And that's the beauty of DIY. You can make speakers that, more often than other speakers, cater to your room, music collection and listening preferences.
Dan
What comes to polar response and types of it - it is good that we have options for different rooms and "tastes of sound" I have dipoels in my living room and waveguide-coaxials in a mancave. Thed do sound diferent but suit the tasks this way very well.
Some people think that this is a beautiful speaker, but not me!
Some people think that this is a beautiful speaker, but not me!
Yeh, I think this is a valid perspective. Definitely, these factors all contribute significantly to "screwing up" (for better or worse) the original acoustic event.
Although I think for close mic'd recordings with minimal or no artificial reverb, most of these factors might end up playing a more secondary role.
Here's the more specific aspect of this question that got me thinking about it in the first place:
Some designers advocate constant-directivity designs, while others advocate designs that provide a kind of controlled directivity that gives a downward sloping polar response (i.e. where directivity increases with frequency).
Do we have examples of real-world instruments (including the human voice) with polar patterns corresponding to either of these two design goals?
Some designers advocate constant-directivity designs, while others advocate designs that provide a kind of controlled directivity that gives a downward sloping polar response (i.e. where directivity increases with frequency).
Do we have examples of real-world instruments (including the human voice) with polar patterns corresponding to either of these two design goals?
Last edited:
^This is actually irrelevant, because the recordings are made to be reproduced with typical monopole loudspeakers.
Instruments' radiation has been investigated well
Musical Instruments' Sound Radiation Patterns
https://users.aalto.fi/~ktlokki/Publs/patynen_aaua_2010.pdf
Instruments' radiation has been investigated well
Musical Instruments' Sound Radiation Patterns
https://users.aalto.fi/~ktlokki/Publs/patynen_aaua_2010.pdf
Speaking of this, and ugly speakers, the latest Wilson's add a rear facing mid (maybe a tweeter too).
Rear-facing drivers have been used to juice the delayed arrivals for a while. Besides Bose, of course, I know Snell did this, to good effect. But "effect" it most certainly is.
As Charles points out, the recording process already had direct and reflected information recorded in 2-channel format. Anything we do like reflecting, or multi-speaker DSP, is an effect, not "more accurate" reproduction. Make yourself happy.
Best,
E
Rear-facing drivers have been used to juice the delayed arrivals for a while. Besides Bose, of course, I know Snell did this, to good effect. But "effect" it most certainly is.
As Charles points out, the recording process already had direct and reflected information recorded in 2-channel format. Anything we do like reflecting, or multi-speaker DSP, is an effect, not "more accurate" reproduction. Make yourself happy.
Best,
E
If you close mic an instrument its radiation pattern is completely irrelevant though, it only plays a role when using ambient/distant mic techniques.
Also bear in mind the abilities and limitations of stereo.
At its best stereo can give you a window into the recording venue* but it will never be able to turn ones living room into a concert hall. For that you need dummy head recordings and headphones.
*by this I mean the optimum result sounds as if your room has an open window to the venue in between your speakers. It will still sound like you are in your room.
I'm not convinced of this. The reason is that you cannot only listen to the recording when played through loudspeakers, but are forced to listen to the recording PLUS the "room sound" which I mean to be the sum of the reflections from the room at the listening position over time until the reverberent field has died out. The music is continually generating these "extra" sounds in the room. So you are NEVER listening only to the recording!As Charles points out, the recording process already had direct and reflected information recorded in 2-channel format. Anything we do like reflecting, or multi-speaker DSP, is an effect, not "more accurate" reproduction. Make yourself happy.
The important question IMO is how to make the room sound pleasant, meaning to as much as possible not detract from the recording. For answers to the questions of how to do this we can turn to psycoacoustics and learn something about how the brain interprets/processes the stimuli that the ears give it and turns that into "hearing" in the brain. This is not such a simple process, really. The brain has a lot of "heuristics" and so on that it applies to what it gets via the auditory nerve.
So how does this relate to loudspeakers? Considering the loudspeaker and the room is important, since they combine into the "listening experience". You cannot just say "horns are best" or "dipoles are best" without the room part of the equation being characterized as well. Sadly that is rarely done, so most of the discussions about different types of radiation patterns from loudspeakers are difficult to sort out in a scientific sense.
Getting back to the OP questions regarding the polar response of instruments, I have seen for example the radiation pattern of a seated violin player as a function of frequency (it changes with freq.). But in a stereo recording what you are capturing is not the instrument but again the sum of the instrument and the reflections of its acoustic output in the space within which the recording was made as picked up by one or more microphones placed in different locations, using different pickup patterns, etc. So the polar response of the instrument is really not a factor and should in no way be considered in the design of loudspeakers. It SHOULD be considered in the placement of microphones for the recording of the instrument, but that is not the question here.
I think people are oversimplifying this. Yes there is added information in the recording and no a polar response is not recorded, but these things only detract from perfection along with a million other variables. A louspeaker that has a similar polar pattern to the instrument being reproduced will get you closer to perfection than if not so. Empirically, I have found this to be one of the largest contributing variables to realism in sound reproduction. It's difficult to identify because most recordings have several instruments playing at once, all played through the same loudspeaker. If we're just going to say it doesn't matter because the recording isn't perfect, I think we are throwing the baby out with the bath water.
Dan
Dan
So lets say I have a stereo recording of a solo cello.
In which way is the central phantom image of the recorded cello improved by having two speakers to either side which each radiate sound like a cello?
I have no problem to see an advantage in this specific case if the recording were mono and only one speaker used for replay but in stereo it looks like would just create a mess and the radiation pattern of both speakers together would in no way be similar to that of one cello.
Could be great for listening to these though:YouTube ;-)
Once again, the stereo phantom image is less perfect than a mono reproduction in terms of realism of a single instrument. It's just a compromise we prefer because it tricks our brain into hearing different locations for the instruments even though all sounds are eminating from only 2 locations. All the way from the recording process through playback protocols to our amps and speakers, technology has made good compromises to get the illusion as close as we can with a certain level of convenience.
Sound waves in a room are always a mess, it's just that some messes our brains identify as one thing and some messes another because of the patterns we perceive. No matter what the polar response of the speaker, each one of a stereo pair is going to produce a cello sound and each will bounce all over your room at varying frequencies/intensities to give your brain added information about the direct sound (and your room, but your brain calibrates pretty quickly to a room). Your brain will identify some of those messes as closer to the original than others and with all else being equal, the closer the radiation pattern to the original, the more readily the brain will identify the pattern and suspend disbelief for your listening pleasure. Even if the reality is that the cello is being reproduced in two speakers at once.
Dan
Sound waves in a room are always a mess, it's just that some messes our brains identify as one thing and some messes another because of the patterns we perceive. No matter what the polar response of the speaker, each one of a stereo pair is going to produce a cello sound and each will bounce all over your room at varying frequencies/intensities to give your brain added information about the direct sound (and your room, but your brain calibrates pretty quickly to a room). Your brain will identify some of those messes as closer to the original than others and with all else being equal, the closer the radiation pattern to the original, the more readily the brain will identify the pattern and suspend disbelief for your listening pleasure. Even if the reality is that the cello is being reproduced in two speakers at once.
Dan
Thanks, these studies are exactly what I was looking for.
I think "irrelevant" is too far. They might be made for monopole loudspeakers, but monopole loudspeakers with what kind of polar response?
Last edited:
The polar response of speakers used in a well treated studio environment is not as important as in our living room. Look what happens at the listening spot. That's what influences our perception of different types of speakers.
The radiation pattern of instruments would be the last thing I would copy with a speaker. But some even try to copy the reverberant nature of instruments with speaker enclosures. If one were able to use a speaker per instrument it could be fun though. But you'd still need the room to match the hall they played in.
I bet this route can still be a lot of fun though, as every speaker pattern you try will change it's interaction with the room and thus have different ambience etc. But it won't be a solution to get closer to the recorded sound i.m.h.o.
The radiation pattern of instruments would be the last thing I would copy with a speaker. But some even try to copy the reverberant nature of instruments with speaker enclosures. If one were able to use a speaker per instrument it could be fun though. But you'd still need the room to match the hall they played in.
I bet this route can still be a lot of fun though, as every speaker pattern you try will change it's interaction with the room and thus have different ambience etc. But it won't be a solution to get closer to the recorded sound i.m.h.o.
But there's a HUGE difference in the reflected sounds coming from the main speakers in front of you and from them coming from the sides of you or behind you (or above you, if Atmos is to be believed). Hearing is very directional itself, not just from stereo and response shaping but from constant head movement of the listener. Reflected sounds coming from the same direction as the direct sound don't at all give the same illusion of envelopment.
True, but I'd go a little further. Given a stereo recording I think you even NEED room sounds of some kind to give any illusion that you're actually in a room where music is being performed. From limited experimentation, I'd say something delayed around 10msec at least and coming from behind or the sides seems to help a lot. Would it be better if it originated from sources with the patterns of the original instruments? Maybe/probably? But "stereo" is the first assumption here, so I don't see how it would even be possible except for maybe a stationary solo instrument.
___
On a somewhat related subject, I've been looking into the "polar pattern" at the other end -- the effects of chairs and surfaces near the ears that are doing the listening. That started because I noticed that when I relax in my recliner and let my head go all the way to the headrest, the feeling of envelopment and the timbre (particularly of voices) goes all to heck. I don't know (yet) whether it's a frequency response effect doing it, or some kind changes of interaural coupling between the two ears, or maybe even a body effect of positioning the head back. Or maybe even imagination, I suppose.
No recliners are used in concert halls or likely in the recording/mixing/mastering rooms either, so if they have effects they aren't likely to be in the direction of perceived accuracy. Seems crazy to get all fussy about a few dB wiggles of speaker response or reflection effects from stuff near the speaker but then ignore the stuff near the listener -- primarily the chair he's in, and perhaps a table or footrest in front.
Last edited: