1. The Gear Page is run by musicians for musicians. We listen, we learn and if we misstep we are not afraid to do the right thing. We proposed some changes to the Emporiums. Based on feedback from members, we have decided to not go ahead with those changes. However, it has also highlighted that we need some community input into what is working and what is not working for members here. Primarily focused on the Emporiums, we'd like input on your thoughts about TGP and how things work in the Emporiums for you and how you'd improve them. The discussion thread on the is here!

    Dismiss Notice

Speaker Directivity

Discussion in 'Amps and Cabs' started by Jay Mitchell, Nov 30, 2008.

Thread Status:
Not open for further replies.
  1. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    Hello All,

    Apparently something in the "Speaker Dispersion" topic triggered a move to the "Manufacturers' and Retailers' Forum." Given the relevance of the material I provided to this forum, I'm starting a new thread with only the free information I provided in the other one.


    Speaker directivity is the bane of a consistent guitar sound. That much is generally recognized. However, there are several audio myths that continually get repeated, and they should be be corrected:

    1. High frequencies come from the center of a cone speaker, and that's why they beam.

    Nope. If you could actually get the high frequencies to be radiated by the "center" of the speaker - say the dust cap - you'd find them being radiated over a greater angle, not a smaller one. A number of past cone speaker designs actually had compliance elements built into the cone in an attempt to create this exact scenario: the size of the cone effectively getting smaller at higher frequencies. They didn't work particularly well.

    2. Placing an obstruction in front of the center of the cone "blocks" these "beaming" frequencies and makes the speaker's directivity (the word "dispersion" does not actually apply to radiation pattern) broader.

    Nope again. The "blocker" will indeed change the on-axis response and directivity of the speaker, but the effect is not consistent or necessarily useful. The directivity will actually be made narrower at some frequencies.

    The mechanisms whereby an obstruction placed in front of a speaker changes its response and directivity are complex and counterintuitive. "Blockage" has nothing to do with it, however.

    An obstruction will reflect sound back towards the cone, which will "re-reflect" the sound forward. The time it takes sound to make this extra trip means that the reflected sound will be delayed by some amount compared to sound that didn't make the extra trip. The response of the combined sound - slightly delayed plus undelayed - will contain interference ("comb filters"). The "phasiness" some folks describe when they install beam blockers is due to these comb filters.

    A portion of the outgoing sound will diffract around the obstruction. This sound is also delayed and will cause an additional set of comb filters. The diffracted sound also has a different radiation pattern from that of the speaker, since it comes from the edge of a small disc. Taken by itself, the diffracted radiation has a strong beam directly in front of the disc. Combined with the other radiation (direct and reflected), the radiation pattern will vary widely with frequency - even more than the pattern of the speaker with no blockage. If you find these effects desirable - as a number of players apparently do - then there is nothing wrong with the use of "blockers," but they don't cause the directitivity changes their makers attribute to them.

    If you want to alter the directivity of a guitar speaker in a favorable and frequency-consistent way, I've developed a means to accomplish that, and it can easily be tried by any reasonably competent DIY type. It is outlined in the next post. I've used this method on my tube amps, and it works beautifully.
     
    Duneland Labs, LordByron and iim7V7I7 like this.
  2. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    Here's how to make your own speaker directivity modifier:

    Cut out a doughnut-shaped piece of acoustically absorbent foam. The diameter should be the same as the speaker cutout in the baffle in your cab, and the diameter of the center hole should be ~3". Attach it to the rear side of your cab's grille using spray contact adhesive (e.g., 3M Super 77, available at Lowe's). Spray a light coating of adhesive on the foam only, and press it against the grille cloth within about 30 seconds of spraying it. You can easily remove the foam with no ill effect on the grille material, if you decide you don't like the effect.

    The material you want is open-cell polyurethane foam in sheet form, and there are a number of sources for it. McMaster-Carr is one. I use acoustic foam that the company I own purchases for use in my loudspeaker designs, but that is a matter of convenience. I have tested and subjectively evaluated two thicknesses: 1/2" and 3/4". The limit on maximum thickness is the thickness of your baffle, so make sure you don't exceed that.

    With the material I use, the 3/4" doughnut produces the most consistent response at different angles. A 12" speaker has huge variations in its response above ~1200 Hz within just 10 degrees of the speaker's axis. With the 3/4" foam doughnut in place, the on axis response and the response at 40 degrees off axis are almost identical. This is a huge improvement.

    If you think about the subject of directivity, you'll easily recognize that there are two ways of saying the same thing: when you say a speaker becomes "beamy" at high frequencies, you're also saying that its on axis response is much brighter than its off axis response. For example, if you equalized the response to be flat on axis (a hypothetical exercise, as that's never what you actually want from a guitar speaker), you'd find that the response off axis falls off pretty rapidly at higher (> 1200 Hz) frequencies.

    The reason for the preceding paragraph is to point out that making directivity more consistent over frequency requires that either the on axis or off axis response change. The foam doughnut causes a change in the on axis response, while leaving the off axis response alone. This means that, if you've tweaked your tone with the speaker aimed at your ears, it's now going to sound darker, and you'll need more treble, presence, and/or midrange, depending on the design of your amp's tonestack and other tone-altering circuits. If you're placing your amp on the floor facing the audience, the response you hear will change little or none, but the response the audience hears will now match what you've been hearing all along.
     
  3. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    Its o.d. should be a match for the baffle cutout (maybe even 1/16" smaller), so that it sits entirely within the cutout without being stuffed. If the cutout isn't round - some cabs have cutouts with a flat - then you should cut the foam to match the actual shape.

    Yes.

    Yes.

    None, if you're accustomed to playing in an off-axis position. Open cell foam in this thickness has almost no effect on sound passing through it at frequencies below ~1kHz. The doughnut will therefore cause no change in off axis response or sensitivity. It will reduce higher-frequency content directly on axis. If you're accustomed to playing with your cab aimed straight at your head, the doughnut will make it sound darker by reducing the on-axis sensitivity of the speaker in the higher frequency ranges. If you've adjusted your tone to remove the "ice pick" effect on axis, you will need to brighten it with the doughnut in place. The advantage you will gain is that your position relative to the cab will no longer matter so much: it will sound the same off axis as when you aim it at your head.
     
    Duneland Labs and LordByron like this.
  4. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    To answer a couple questions: to try the idea, you could temporarily attach the doughnut to the front side of the grille. Just make sure to center it over the speaker cutout. If your cab has metal speaker protectors, you can attach the doughtnut to those.

    The spacing of the foam off of the speaker cone is not critical, but you do not want it to make contact with the cone, and you also do not want there to be a flanking path (i.e., a large gap around the outer edge of the doughnut) that sound can take without passing through the foam.

    In order to understand why/how it works, you have to understand a bit of the geometry and acoustics in a cone transducer. The cone is driven very near its center (where the voice coil former is attached to the cone). The mechanical excitation (the vibration of the cone) is not instantaneous. It moves outward in the cone with its own characteristic velocity, which is geater than the velocity of sound in air. As the excitation moves outward, the air in contact with the segment of the cone that is moving is also excited, and sound is radiated from that segment. If the angle of the cone is "just right" for the properties of the cone material, the sound that is radiated from each section will be almost perfectly in time with the sound that was radiated earlier, from a portion of the cone that is set further back.

    When all this sound adds up out in front of the cone, you get a coherent wavefront to a relatively high frequency (as high as ~5kHz), but only on axis. At off-axis positions, the synchronization falls apart very rapidly, to the extent that a typical guitar speaker will see its output above 1200 Hz fall off by as much as 15-20dB within just a few degrees off axis. The effects on directivity due to mis-synchronized off-axis arrivals are much less at lower frequencies, because the wavelengths of sound are larger than the cone, and the speaker is therefore not "beamy" at these frequencies.

    Placing an absorber doughnut in front of the cone has almost no effect below 1kHz, because the material is too thin to have a significant effect at lower frequencies. This is good, since, as we've seen above, the speaker's behavior below 1kHz is not a problem. At higher frequencies, the radiation from the outer portions of the cone is progressively absorbed, whereas the portion of the radiation from the center of the cone which moves straight forward is allowed to pass through the opening in the doughnut unattenuated. The result is that there is a net reduction in high-frequency content on axis, but the high frequencies that do get through are now radiated by a virtual source - the opening in the doughnut - which is much smaller than the actual speaker, and they are therefore radiated over a greater angle. The speaker's directivity is now essentially the same over the entire range of interest for electric guitar, and you're out a few bucks for a piece of foam and a few minutes to cut out and attach the doughnut.
     
    Duneland Labs and LordByron like this.
  5. rhythmrocker

    rhythmrocker 1966 Battle of the Bands Supporting Member

    Messages:
    2,876
    Joined:
    Jul 19, 2006
    Location:
    City by the Bay, CA
  6. kimock

    kimock Member

    Messages:
    12,589
    Joined:
    Mar 12, 2005
    Location:
    Where the Palm Tree meets the Pine
    Thanks chief, excellent work.
     
  7. teleamp

    teleamp Senior Member

    Messages:
    3,508
    Joined:
    Aug 22, 2006
    Location:
    Central Texas
    Thanks for an excellent post. I will be doing some experimenting, thanks.
     
  8. rockon1

    rockon1 Member

    Messages:
    12,058
    Joined:
    Aug 12, 2005
    Location:
    Stamford CT
    Thanks Jay excellent stuff here! Sooooooo t o make sure Ive got this correct.....recap?

    1) Use McMaster-Carr part # 85735K72 (or similar product)

    2) Cut foam to size of baffle cut out.

    3) Make 3" hole in center of foam.

    4) Attacth to inside of grill with contact spray adhesive.

    I'll be busy soon! Thanks Bob
     
  9. zzmoore

    zzmoore Member

    Messages:
    6,820
    Joined:
    Sep 22, 2006
    Location:
    USA, California, 94585
    Would you like to post some pictures of your work?
    Thank You
     
  10. Zelja

    Zelja Member

    Messages:
    1,441
    Joined:
    Aug 6, 2007
    Location:
    Sydney, Australia
    Thanks Jay, great info & things to try!

    In your posts , the negative effects of comb filtering are mentioned. I understand that having two or more speakers in a box will cause comb filtering - will the "doughnut" eliminate the comb filtering caused by 2 or more speakers in one cabinet (using one per speaker of course)?

    Also, have you ever used this method in a detuned cab scenario i.e. a closed-back cab with one speaker but basically another speaker sized cut out in the front panel as well? I have experimented by taking out a speaker from a 2 x 12 & found the distorted tones to be better (not as harsh) & "beaming" to be less of a problem. Any thoughts?
     
  11. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    For tone producers, as opposed to reproducers, comb filters are not always bad. I mentioned that beam "blockers" can cause comb filtering, but I did not say that this effect is generally a negative one. Some players may like it.

    No.

    No, but the effect on the speaker's beaminess will be the same, regardless of the enclosure.
     
    Duneland Labs likes this.
  12. Chuck Snider

    Chuck Snider Supporting Member

    Messages:
    3,965
    Joined:
    Jan 4, 2002
    Location:
    Limbo
    yea,
    could you do that?
    thanks for your time
     
  13. gag halfrunt

    gag halfrunt A fellow of infinite jest Supporting Member

    Messages:
    1,928
    Joined:
    Aug 30, 2005
    Location:
    Marquette, Michigan
    Jay,

    How would you address the beaminess of a tweeter? One of my gig rigs is an Axe-FX, and I want to use a 2-way, full range speaker with it (to take advantage of cab simulations), but would like to get rid of the harsh highs on axis. Thanks.
     
    Last edited: Dec 1, 2008
  14. bruce egnater

    bruce egnater Member

    Messages:
    410
    Joined:
    Sep 10, 2004
    Jay,
    I don't understand how your foam piece with the hole is achieving the same goal as the foam disc in the center I mentioned earlier. The center foam disc attenuates (does not block) the beaming high frequencies at the center while leaving the edges unaffected. It would seem that since the "edge sound" is the desireable tone, it would be best to leave that area of the driver unobstructed. Measured response across the front of the speaker from edge to center with no disc shows a substantial increase in high frequency output as you move towards the center. With the foam disc installed to attenuate the high frequencies at the center, the measured response is essentially uniform across the entire front of the cone from edge to edge. Can you explain how the foam ring at the edge would stop the beaming effect of the highs from the center. Not disputing your results, just not understanding.
     
  15. Tone_Terrific

    Tone_Terrific Supporting Member

    Messages:
    25,865
    Joined:
    Dec 26, 2004
    Location:
    Canada-GTA

    The multi-thread aspect of this subject may have anyone missing this post where it is quite well explained but doesn't address your particular point.
    So, just as a reference point to those who have missed it:

    http://www.thegearpage.net/board/showpost.php?p=5094853&postcount=26

    This really is a must-try for anyone who is concerned with on-axis beaminess. I am trying to source this type of poly foam locally. No luck finding it, yet.
     
  16. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    No, it does not. The "beaming" high frequencies come from all parts of the speaker cone. They beam specifically for this reason: the size of the radiator (the cone) is large compared to a wavelength of sound. This is the part that folks get completely wrong, and it is this mistake that leads to the audio myth. If high frequencies only came from the center of the speaker, the speaker would not be "beamy" at those frequencies.

    It is not. See my original explanations of how and why this works. If you still have trouble understanding why, try it out. There is no question that it does work.

    Wrong measurement, wrong distance. You must measure the response in the farfield in order to see directivity.

    Again, wrong measurement, wrong distance. You need a suitable test area for loudspeaker measurement, and you must place the test microphone at least 6 feet away from the speaker. Your ears will never be against the grille cloth. If you want measurements to be indicative of what you will hear, you can't put the mic in that position either.

    I already did, in some detail. See my earlier post in this thread.
     
    Duneland Labs and yeky83 like this.
  17. stephen sawall

    stephen sawall Member

    Messages:
    5,779
    Joined:
    Nov 22, 2008
    Location:
    Seattle WA USA
    I also would like a photo to make sure I am seeing this right in my mind. Thanx.
     
  18. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    I made two "doughnuts" and installed them in two of my amps more than a year ago. I did not photograph anything at the time, and I have no plans to remove the speaker from either amp in the near future. If I have occasion to do so, and if a camera is nearby when this happens, I may take photos.

    This entire setup is so simple as to require no photos to visualize or create, however. It really is every bit as simple as I describe.

    Once again, in a nutshell: Remove the speaker (presumably rear-loaded) from your cabinet. Look at the hole in the baffle on which it is mounted. Cut out a piece of foam that fits that hole. Cut out a 3"-diameter circular hole so that it will lie directly in front of the center of the speaker. Attach foam to the rear of the grille, making sure that it is not thicker than the baffle. Remount and reconnect speaker. Play.
     
    Duneland Labs and ericb like this.
  19. bruce egnater

    bruce egnater Member

    Messages:
    410
    Joined:
    Sep 10, 2004
    Thanks Jay. I guess I am one of those "engineers" who should probably stick to their area of expertise, which obviously is not psychoacoustics. Thanks for all the info. I will try your method. There is a place called the Foam Factory in Troy, MI who sells all kinds of "acoustic" type foam. Maybe they would have the recommended material? The "Tech Notes" I offer are really just to help guitar players to address issues without getting too technical. Maybe Tech Notes is the wrong title? I do apologize if it appears I am propogating inaccurate information. I just want to help players get the most enjoyment out of playing music and be aware of some of the pitfalls we all run into. For those more technically oriented (including myself), could you suggest places to find plots/graphs demonstrating the dispersion patterns in this discussion? I am not having any luck at the guitar speaker mfg. websites (Celestion, Eminence, etc.)

    Thanks for your knowledge and insight.
    Bruce Egnater
     
  20. Jay Mitchell

    Jay Mitchell Member

    Messages:
    4,874
    Joined:
    Oct 13, 2007
    And you won't. Those companies manufacture transducers, and they never take directivity data on their products as used in actual enclosures. As an OEM customer for transducers, I can tell you that none of the OEM manufacturers are equipped to take the kind of data you're asking about. They consider it the responsibility of their customers to take such data, and I'm inclined to agree.

    When I was doing the R&D for this "project," I took on- and off-axis response measurements of the untreated speaker and the two thicknesses of foam. I may have saved that data. If so, I'll try to generate some comparative graphs and make them available. I can't make any promises until I get to the office and see what I've saved from that work.
     
Thread Status:
Not open for further replies.

Share This Page