I'm going to be building a pair of 2-way speakers (with some drivers that I already have), and it is looking like l'll want to use them for nearfield listening too. I think this may be okay; the bass drivers are the scanspeak 15w/8530k01, which have a fairly modest diaphragm diameter of 4.3", and the D2608/913000 tweeters are going to be in Pellegrene waveguides. I feel this combo integrates fairly well already, and I'll also be using active crossovers and bi-amping so there is good scope for upstream tweaking of the driver responses.
I've not ventured into nearfield stuff before. As I understand it, the reduced contribution from the room will mean the bass should be made to roll off less, and small differences in time-alignment and also tweeter-to-bass distance become proportionally more critical. So, I'm starting to think that this project will become more involved with DSP (probably computer-based) than I'd originally intended and that also the baffle might be made with a bit more attention to detail.
But designing for nearfield is new to me; am I thinking along the right lines and/or is there anything else that I should consider or look out for with the design?
Thanks,
Kev
I've not ventured into nearfield stuff before. As I understand it, the reduced contribution from the room will mean the bass should be made to roll off less, and small differences in time-alignment and also tweeter-to-bass distance become proportionally more critical. So, I'm starting to think that this project will become more involved with DSP (probably computer-based) than I'd originally intended and that also the baffle might be made with a bit more attention to detail.
But designing for nearfield is new to me; am I thinking along the right lines and/or is there anything else that I should consider or look out for with the design?
Thanks,
Kev
My definition is not too specific at this stage, as this method of listening is new to me; I've much to learn and discover about it. However, as a starting point I'm envisaging the speakers quite close to the listening position and not very close to walls - maybe a meter (or two at most) from the ears.
Perhaps more usefully (for people who know what they're talking about!) my aims or aspirations are all somewhat related: to reduce the SPLs needed from the speakers, to reduce the sound levels, gains and vibrations kicking around the room, and to reduce the effects of the room on the sound perceived by me the listener.
The situation is a modestly sized room with limited options for customisation and optimisation, plus thin walls and a neighbour who works nights. Headphones are going to be part of the strategy, but I enjoy speakers and so am thinking (basically) closer and quieter will let me use them more often whilst also being less influenced by the sub-optimal room. If this is naive or flawed thinking though, hopefully someone more experienced might correct me!
Thanks,
Kev
Perhaps more usefully (for people who know what they're talking about!) my aims or aspirations are all somewhat related: to reduce the SPLs needed from the speakers, to reduce the sound levels, gains and vibrations kicking around the room, and to reduce the effects of the room on the sound perceived by me the listener.
The situation is a modestly sized room with limited options for customisation and optimisation, plus thin walls and a neighbour who works nights. Headphones are going to be part of the strategy, but I enjoy speakers and so am thinking (basically) closer and quieter will let me use them more often whilst also being less influenced by the sub-optimal room. If this is naive or flawed thinking though, hopefully someone more experienced might correct me!
Thanks,
Kev
That seems a useful description of what you want. I'll start the ball rolling by saying that sources of diffraction eg on the speaker face will approach from a greater angle when you are close.
Great, thank you very much! Differences in diffraction weren't something that I'd even considered; it does seem to be logical though.
Cheers,
Kev
Cheers,
Kev
I got a rare day off today, so was able to do a bit of testing with my existing 2-ways placed in (my interpretation of) near-field. First and foremost, I'm actually excited by some of the results, which is quite a surprise. I expected it to be about making the best of a poor situation, but at best it seems very good indeed - quite a good soundstage, and I'm hearing aspects of the music that had passed me by before.
Though almost everything seems more critical. The positioning took hours to arrive at a particularly good setup, and the perceived stereo imaging still varies in success between different recordings. The time/distance mis-alignment between tweeter and woofer is quite noticeable too, as is the reduction in apparent bass (without much spl or room reinforcement). I was also quite surprised by how physically large and invasive the modest bookshelf speakers became, when placed within arms reach.
So that was a useful afternoon; I now know that this has a great deal of potential for me, and have also started to identify some things which could stand improvement for a near-field application. Some might be harder to address than others!
Though almost everything seems more critical. The positioning took hours to arrive at a particularly good setup, and the perceived stereo imaging still varies in success between different recordings. The time/distance mis-alignment between tweeter and woofer is quite noticeable too, as is the reduction in apparent bass (without much spl or room reinforcement). I was also quite surprised by how physically large and invasive the modest bookshelf speakers became, when placed within arms reach.
So that was a useful afternoon; I now know that this has a great deal of potential for me, and have also started to identify some things which could stand improvement for a near-field application. Some might be harder to address than others!
Last edited:
On further reading, I'm wondering if it might be worth beginning with a blank sheet on this. There are various things that seem to be more important than normal when close up - for instance the tweeter-woofer separation, beaming, diffraction and time-alignment. Even cabinet size, given how invasive they can be when so close to you.
There appear to be good arguments for a full-range driver, though equally there are questions over both high frequency beaming (which might translate to mere inches when sitting very close) and bass response (especially without room reinforcement in the near-field). So maybe a coaxial driver might be better.. though good ones seem rare for DIYers where I am.
Or maybe a relatively tiny mid-bass with a small tweeter combo for wide off-axis capability and good integration, that 'in some way' gets support from a modest bass-driver coming in below where localisation is quite so critical (though I'm not really thinking subwoofer). I'm currently feeling quite keen on this concept.. until/unless more knowledgeable people can say otherwise?
There appear to be good arguments for a full-range driver, though equally there are questions over both high frequency beaming (which might translate to mere inches when sitting very close) and bass response (especially without room reinforcement in the near-field). So maybe a coaxial driver might be better.. though good ones seem rare for DIYers where I am.
Or maybe a relatively tiny mid-bass with a small tweeter combo for wide off-axis capability and good integration, that 'in some way' gets support from a modest bass-driver coming in below where localisation is quite so critical (though I'm not really thinking subwoofer). I'm currently feeling quite keen on this concept.. until/unless more knowledgeable people can say otherwise?
Last edited:
For bass response your seating/ear position in the room can be critical due to cancelations from reflected surfaces based on the room dimensions. Try to be seated a distance from the wall you are facing (looking at your speakers) by about 38% of the length of the room. It is easy to put lots of bass into the room without hearing it if you are sitting in a null point. Your neighbours will hear the bass quite clearly though which is not good for you.
Thanks. I can't quite arrange the room to give optimum listening (or speaker) positions (which is partly why nearfield listening is attractive for reducing room interaction). But the layout could certainly be made some way towards that.
Though of course physics doesn't magically cease to exist. Maybe I could also make things more easily moveable so that the compromises might be reduced for more serious listening sessions.
Cheers,
Kev
Though of course physics doesn't magically cease to exist. Maybe I could also make things more easily moveable so that the compromises might be reduced for more serious listening sessions.
Cheers,
Kev
You can set your rear-vision mirror so that it works for you without touching it. If you adjusted it in the morning it might be set a little high, but the variation isn't that great.
Ha, I'm afraid that my familiarity with this field is so limited that I don't even understand your meaning there, Allen.
On another note though, I found this thread on a slightly different question but with a number of responses that are very relevant: Desirable traits for nearfield/desktop setup. It has me thinking that my existing 6" driver and tweeter in its fairly large waveguide might be less suitable than first thought. A 3" to 4" driver and small tweeter might be much better for close listening, both in audio characteristics and in speaker size. It could be quite satisfying to make something decicated that is small but high quality.
Though that said, it does appear that a fair bit of EQ and therefore excursion might be needed at the low end, which might not be great for quality from the smaller drivers. So possibly they might still be teamed with an additional pair of (very modestly sized) bass drivers, that come in below where localisation is quite so critical; sort of helper-woofers, as it were. If i understand correctly, our ability to localise is helpfully not very acute at low frequencies, and begins to rely more on time alignment and SPL, which could be tweaked with positioning and/or DSP.
On another note though, I found this thread on a slightly different question but with a number of responses that are very relevant: Desirable traits for nearfield/desktop setup. It has me thinking that my existing 6" driver and tweeter in its fairly large waveguide might be less suitable than first thought. A 3" to 4" driver and small tweeter might be much better for close listening, both in audio characteristics and in speaker size. It could be quite satisfying to make something decicated that is small but high quality.
Though that said, it does appear that a fair bit of EQ and therefore excursion might be needed at the low end, which might not be great for quality from the smaller drivers. So possibly they might still be teamed with an additional pair of (very modestly sized) bass drivers, that come in below where localisation is quite so critical; sort of helper-woofers, as it were. If i understand correctly, our ability to localise is helpfully not very acute at low frequencies, and begins to rely more on time alignment and SPL, which could be tweaked with positioning and/or DSP.
I was talking about angles, and things that are close to you. I thought you were talking about lobing when you mentioned full range drivers. Frankly I would be confident about being able to bring that together.
Since you've mentioned sitting close you should know that the biggest reflective surface is going to be the desk. Just a thought but if you did decide to use a waveguide, you might point a crossover lobe null at the reflection point to extent the directivity below the cross.
Since you've mentioned sitting close you should know that the biggest reflective surface is going to be the desk. Just a thought but if you did decide to use a waveguide, you might point a crossover lobe null at the reflection point to extent the directivity below the cross.
Ah i see, thanks for the info and for the explanation; it makes complete sense now!
Hmm yes, that is a good point, thank you. In theory I could probably avoid using a desk, and just use a handheld tablet or phone to control the music. Or maybe a large screen could be placed somewhat behind the speakers and a wireless keyboard used on my lap or something like that; closer than a strong reflection point to the ears. But yes, it would be more convenient to have a desk and avoid sound reflecting from it.
In fact.. as a key goal is to avoid exciting parts of the room, I don't actually want soundwaves directed downwards at all really, desk or no desk. My intentions already preclude using the floor for bass reinforcement, so there seems little to be gained from sound going very much below the horizontal.
Hmm yes, that is a good point, thank you. In theory I could probably avoid using a desk, and just use a handheld tablet or phone to control the music. Or maybe a large screen could be placed somewhat behind the speakers and a wireless keyboard used on my lap or something like that; closer than a strong reflection point to the ears. But yes, it would be more convenient to have a desk and avoid sound reflecting from it.
In fact.. as a key goal is to avoid exciting parts of the room, I don't actually want soundwaves directed downwards at all really, desk or no desk. My intentions already preclude using the floor for bass reinforcement, so there seems little to be gained from sound going very much below the horizontal.
Most effective speaker system to limit vertical dispersion is a line array. Open baffle becomes next, quite effective nulls to very low frequency with quite easy and relatively small construct. Horns / waveguides get big quick as you calculate what kind of an angle the floor / ceiling is and how big device would need to be in comparison to open baffle system for example, for nearfield the angle towards floor and ceiling gets bigger and waveguides start to be effective even with wide nominal dispersion. There is of course downsides for all of the systems so remember to consider all aspects of your system as you think it through. While reducing reflections would make sense don't go too far so that some other perhaps more important aspects get compromised.
I've posted some thoughts and demonstrations on the reflections few months ago and while any attenuation towards reflections would make difference you'd need 20db attenuation on the reflected signal to really reduce the effect at listening position, few db attenuation hardly makes a difference as could be the case with 90 deg waveguide with +2m listening distance for example. These experiments don't take account direction of the reflections (from which they arrive, and how we would perceive them) or other reflections that happen at the same time, which there are plenty off, so main takeaway from simple experiments has been to me at least that the remarkable hearing system of ours is quite effective filtering out the reflections and all we need to do is to help that, what ever the help would be.
Here is the thought process: Lets remember there is say one first reflection per every boundary of your room making it six making direct sound only 1 / 7 of the sound you hear within few milliseconds of first arrival. If one thinks what this would sound like it would sound like a single sound, we don't perceive small delays like this. Now if one takes multiple signals and superimpose them you'll notice the main features dominate the end result, I mean if the direct sound is flat but all the reflections have dip at 1kHz then the end result has a dip at 1kHz. In fact, it wouldn't perhaps matter what the direct sound was if all the reflections would have flat frequency response since they would completely dominate the superimposed response. This is simplified thinking of course, not taking account how the various arrival times actually affect the perceived sound, and how the higher order reflections affect, how much attenuation there is and so on, so perhaps don't take my message as scientific truth 🙂
I'd like to think that making the reflections sound similar to direct sound, but attenuate them as much as feasible (wide bandwidth), would make most sense to any loudspeaker design that tries to address the reflections somehow. I haven't dug too deep on research done on this so this is something I've come up with while thinking it through, everyone should make their own research and thinking. For example, making a small bandwidth null point to a reflection hardly makes a difference at listening spot (inspected from interference pattern) because rest of the bandwidth still interacts with the direct sound. While a dip towards reflection might help, especially if one knew that it was at some important frequency and time delay to help the perception, there might very well be detrimental effect as well, perhaps brain has to use more processing to figure out it is a reflection since the signature is not the same. For example two identical signals cause comb filter that has distinct pattern to it, dips at certain interval, but if one of the interfering signals has a dip on it the resulting comb filter dips interval is now mixed up around that. Here are few of the experiments I mentioned that demonstrate what I'm rambling about https://www.diyaudio.com/community/threads/faital-18-sound-3-way.377789/post-6942037
https://www.diyaudio.com/community/threads/efficient-2-way.380822/post-6897205
Anyway, just trying to give food for thought. We usually cannot get the reflections go away with home listening situation so best thing to do is be friends with it and try to balance the system so that the overall performance fits the bill. Headphones would take out all the reflections so in loudspeaker land there is no need to I guess, otherwise just use headphones. And don't forget take account room acoustics, or acoustic treatment, as part of the speaker design since the speaker and room works together to make best sound. For nearfield perhaps not as important as ratio of direct to reflected sound is higher.
I've posted some thoughts and demonstrations on the reflections few months ago and while any attenuation towards reflections would make difference you'd need 20db attenuation on the reflected signal to really reduce the effect at listening position, few db attenuation hardly makes a difference as could be the case with 90 deg waveguide with +2m listening distance for example. These experiments don't take account direction of the reflections (from which they arrive, and how we would perceive them) or other reflections that happen at the same time, which there are plenty off, so main takeaway from simple experiments has been to me at least that the remarkable hearing system of ours is quite effective filtering out the reflections and all we need to do is to help that, what ever the help would be.
Here is the thought process: Lets remember there is say one first reflection per every boundary of your room making it six making direct sound only 1 / 7 of the sound you hear within few milliseconds of first arrival. If one thinks what this would sound like it would sound like a single sound, we don't perceive small delays like this. Now if one takes multiple signals and superimpose them you'll notice the main features dominate the end result, I mean if the direct sound is flat but all the reflections have dip at 1kHz then the end result has a dip at 1kHz. In fact, it wouldn't perhaps matter what the direct sound was if all the reflections would have flat frequency response since they would completely dominate the superimposed response. This is simplified thinking of course, not taking account how the various arrival times actually affect the perceived sound, and how the higher order reflections affect, how much attenuation there is and so on, so perhaps don't take my message as scientific truth 🙂
I'd like to think that making the reflections sound similar to direct sound, but attenuate them as much as feasible (wide bandwidth), would make most sense to any loudspeaker design that tries to address the reflections somehow. I haven't dug too deep on research done on this so this is something I've come up with while thinking it through, everyone should make their own research and thinking. For example, making a small bandwidth null point to a reflection hardly makes a difference at listening spot (inspected from interference pattern) because rest of the bandwidth still interacts with the direct sound. While a dip towards reflection might help, especially if one knew that it was at some important frequency and time delay to help the perception, there might very well be detrimental effect as well, perhaps brain has to use more processing to figure out it is a reflection since the signature is not the same. For example two identical signals cause comb filter that has distinct pattern to it, dips at certain interval, but if one of the interfering signals has a dip on it the resulting comb filter dips interval is now mixed up around that. Here are few of the experiments I mentioned that demonstrate what I'm rambling about https://www.diyaudio.com/community/threads/faital-18-sound-3-way.377789/post-6942037
https://www.diyaudio.com/community/threads/efficient-2-way.380822/post-6897205
Anyway, just trying to give food for thought. We usually cannot get the reflections go away with home listening situation so best thing to do is be friends with it and try to balance the system so that the overall performance fits the bill. Headphones would take out all the reflections so in loudspeaker land there is no need to I guess, otherwise just use headphones. And don't forget take account room acoustics, or acoustic treatment, as part of the speaker design since the speaker and room works together to make best sound. For nearfield perhaps not as important as ratio of direct to reflected sound is higher.
Last edited:
For completeness sake here is link to some Bech & Soren papers that study audibility of reflections posted by fluid in response to some of my post on another thread https://www.diyaudio.com/community/...to-one-forum-member-an-ob.374532/post-6892249
The other paper ( Timbral aspects of reproduced sound in small rooms. II ) investigates detection threshold of single reflection and concludes the floor reflection would be easiest to detect and that in general attenuating the reflection with acoustic treatment and speaker directivity raise the detection threshold on mid-high frequencies. There is also mention that detection threshold was already high with speech signal but lower with noise signal etc. Perhaps there are some specifics that could be exploited in a loudspeaker design but to me this just reads that attenuating the reflections by speaker directivity is the best thing to do if one wants to address the reflections in a speaker design and if there is wide bandwidth attenuation perhaps the dips / lobing nulls don't matter mucho.
Fast read on the papers gave impression to me that directivity should be controlled down to 500Hz in order to make the reflections less detectable. This is very low frequency to control, especially for the floor reflection in far field ( +2m ) since it is in very low angle, perhaps something like 30 degrees. This calls for multiple drivers stacked vertically (MTM or variation of, or line array) or rather bulky waveguide. Nearfield ~1m the angle towards floor can be >50deg, which gives more attenuation but still if want to maintain the pattern control down to 500Hz the device needs to be almost the same size (due to the angle changing there is some leeway for same attenuation toward reflection at particular frequency so for the higher angle the device could be little smaller for same effect).
The other paper ( Timbral aspects of reproduced sound in small rooms. II ) investigates detection threshold of single reflection and concludes the floor reflection would be easiest to detect and that in general attenuating the reflection with acoustic treatment and speaker directivity raise the detection threshold on mid-high frequencies. There is also mention that detection threshold was already high with speech signal but lower with noise signal etc. Perhaps there are some specifics that could be exploited in a loudspeaker design but to me this just reads that attenuating the reflections by speaker directivity is the best thing to do if one wants to address the reflections in a speaker design and if there is wide bandwidth attenuation perhaps the dips / lobing nulls don't matter mucho.
Fast read on the papers gave impression to me that directivity should be controlled down to 500Hz in order to make the reflections less detectable. This is very low frequency to control, especially for the floor reflection in far field ( +2m ) since it is in very low angle, perhaps something like 30 degrees. This calls for multiple drivers stacked vertically (MTM or variation of, or line array) or rather bulky waveguide. Nearfield ~1m the angle towards floor can be >50deg, which gives more attenuation but still if want to maintain the pattern control down to 500Hz the device needs to be almost the same size (due to the angle changing there is some leeway for same attenuation toward reflection at particular frequency so for the higher angle the device could be little smaller for same effect).
Last edited:
Crikey, thank you very much indeed for all that info and the links to more; there is a great deal to take in there! I shall work my way through it and see what conclusions I feel able to draw from it all.
Cheers,
Kev
Cheers,
Kev
There is (as always) quite a lot to consider with this project, so I'm trying to avoid narrowing down to specific options yet. However after some quick simple trials I've certainly arrived at one promising possibility - speakers positioned very close to an armchair, rather like wing-mirrors on a car.
The layout seems to mean wide dispersion or off-axis response doesn't then become a big requirement, after all. Sitting close to the speakers inevitably means that even small distances off axis can create quite large angles, but in such a chair setup the speaker's relationship to the head are naturally and comfortably maintained within a relatively small, known volume. So contrary to what I'd initially imagined, it may not be necessary to get into things like unusually low crossover points or small or closely-positioned/coaxial/full-range drivers etc. to maintain consistency throughout the listening position. The close proximity over soft furnishings (and my legs) are also good for reducing early reflections and floor resonances, too.
But I wonder if there are any new driver-integration factors likely to come in, instead. For example, even if any beams in dispersion or vertical lobes are directed where one wants them, it must presumably still take 'some' distance for adjacent drivers to merge into one apparent source.. but I'm not at all sure what that distance might be?
Thanks, Kev
The layout seems to mean wide dispersion or off-axis response doesn't then become a big requirement, after all. Sitting close to the speakers inevitably means that even small distances off axis can create quite large angles, but in such a chair setup the speaker's relationship to the head are naturally and comfortably maintained within a relatively small, known volume. So contrary to what I'd initially imagined, it may not be necessary to get into things like unusually low crossover points or small or closely-positioned/coaxial/full-range drivers etc. to maintain consistency throughout the listening position. The close proximity over soft furnishings (and my legs) are also good for reducing early reflections and floor resonances, too.
But I wonder if there are any new driver-integration factors likely to come in, instead. For example, even if any beams in dispersion or vertical lobes are directed where one wants them, it must presumably still take 'some' distance for adjacent drivers to merge into one apparent source.. but I'm not at all sure what that distance might be?
Thanks, Kev
My solution for small room near field listening was to build the speakers 'upside down'.
If the speaker were conventionally laid out, then effectively the bass driver offset would be huge and 'lobing' pointing down at the floor. Measure the offset from the ear position and the bass driver would be much farther away - something you'd have to deal with in the crossover. Having the bass driver away from the floor eliminates floor bounce too.
With the upside down configuration offset is only about 20mm with the speakers tilted back slightly - about 15° - as measured. A much simpler crossover is required and everything suddenly becomes easier to design.
The only issue with small rooms is uneven response due to room interference. I have the balance control set at the 10 O'clock position and the treble is still off to the right a little. Leaning forward corrects the treble anomaly but the room just isn't large enough to listen in that position. Other than changing room there's nothing I can do about it. The two shelf supports L&R aren't there to support the shelf, they are there to interrupt midrange wall reflections - they do work!
One advantage of a small room is the cabin effect. The sub has a flat response down to where it makes your internal organs vibrate...
Welcome to my man cave 😛
If the speaker were conventionally laid out, then effectively the bass driver offset would be huge and 'lobing' pointing down at the floor. Measure the offset from the ear position and the bass driver would be much farther away - something you'd have to deal with in the crossover. Having the bass driver away from the floor eliminates floor bounce too.
With the upside down configuration offset is only about 20mm with the speakers tilted back slightly - about 15° - as measured. A much simpler crossover is required and everything suddenly becomes easier to design.
The only issue with small rooms is uneven response due to room interference. I have the balance control set at the 10 O'clock position and the treble is still off to the right a little. Leaning forward corrects the treble anomaly but the room just isn't large enough to listen in that position. Other than changing room there's nothing I can do about it. The two shelf supports L&R aren't there to support the shelf, they are there to interrupt midrange wall reflections - they do work!
One advantage of a small room is the cabin effect. The sub has a flat response down to where it makes your internal organs vibrate...
Welcome to my man cave 😛
Here's some ideas. The advantage of nearfield listening is it is less room dependant. I think it sounds clearer too.
Speakers away from the wall reflections.
IMO, there are some even better refinements. Doubled drivers actually put only half the power into the room for same SPL at the on-axis listening position.
SEAS use this sort of idea for overly lively and echoey conference rooms.
http://www.seas.no/index.php?option=com_content&view=category&layout=blog&id=86&Itemid=399
An MTM with a directional tweeter.Very focussed,
Also good IMO, is this sort of Logitech Z200 computer speaker. Again, overall less power for the same SPL. You could add a sub-woofer if bass is your bag.
In practise you could wire a couple of 4 ohm fullrangers in series to the same effect. MarkAudio 3" CHN-50 fits the bill in about 4L reflex for two. Bass can be adequate according to the buffs in the multiway forum.
Speakers away from the wall reflections.
IMO, there are some even better refinements. Doubled drivers actually put only half the power into the room for same SPL at the on-axis listening position.
SEAS use this sort of idea for overly lively and echoey conference rooms.
http://www.seas.no/index.php?option=com_content&view=category&layout=blog&id=86&Itemid=399
An MTM with a directional tweeter.Very focussed,
Also good IMO, is this sort of Logitech Z200 computer speaker. Again, overall less power for the same SPL. You could add a sub-woofer if bass is your bag.
In practise you could wire a couple of 4 ohm fullrangers in series to the same effect. MarkAudio 3" CHN-50 fits the bill in about 4L reflex for two. Bass can be adequate according to the buffs in the multiway forum.
Attachments
The idea of using your environment's nearfield is you don't have to worry about these things. It will be OK when you sit in front on the design axis.
In a real world situation there is always other involvement, but for the topic at hand you're overthinking it.