Baffle Difraction &

[ralph-bway]

Hi All,

I built a set of arrays using the PE 4.5" 49 cent drivers very similar to these:

http://ceramicsubs.web1000.com/4x4x42/index.htm
(sorry this is the best pic available to show the cabinet)

Now as you can see the front baffle is just wide enough to house the drivers. How does such a narrow front baffle affect the imaging and soundstage of the speaker.

Also I'm interested in doing an open dipole design for front mains to operate from aproximately 250HZ - 4KHZ using six of these drivers I have on hand for each side:

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=299-434

From all the reading & searching I've done it seems the wings should be at least as wide as the front baffle.

I have 80 of the Onkyo tweeters to cover above 4KHZ and thought if I mounted them on a seperate baffle and if angled properly the Audax & Onkyos would be on axis to my listening chair. Does this seem reasonable or am I missing something.

I also have 4 Peerless drivers for the subs (right & left).

 

[cc00541]

The narrow baffle is an aid to imaging as the time delay between the direct and reflected sound from the baffle edges is very small.

To have a dipole accommodate a flat response down to 250 Hz, each of your wings would have to be around 24” wide, although you could actively equalize the 6 dB / octave drop off characteristic of dipoles and get by with a smaller baffle. Another option would be to angle the baffles back to minimize the footprint of the array.

SL is the guru, and much more eloquent than I, on the subject of dipoles. Here’s a link to his dipole discussion:

http://www.linkwitzlab.com/models.htm#A

C

 

[johninCR]

I don't know how C came up with the 24" wing figure. The wavlength of 250hz is 54", so a 54" wide flat baffle would give you a peak in response at 250hz, not flat response. To get only down to 250hz requires minimal baffle and much depends on the shape and placement with the sides swept back. Most likely just cutting off the back of the array cab in your pic will get you down to 250hz. Having the baffle that is big enough to be stable should be enough. You don't want 90 degree sides due to cavity resonances and make your front baffle big enough to have the tweeter array on the same front baffle. Having the tweeter array on a different angle of baffle is asking for trouble.

If great imaging is a goal, then you should consider the tapered array configuration mentioned in Dr. Griffins paper. Also with regard to imaging, I find that narrow front OB arrays image better than wide fronts.

 

[BillFitzmaurice]

The CTC from the midbasses to the tweeters should be as close as possible, so by having them in separate boxes you've got problems from the get go. Do the best you can to get them as tight as possible. If you want to run the Onkyo's that low you'll need to run at least a 4th order HP filter.

 

[cc00541]

I don't know how C came up with the 24" wing figure. The wavlength of 250hz is 54", so a 54" wide flat baffle would give you a peak in response at 250hz, not flat response.

Yes, you are correct about the peak, but below that point the response will roll off at 6 dB/octave summed with the natural response of the driver. Above that point would be a dip @ 500 Hz, etc.

IMO, the peak is not as significant as the 6 dB roll off below it, and I assumed Ralph would read SL's documentation, so I did not mention it. I also assumed a 6" baffle width for the drivers and 24" wings. A 6" wide baffle with no wings would start the characteristic 6 dB roll off at around 1130 Hz.

C

 

[ThorstenL]

Konnichiwa,

Yes, you are correct about the peak, but below that point the response will roll off at 6 dB/octave summed with the natural response of the driver. Above that point would be a dip @ 500 Hz, etc.

IMO, the peak is not as significant as the 6 dB roll off below it, and I assumed Ralph would read SL's documentation, so I did not mention it. I also assumed a 6" baffle width for the drivers and 24" wings. A 6" wide baffle with no wings would start the characteristic 6 dB roll off at around 1130 Hz.

You know what. You are eloquent in quoting theory.

Now here is the practice. I have a driver with the following parameters:

Fs = 55Hz
Qts = 0.7

It is placed on a baffle which is around 2' (24") wide and thus, based on your "milkmaid" calculation rolls off at 500Hz. Now my problem is that I get in room a measured response that places -6db @ 50Hz and is free from drastic ripples.

Explain.

Sayonara

 

[johninCR]

Here's another example. My 15" Seleniums, Qts .62 , start their dipole bass roll off at 300hz WITH NO BAFFLE, just propped up using some paper egg crates.

 

[ThorstenL]

Konnichiwa,

Now as you can see the front baffle is just wide enough to house the drivers. How does such a narrow front baffle affect the imaging and soundstage of the speaker.

It affects it very badly. Assuming the wavelength is such that the driver (array) does not show significant beaming anything having a wavlength longer than 2 X the width of your baffl will in eaffect "wrap around" the baffle and will creat a very wide dispersion, putting ound anywhere and everywhere in addition to directing it at listening position.

Further, such narrow faffles invariably louse up instrument tonality. Evaluation of wavelength and formants of instruments dictates that a speaker should have a baffle with the smallest dimnsion no les sthan around 10", preferably more.

From all the reading & searching I've done it seems the wings should be at least as wide as the front baffle.

Wings? I have a simple rule that indicates the EFFECTIVE baffle width for use. That is that the baffle is as wide as it is wide visually. Sweep back "wings" from a plan baffle and you losse progressively the actual "surface extension". Once you fold back much past around 60 degrees the acoustic system does not operatle really any longer as classic dipole and other approaches must be taken to get predictions to work well.

I would suggest that if you make your baffles at 10" Wide (visually) and your drivers have a flat response loaded into "open air" (eg Qt = 0.7 or higher) and sufficiently low resonance you will be fine, for anything but serious fundamental bass in the lower notes, or wherever the self resonance of the driver places the rolloff.

Sayonara

 

[johninCR]

Re: Re: Baffle Difraction &

Wings? I have a simple rule that indicates the EFFECTIVE baffle width for use. That is that the baffle is as wide as it is wide visually. Sweep back "wings" from a plan baffle and you losse progressively the actual "surface extension".

I know that we have discussed this before, but I still don't understand this "visual width" idea. While I understand that other factors come into play with swept back wings, I only want to talk about bass extension. Simple geometry tells me that when you bend back the sides of a flat baffle, the rear wave must travel farther to reach your ears than with the baffle being flat. Bass extension from H & W baffle shapes also appears to contradict this "visual width rule".

Can you please explain "surface extension", maybe that's the part I'm missing.

 

[ThorstenL]

Re: Re: Re: Baffle Difraction &

Konnichiwa,

I know that we have discussed this before, but I still don't understand this "visual width" idea.

It is simply a ballpark idea. Seems to work. I explained before that you need to start to think in terms of waves, not pressure. I can't help anymore.

Bass extension from H & W baffle shapes also appears to contradict this "visual width rule".

Actually, they don't, because they are much more copmplex than the simple dipole the "visusal width" rule applies to. And due to their resonances they are invariably operates strictly below the actual "baffle cutoff" and not around it.

Can you please explain "surface extension", maybe that's the part I'm missing.

To understand the way an open baffle (not any more complex things, a plain baffle with or without wings) works you need to get the point that you are attempting to "launch" a soundwave into the room, into a specific direction.

Then take stones to a quiet pond and place some 2-dimensional "baffles" in the way of the waves your pebble causes, preferably you drop it extremely close in front of the "baffle".

Sayonara

 

[johninCR]

Waves is exactly how I visualize it. Drop 2 stones one in front and one in back. With swept back wings the rear wave must travel farther to reach the listening position assuming an equal surface area and height. The rear wave must travel to the baffle edge, then it goes an equal remaining distance as the front wave with a flat baffle. With the swept back wings it must travel to the edge of the baffle PLUS forward to the front plane of the baffle then the remaining distance is equal. The distance differential is greater, so the bass goes deeper with swept back wings despite being visually narrower.

The closer the wings are to the driver(s) the bigger the difference. Short wings at the ends of a wide baffle offer little benefit. On the other hand close to the drivers results in a narrow front which is apparently detrimental to SQ.

 

[cc00541]

You know what. You are eloquent in quoting theory.

Thank you. I understand a bit of it as well...

Now here is the practice. I have a driver with the following parameters: Fs = 55Hz Qts = 0.7 It is placed on a baffle which is around 2' (24") wide and thus, based on your "milkmaid" calculation rolls off at 500Hz. Now my problem is that I get in room a measured response that places -6db @ 50Hz and is free from drastic ripples. Explain. Sayonara

Since you didn't specify, I would surmise you have the baffles placed close to the front wall. This would cause cavity effects and mitigate operation as a dipole as well. The close proximity of the front wall would tend to shore up the low end, however.

Your quote about narrow baffles and imaging:

It affects it very badly. Assuming the wavelength is such that the driver (array) does not show significant beaming anything having a wavlength longer than 2 X the width of your baffl will in eaffect "wrap around" the baffle and will creat a very wide dispersion, putting ound anywhere and everywhere in addition to directing it at listening position.

But doesn't a dipole by design have a figure 8 response pattern, with as much out the back as directed to the listener? And as a consequence, isn't wide dispersion one of the advantages of a dipole? For that reason imaging is not the strong suit of the dipole. If you want imaging, build a monopole. If you want a big spacious sound, the dipole will give that to you.

Wings? I have a simple rule that indicates the EFFECTIVE baffle width for use. That is that the baffle is as wide as it is wide visually. Sweep back "wings" from a plan baffle and you losse progressively the actual "surface extension". Once you fold back much past around 60 degrees the acoustic system does not operatle really any longer as classic dipole and other approaches must be taken to get predictions to work well.

Due to the cavity resonances...



From what I have read, most of your rhetoric appear to be purely subjective. I'll toss the ball in your court. I'd enjoy some clarification of the rational behind all those suppositions.

C

 

[wintermute]

Re: Re: Baffle Difraction &

Konnichiwa,

Further, such narrow faffles invariably louse up instrument tonality. Evaluation of wavelength and formants of instruments dictates that a speaker should have a baffle with the smallest dimnsion no les sthan around 10", preferably more.
Sayonara

Hi Kuei Yang Wang,

Can you post any url's or references for research that suggests this. I was about to use a baffle width of around 7" on some new speakers (final dimentions not determined yet). I have seen plenty of designs with baffles this narrow or narrower and had until this point never come across any guidelines that said the minimum baffle dimention should be 10 or more".

Tony.

 

[ThorstenL]

Konnichiwa,

Since you didn't specify, I would surmise you have the baffles placed close to the front wall.

Define "close". How close? In conditions other than at my own place I had them as far as 1.5m from teh wall behind them, they still measured about the same.

Your quote about narrow baffles and imaging:

But doesn't a dipole by design have a figure 8 response pattern, with as much out the back as directed to the listener?

EXCEPT, in the specific question the enquirer did not ask about a dipole, but about a monopole. The Dipole crops up in a seperate question.

And as a consequence, isn't wide dispersion one of the advantages of a dipole?

It is? Do you actually have the faintest how dynamic driver dipoles REALLY work? For how wide a bandwidth does a traditional dynamic driver actually behave as dipole?

From what I have read, most of your rhetoric appear to be purely subjective.

Actually, it is purely objective, as I have measured similar size dipoles in various versions and with various drivers (from $ 20 Audax Fullrangers to very expensive fieldcoil drivers) in a wide variety of locations and placements (3 HiFi Shows, my own living room, my business partners living room and a friends place who build himeself some very similar baffles) and they all agreed in terms of LF response within the limits of expermental error.

Equally, they also agreed with the mathematics presented in the AES Preprint 5025.

You may find studying the current, up to date literature on the topic illuminating.

So, basically your position is "The Milkmaid calaculations are right and with 1.5m or more free space behind dipoles cavity resonances increase the LF output." Not what I'd call convincing.

Sayonara

 

[ThorstenL]

Re: Re: Re: Baffle Difraction &

Konnichiwa,

Can you post any url's or references for research that suggests this.

Not really. Much of the fundamentals are found in a range of older publications on acoustics, though as far as I know no-one has ever bothered formally to overlay the typhical Baffle diffraction regions with the instruments and Vocals Formant regions (except myself that is). If you understand that formants actually MAKE the sound of an instrument you understand why the acoustic envoironment seen by the wave should not change.

BTW, I made a big mistake, I someohow messed up halve waves and halve dimensions. The minimum baffle width that ensures that the formants of most instruments and the human voice can develop waveforms undisturbed is of course 20", not 10". Mea Culpa.

Eve wondered why most studios have soffit mounted Monitor speakers (fluns mounted into the wall)? The resonans are acoustic, not interior architecture determined.

I was about to use a baffle width of around 7" on some new speakers (final dimentions not determined yet). I have seen plenty of designs with baffles this narrow or narrower and had until this point never come across any guidelines that said the minimum baffle dimention should be 10 or more".

Yes, there are plenty of narrow front speakers. Usually with their on-axis response equalised flat supposedly offsetting the baffle loss.

And how do they usually sound?

Compare at some time a narrow baffle speaker equalised to flat response on axis to a wide baffle one (20" Minimum) also equalised to flat response on axis. You might find listening to the differences amusing.

Sayonara

 

[pooge]

Konnichiwa,

Yes, there are plenty of narrow front speakers. Usually with their on-axis response equalised flat supposedly offsetting the baffle loss.

And how do they usually sound?

Compare at some time a narrow baffle speaker equalised to flat response on axis to a wide baffle one (20" Minimum) also equalised to flat response on axis. You might find listening to the differences amusing.

Sayonara

This is likely due to the fact that if you equalize the on-axis response of the narrow loudspeaker, you raise the power response above flat. This is due to the fact that the acoustic load of the narrow loudspeaker changes on either side of the baffle-loss frequency, whereas the wide baffle maintains the acoustic load. If equalization is used to correct baffle loss, the on-axis correction such get back only 3db rather than the full 6db in order to maintain flat power response, ignoring other load changes such as room boundaries. If the full 6db on-axis response is equalized back to flat, the power response will be up 3db. Boom Boom.

 

[ThorstenL]

Re: Re: Baffle Difraction &

Konnichiwa,

This is likely due to the fact that if you equalize the on-axis response of the narrow loudspeaker, you raise the power response above flat.

You could equally equalise for flat power response and observe differences.

Perhaps the best way of saying what is on is that the directivity index of the speaker should change only minimally throughout the formant region. And the issues are not frequency response per se either.

Sayonara

 

[wintermute]

Thanks Kuei Yang Wang,

I guess building loudspeakers is all about compromises, I'm not really willing to make 20" baffles for my speakers so, narrow baffles will have to be one of my compromises (I suspect too that having wide baffles while it may address some problems, may also cause others, as is the nature speakers, improving one aspect allmost allways detracts from another).

I'll roll my dice and see what happens..... plenty of room for tweaking in the future!!

I've downloaded the edge and the excell based baffle diffraction simulator from the frd consortium so I'll play with those and see what I come up with.

Tony.

 

[LineSource]

In Jan 2005 Stereophile Michael Fremer reviews the Sonus Faber Stradivari speaker which is 53.5" high, 25.5" wide and 19.7" deep and noted:

" The Stradivari's wide baffle produced a singular sonic picture. Instead of the more common narrow-baffle, low-diffraction sound, in which a speaker "disappears" to leave behind a ghostly apparation of a three-dimensional sound picture. the Stradivari presented a more weightly, unusually solid picture that seemed to be a three dimensional curtain wrapped behind the baffles and extending well back into virtual space. While more conventional baffles have produced wider, more transparent soundstages and perhaps more focused and upfront images, none has delivered such a solid and physically believable three-dimensional soundstage in my room."

 

[wintermute]

Thanks Linesource,

interesting info! if I'm interpreting correctly, the wider baffle apparently gives the soundstage more depth and solidity (but less width).

Though without testing with everything else equal (except perhaps BSC) and narrow baffles, attributing the soundstage simply to the baffle size is probably somewhat subjective, as there may be other factors in the speakers design affecting the soundstage.

But if the reviewer is correct then if a solid well defined soundstage is your goal, perhaps wide baffles is indeed the way to go.

perhaps my existing 3 ways with their (what I considered extremely wide) probably at least 20 inch baffles are the way to go afterall I must admit on occasion I have been fooled into thinking that there were noises coming from behind me that were in fact coming from the speakers!! (stuff on tv or dvd, no surround processing involved)

Tony.

edit: Just when you think you have everything worked out along comes something to upset everything. I'm sure I had read heaps of stuff that suggests that narrow baffles are the best for imaging.... maybe some test boxes are in order