I dont understand this
Please explain this?
I use DSP and Eq more for mixing than correcting speakers but here is my understanding.
Lets assume a driver has a bad resonance problem at 1,000Hz and it shows on a frequency response as a big bad 10dB spike with a narrow Q ....An extreme example but I have seen this.
If I notch out this spike ( using Fab Filter) to flatten the curve I remove the problem...Yes?
As far as energy is concerned it does not matter if the spike is caused by a "pure" 1,000Hz tone, or a 500Hz tone or a 250Hz tone on second or third harmonics....The bottom line is that the energy spike emerges from the cone / suspension system at 1,000 Hz.....It remains inaudible AND un-measurable at lower or higher frequencies.....
If other spikes emerge, I would flatten them where the emerge.....
This is how every studio guy uses Eq.
iF i have missed an audiophile trick please do tell!
Please explain this?
I use DSP and Eq more for mixing than correcting speakers but here is my understanding.
Lets assume a driver has a bad resonance problem at 1,000Hz and it shows on a frequency response as a big bad 10dB spike with a narrow Q ....An extreme example but I have seen this.
If I notch out this spike ( using Fab Filter) to flatten the curve I remove the problem...Yes?
As far as energy is concerned it does not matter if the spike is caused by a "pure" 1,000Hz tone, or a 500Hz tone or a 250Hz tone on second or third harmonics....The bottom line is that the energy spike emerges from the cone / suspension system at 1,000 Hz.....It remains inaudible AND un-measurable at lower or higher frequencies.....
If other spikes emerge, I would flatten them where the emerge.....
This is how every studio guy uses Eq.
iF i have missed an audiophile trick please do tell!
No, you're hiding it, along with any program material at this point.
The "bell" is still there, waiting for someone to ring it. Meahwhile, you have a big hole in your music.
It will only measure "flat" if the bell rings enough to fill out the hole.
You're assuming that the system is linear. If you're a studio guy you'll know all about snares & high-hats rattling in sympathy to some loud bass note or drum kick. Same set of problems, just inside your speaker
Hydrogen,
As thoglette just described cutting the electrical signal only hides the problem and then only if there is not a harmonic frequency that is not a multiple of the resonance. You have done nothing that physically stops the resonance from being excited, the resonant property of the device has not been changed. Just as thoglette has just pointed out if you take two drums and have them in the same room if you hit one the other will also resonate, it is a mechanical coupling, it needs no signal input from hitting the second device. A harmonic in the electrical signal is all that is needed to excite the resonance at the frequency you have cut from the source material. So many think you can notch this out electrically but as was just said you are only hiding the problem but if you look carefully you will typically see this as a long decay time at that frequency in a waterfall plot.
As thoglette just described cutting the electrical signal only hides the problem and then only if there is not a harmonic frequency that is not a multiple of the resonance. You have done nothing that physically stops the resonance from being excited, the resonant property of the device has not been changed. Just as thoglette has just pointed out if you take two drums and have them in the same room if you hit one the other will also resonate, it is a mechanical coupling, it needs no signal input from hitting the second device. A harmonic in the electrical signal is all that is needed to excite the resonance at the frequency you have cut from the source material. So many think you can notch this out electrically but as was just said you are only hiding the problem but if you look carefully you will typically see this as a long decay time at that frequency in a waterfall plot.
Member
Joined 2003
- "All horns are pressure controlling devices..."
"The reciprocating flare horn is one which the rate of expansion of the horn remains constant, but the direction in which it expands changes from one plane to another for purposes of contouring the pressure fronts within the horn. Reference to Fig. 3-19 will indicate that the reciprocating horn firsts expands rapidly in the vertical direction with practically no expansion in the horizontal direction. A cut taken across this first horn section would show it to be a fairly well elongated rectangle oriented vertically. The sound pressure traveling down a channel of this type cross-section finds it relatively easy to to expand in the vertical direction, for the walls flare away in that direction. In trying to expand in the horizontal direction, however, the sound pressure meets the side walls of the horn, which restrict the wave from expanding, and consequently the sound pressure finds itself building up at these side walls rather than being allowed to freely expand. We might therefore say that looking into this rectangular section of the horn, there are high pressures against the unflared vertical walls and low pressure against the flared horizontal walls of the horn"
"Now, any one cross-section of a horn may be considered to be the "throat" area of the horn section in front of it: that is, one section feeds the next, and so forth. Thus we may look upon one horn section as the "driver" for the next section, with the very important attribute that if it is a driver that is pushing greater sound pressure against the side walls than against the top and bottom walls. If, at some discrete point along the horn, we change the direction of the flares so there is little expansion into the vertical direction and more into the horizontal direction, we have in effect a new horn with a flared directional advantage for horizontal dispersion, energized by a "driver" with already existing high sound pressure against the vertical walls anxious to be relieved into the horizontal direction. The overall effect of this reversal of the flare direction and differential pressures at the throat of the reversed section contributes to exceptionally wide horizontal dispersion"
"In the design of the horn, care is taken to so proportion the horn expansion that the final mouth of the horn is large compared with the wavelength of the frequency to be radiated. By this precaution, minimum diffraction of energy into the vertical plane takes place; it is all concentrated into the horizontal plane. In consequence of the energy being restricted from radiating into the vertical direction in free space and being concentrated in the horizontal plane, the overall efficiency of the horn is increased" etc
HI-FI Loudspeakers and Enclosures by Abraham B. Cohen, copyright 1956 John F. Rider Publisher, Inc. New York
I have several pairs of 811 horns and have used one pair as an on going experiment.
The first thing I tried was to cut the welds on the vanes because some people claimed it would reduce ringing. It didn't do as much as pressing a block of duct seal into the back of the horn bells. I also tried applying damping material to the outside of the throat area up to the mounting flange, which works OK if you don't have the horn mounted. If it is mounted, that provides a greater reduction in ringing than pretty much anything else.
The most effective mod I tried was to fill the internal weld seams with body filler and sand them smooth. This produced substantial audible improvements in the reduction of the sometimes harsh sound of the 811 horns. I switched back and forth between unmolested 811s and the filled versions and the difference was repeatable. I used Model 19s as the test bed.
It's surprizing how large the groove between casting halves is and how much misalignment there is between the halves. There is often quite a bit of weld spatter on the interior as well. I've been told that smooth is the answer for horn interiors so it can't hurt to clean them up at the very least.
The first thing I tried was to cut the welds on the vanes because some people claimed it would reduce ringing. It didn't do as much as pressing a block of duct seal into the back of the horn bells. I also tried applying damping material to the outside of the throat area up to the mounting flange, which works OK if you don't have the horn mounted. If it is mounted, that provides a greater reduction in ringing than pretty much anything else.
The most effective mod I tried was to fill the internal weld seams with body filler and sand them smooth. This produced substantial audible improvements in the reduction of the sometimes harsh sound of the 811 horns. I switched back and forth between unmolested 811s and the filled versions and the difference was repeatable. I used Model 19s as the test bed.
It's surprizing how large the groove between casting halves is and how much misalignment there is between the halves. There is often quite a bit of weld spatter on the interior as well. I've been told that smooth is the answer for horn interiors so it can't hurt to clean them up at the very least.
Last edited:
That looks like an EV CD horn with the slotted throat. It looks like the flare changes where these keep constant
Like the Emilar horns in this link ƒ~ƒbƒh�EƒŒƒ“ƒW�Eƒz�[ƒ“
University did it a lot with treble horns
what i'm working with is the EH153 Emilar horn with B&C 8ep1
The smaller EH330 http://i.ebayimg.com/00/s/MTYwMFgxMjM2/z/rc8AAOSw8lBTlmKm/$_1.JPG
Last edited:
When cutting out the vane welds, the gaps ideally need rubber wedges inserted, though most just apply some sort of rubbery silicone caulk. If done to the 511, then the small vertical brace further up the horn needs to be removed and some claim it should be done regardless.
Right, the entire inside of the horn ideally needs to be smooth, sealed to a very low friction; ditto the mouth transition/mounting flange; ditto matching driver, horn throat via align honing, replacing the cardboard washer with a machined one and all with wax paper or similar gaskets, which yields the greatest overall improvement IME.
Note too that mounting the horn to a proper baffle goes a long way to damping the mouth by itself, ditto adding a foam mouth 'extension' similar to the M19's, later Valencias, etc., though I found I preferred somewhat more damping than the more decorative ones Altec used.
GM
Right, the entire inside of the horn ideally needs to be smooth, sealed to a very low friction; ditto the mouth transition/mounting flange; ditto matching driver, horn throat via align honing, replacing the cardboard washer with a machined one and all with wax paper or similar gaskets, which yields the greatest overall improvement IME.
Note too that mounting the horn to a proper baffle goes a long way to damping the mouth by itself, ditto adding a foam mouth 'extension' similar to the M19's, later Valencias, etc., though I found I preferred somewhat more damping than the more decorative ones Altec used.
GM
Pooh,
that description if your read it the way it sounds seems to very well describe a bi-radial horn like the Altec mantaray or the JBL Bi-radial design where the horn flare trades vertical for horizontal expansion as it moves forward.
Zelgall,
I did an experiment many years ago and copied an Altec 311 horn and changed the material from the die cast aluminum to a foamed plastic urethane material and it was night and day difference in how they sounded. Just getting rid of the metallic surface made a huge audible difference. This started me down the road to making and producing urethane horn lenses. That was back in 1976
that description if your read it the way it sounds seems to very well describe a bi-radial horn like the Altec mantaray or the JBL Bi-radial design where the horn flare trades vertical for horizontal expansion as it moves forward.
Zelgall,
I did an experiment many years ago and copied an Altec 311 horn and changed the material from the die cast aluminum to a foamed plastic urethane material and it was night and day difference in how they sounded. Just getting rid of the metallic surface made a huge audible difference. This started me down the road to making and producing urethane horn lenses. That was back in 1976
That's what first came to mind when I read it. I was looking for a rear end shot of a Manta horn, that's the best I could find. The Emilar horn seems to change flare rate in only 1 dimension, so is probably closest to the description.
I thought somebody would have addressed this by now. Even if the harmonic distortion excites the resonance, HD is supposed to be low in a proper design. For example, if there is a 2nd order HD product of 500 Hz that is exciting a 1000 Hz resonance, if the product is below 1%, it will almost completely be masked.
Also, fixing a resonance generally fixes the time domain also, assuming the system is minimum phase in that region. For ages, people have been using notch filters on woofers to tame the out of band peaks. Of course we would prefer to have drivers with no problematic resonances in the passband. Most of the time, it is difficult to properly tune notch filters to exactly reverse the resonance. This can lead to audible problems. Also, sometimes the peaks are due to breakup in the cone, i.e., different regions on the cone not moving together. In that case, the peak can be seen on one axis but not on other angles. This you can't fix with EQ.
Also, fixing the peak in amplitude also fixes it in the time domain. If your subwoofer is underdamped and producing too much output near it's resonance, electrically correcting the magnitude will fix the ringing in the time domain. What you can't fix electrically is the directivity of drivers or if the peak/trough is the result of a reflection. The system is non-minimum phase at that point and EQ will not make any change in that situation.
Last edited:
I should have mentioned the silicone filler. I did that as well based on a recommendation by you. I eventually cut the vanes out completely and ground them flush. As I was just using this pair as a tester, it didn't seem like a loss at that point. I found no difference in how they sounded with the vanes removed compared to my unaltered 811s. I may do this to a pair of 511s eventually.
I also removed the cardboard gasket and used waxed paper in it's place. The throat in my 802-8Gs have the same exit diameter as the entrance to the horn throat so the thinner no stick gasket was fine. The mating surfaces on both were very flat and I could detect no air leakage.
It's amazing how much better you can make a perennially maligned horn like the 811 sound with a little experimentation and elbow grease.
Interesting. I never went to that much trouble with the 811 or 511, just moved on to the multi-cell horns. Nice to know real improvements can be made. 
The most I did was to add a soft edge of foam or beach towels - and that made a slight audible difference and was easy to see in measurements.
The most I did was to add a soft edge of foam or beach towels - and that made a slight audible difference and was easy to see in measurements.
It's the non-minimum phase assumption where things go wrong. If the diaphragm is "well-behaved" ... that is, moving as one unit, with no funny business from the dust cap, surround, or spider, yes, then it's minimum phase. But it's also usually flat if it's well-behaved, so we're right back where we started. The minimum phase assumption tends to be true when the (single) peaks are broad and gentle, not true when we see multiple narrowband peaks or an overall chaotic region. (It takes accurate time-domain measurements that are free of local reflections, which includes mike-stand reflections, to see the chaotic regions.)
Similarly, when horns are non-flat, the error is not always minimum-phase. It is not minimum phase when there are multiple paths through the horn, if standing waves are present, or if there's diffraction at the horn mouth or created by sharp-edged features inside the horn. These are not minimum-phase, and frequency equalization will make the time domain problem worse, not better. Problems of this sort are audible with vocals or string tone, which become very raspy and sibilant when there's time domain energy storage.
The figure-of-merit that I use is the overall decay time (less than 2mSec preferred), and whether it has a chaotic appearance or not. The waterfall/CSD is a useful complement to see if the resonances are narrowband or chaotic.
The point about directivity is a good one. It's one thing when the directivity changes smoothly, quite another when there are sharp spikes and nulls that are frequency-dependent. These are the result of diffraction, radiation problems from the horn, or non-pistonic motion from the diaphragm. Horn theory assumes uniform behavior from the diaphragm; things fall part when the diaphragm starts moving in several directions at once.
My feeling is that measurements are most useful when we use them to find problems in the physical systems of the horn, diaphragm, or direct-radiator loudspeaker. Once discovered, the most thorough solution is at the mechanical and acoustical level, where the problem is happening.
When we hear a loudspeaker or live music, we perceive it as a unitary sensation, not as THD, frequency response, or time-domain errors. It just sounds wrong, or unnatural, or different from a physical, real-world musical instrument. It sounds fake, or as I sometimes put it, a sound not found in nature. Phasey sounds don't happen in the real world; they're artifacts of large phase angles between widely spaced radiators. Diffraction artifacts can really mess up the sound of voices, giving them a very electronic and unnatural sound, and this is a very common problem in expensive high-end speakers. I hear it in nearly every room at hifi shows, and don't like it.
Correlating these impressions of "wrongness" to specific measurements is not always easy, and sometimes no correlation can be found. But we can use the same measurements to find system errors, and get rid of them.
Last edited:
Some horns have more diffraction and reflection than others, some have less. Audibility of energy-storage artifacts probably depends where it falls in the audio spectrum, and to a lesser degree, the individual listener. My sensitivity to time artifacts is probably higher than average.
As for a potential 200~800 Hz horn, I'll be contacting my friends Bjorn Kolbrek and Gary Pimm, and see what they think. It'll come down to that or a pair of 15" drivers in closed boxes. The latter isn't as glamorous, but will meet the technical goals of sensitivity and headroom.
As for a potential 200~800 Hz horn, I'll be contacting my friends Bjorn Kolbrek and Gary Pimm, and see what they think. It'll come down to that or a pair of 15" drivers in closed boxes. The latter isn't as glamorous, but will meet the technical goals of sensitivity and headroom.