Lynn,
Sorry to disappoint, but your blog posting seems very sensible, and puts in cogent form ideas that have been swirling round for a while. I think much of what you say fits in very closely with things that Linkwitz has said recently as well.
If one says that a consistent polar pattern, as well as a flat or slightly down sloping frontal frequency response is needed, you have a fairly small range of possible strategies.
1) Omni - e.g. like Linkwitz Pluto.
2) Wide baffle - PMS, Sonus Faber Stradivari, where the radiation is confined to frontal hemisphere, at least down to 300Hz or so.
3) Point source Dipole
4) Monopole line source - probably has to be more or less floor to ceiling to maintain a cylindrical wavefront to low enough frequencies, and very hard to keep horizontal dispersion good all the way up. May be impossible to realize.
5) Half - space line source - like 2, but vertically extended. Will be really big, probably the big infinity systems were like this.
6) Dipole line source. Again has to be very tall.
There might be others, but I can't think of them. Other classes of design are incapable of maintaining a consistent polar pattern across the whole audio spectral range.
Sorry to disappoint, but your blog posting seems very sensible, and puts in cogent form ideas that have been swirling round for a while. I think much of what you say fits in very closely with things that Linkwitz has said recently as well.
If one says that a consistent polar pattern, as well as a flat or slightly down sloping frontal frequency response is needed, you have a fairly small range of possible strategies.
1) Omni - e.g. like Linkwitz Pluto.
2) Wide baffle - PMS, Sonus Faber Stradivari, where the radiation is confined to frontal hemisphere, at least down to 300Hz or so.
3) Point source Dipole
4) Monopole line source - probably has to be more or less floor to ceiling to maintain a cylindrical wavefront to low enough frequencies, and very hard to keep horizontal dispersion good all the way up. May be impossible to realize.
5) Half - space line source - like 2, but vertically extended. Will be really big, probably the big infinity systems were like this.
6) Dipole line source. Again has to be very tall.
There might be others, but I can't think of them. Other classes of design are incapable of maintaining a consistent polar pattern across the whole audio spectral range.
Thinking more about compression drivers, an interesting marriage of old and new might be to take the 4" phenolic diaphragm from the JBL 2482/2485 (pn D16R2482, iirc) and drop it in a 2451 or 2452.
These are NdFeB-magnet motor assemblies with the new-tech shorter throat for proper mating to rapid flare-rate horns and modern waveguide designs, which reaps dividends in the treble. (The 2451 has the "coherent wave" curved-slot phase plug, but the newer 2452 seems to have gone back to straight slots.)
I don't know how high the D16R2482 diaphragm can be coaxed, but it would be interesting to see.
These are NdFeB-magnet motor assemblies with the new-tech shorter throat for proper mating to rapid flare-rate horns and modern waveguide designs, which reaps dividends in the treble. (The 2451 has the "coherent wave" curved-slot phase plug, but the newer 2452 seems to have gone back to straight slots.)
I don't know how high the D16R2482 diaphragm can be coaxed, but it would be interesting to see.
PigletsDad said:
Corner placed controlled/constant directivity will also work - this is the setup I'm currently working with. I have Yorkville U15's which are a 60x60 conical operating down to 300 mated to a 15" midbass. They are placed directly in the corners, aimed to cross just in front of the listening position. Directivity gently broadens from 60x60 to 90x90 which gives you a taper in your power response with basically flat on-axis. Obviously very installation dependent and requires dealing with corner-mode excitation and first reflections, but IMHO this can be a more successful approach in small/difficult rooms than the typical compromised 'a little bit out in the room' positioning.
One catch with that driver might be the long gradual-taper (no taper?) throat that's built into the driver. Looks like 3"-4". The initial flare rate might be <200Hz. So mating it with a smaller, horn/WG could result in a mis-tuned system.
I'm a big fan of Radian, Plus-One Engineering, etc. designs that build a rapid initial flare rate right into the driver.
Magnetar said:
I'm getting very close to ordering drivers this month. High probability of getting the following:
HF:
2X Double-height RAAL tweeters, 102 dB efficient, already on hand
Mids:
4X 18Sound 6" 6ND410
2X GPA 390 Phenolic compression driver
2X AH-550 Azurahorn with 1-3/8" adapter plate for Altec 288/290
Midbass:
2X 18Sound 12" 12NDA520
2X Tone Tubby 12" Alnico 16-ohm
2X Tone Tubby 12" Ceramic 16-ohm
Bass:
4X Selenium 15" 15PW3-SLF
4X Madison 12" Knight
They won't be all used at the same time, of course - more like a process of evaluation, measurement, audition, and seeing which ones play well together without an absurdly complex crossover (no plans for 24dB/oct active crossover, except possibly for the lowpass filter for the Bass section).
P.S. Magnetar, good find on the Eminence Li'l Buddy 10" hempcone driver - way better than the Eminence 12" hempcone, that's for sure. Looks like your experience with the Li'l Buddy mirrored mine with the 12" Tone Tubby - vivid and expressive tonality without the severe colorations of most musical-instrument or whizzer-cone speaker drivers.
P.P.S. Thanks for the information about the Clarisonus site - I've passed it along to John Atwood, blogmaster. I don't know who he is using for hosting the site - maybe John needs to change hosts.
I use a Mac Pro Quad Xeon, and have noticed the site doesn't stop loading - this is a behavior I've seen before when sites try to force ActiveX scripts into the unreceptive OS X environment (which don't run ActiveX, don't have a Windows folder, and don't have a system-wide Registry file).
Lynn Olson said:
Looking good! Looks like it might be within reach $$ of most builders too. I tried a bunch of different things here in the last few weeks and still can't find anything that satisfies like the open baffles. With your expertise in speaker design with that group of drivers this speaker will be dynamic, toneally correct and image like a SOB.. a real tour de force.
The baffles - I like Maple plywood braced with pine - try and isolate the drivers from the baffle - I use lots of mortite - there has to be a better way?
What will you name it?
Magnetar said:
Well, the Mother of Tone site is worthy of very careful reading. I thought there was a pretty high BS factor until a before-and-after treatment of a compressed-carbon turntable mat - and whoa, nellie, what a difference!
Without the lacquer, you get HiFi; after the lacquer, you get music, with beautiful tonality and very natural, lifelike dynamics (crisp and vivid, not slow at all). Sounds stupid but it's really true. By far the most cost-effective tweak for any system. Just slap this stuff on your plywood and you will be flabbergasted - it can be a bigger difference than the drivers themselves. Stupid but true.
There's something going on with organic-sourced materials - hemp cones, phenolic diaphragms, wood lacquers, wool damping - that is hard to describe but instantly audible. The sad people who only listen to Stereophile test records and never to live music (and I'm sure we all know people like that) can't hear this, but anyone who loves live music performed with enthusiasm hears it instantly.
My only guess is that human beings evolved listening to sounds in the forest, and are accustomed to the sound of certain materials, and don't care for the sound of others. Aluminum cymbals? Yech. Fiberglass violins? Urp.
I'm of the school that thinks coloration, per se, can never be fully removed - so the residue has to be a type that is sonically pleasing, and is compatible with the fabric of the music. Not that I think deliberate coloration is desirable - no, not at all. But there's always going to be some that's there - in a preamp, CD player, power amplifier, and loudspeaker. The goal of complete elimination is noble but also fruitless, and can lead to an intellectual blind spot regarding the inevitable residue. The residue of coloration - which is certain to be present - must be musically consonant.
James Boyk, a former CalTech professor, puts it very well in his Capturing Music seminar:
The Ear of the Beholder essay is especially good and a principle I've always felt very strongly about: that beauty is one of the most important qualities of recorded sound, and is sadly very rare in hifi systems.
"The goal of complete elimination is noble but also fruitless, and can lead to an intellectual blind spot regarding the inevitable residue. The residue of coloration - which is certain to be present - must be musically consonant."
Sadly the same blindspot appears to have resulted calling reproduced music which lacks tone colour and harmonic development "neutral" or worse "accurate"
I recall attending a friend's house where he had inserted another friend's TACT room correction/ DAC device. Whilst the room correction benefits were clearly audible the system also seemed to have the life blood sucked from it. So; the system with the TACT benefitted from a more linear frequency response yet sounded a significant degree further removed from the sound of real music simultaneously. I certainly would not choose to live with the results.
Regards,
Rob.
Sadly the same blindspot appears to have resulted calling reproduced music which lacks tone colour and harmonic development "neutral" or worse "accurate"
I recall attending a friend's house where he had inserted another friend's TACT room correction/ DAC device. Whilst the room correction benefits were clearly audible the system also seemed to have the life blood sucked from it. So; the system with the TACT benefitted from a more linear frequency response yet sounded a significant degree further removed from the sound of real music simultaneously. I certainly would not choose to live with the results.
Regards,
Rob.
I think the problem really is how one uses technology to appropriately address the issue of reproduction of a live performance. Many people misuse tools and misinterpret data, thus resulting probably the TACT type system previously mentioned.
One thing I would never do is use electronics means to solve what seems to be room problems. If we look at how the signal itself is altered without considering any math processing conducted on the signal, it should be quite clear. I'm sure if one recorded what is reproduced by the said TACT system and compared it with the original signal, it would probably be so bad that you could not line them up in time.
One thing I would never do is use electronics means to solve what seems to be room problems. If we look at how the signal itself is altered without considering any math processing conducted on the signal, it should be quite clear. I'm sure if one recorded what is reproduced by the said TACT system and compared it with the original signal, it would probably be so bad that you could not line them up in time.
The "moral hazard" of digital correction is the irresistable temptation (it's automated, after all) to correct the frequency and time domains - for one point in the room. The biggest hazard of all is HF correction above 3 kHz - these corrections are so physically small it only applies to one ear, and not the other!!!
This single-point correction unfortunately degrades everything outside the measurement point. At low frequencies, we're in good shape - the wavelengths are long enough to cover much of the room, and all of the listening area. This is the below-300 Hz correction I recommend - with a four-foot wavelength, the correction area is large enough to encompass most of the listening area.
But at a frequency ten times higher, things aren't as pretty - wavelengths are now inches long, and the ear reaches peak sensitivity to lateral differences in phase and amplitude in the 3 kHz region. Attempting to correct for faults in driver resonances, internal reflections from abrupt boundaries in a horn, diffraction from the horn-mouth and cabinet edges, or inter-driver phasing errors (from poor crossover design) is asking for trouble. Problems must be solved in the domain where they occur, not after the fact.
Time domain problems must not be solved in the frequency domain - they have to be solved at the source, which will be reflection and diffraction as the wavelength expands from the diaphragm, as well as standing-wave modes on the diaphragm and surround. Frequency domain problems result from resonances in diaphragms and acoustical standing waves in cabinets and horns. Nonuniformities in dispersion (narrow spikes) are caused by reflections and diffractions from sharp boundaries as the wavefront expands from the diaphragm. The real power of MLS, FFT, and TDS measurement techniques is being able to find these problems and resolve them (at the source).
It is foolish and counterproductive to attempt to solve these problems "after the fact" - very much akin to disguising the taste of spoiled food with artificial flavors. It may sell, it may be profitable, but it's not ethical, nor is it a quality product.
High-quality audio, like high-quality food, is the result of good ingredients skillfully put together in a pleasing manner. You're not going to get to beauty by not using good ingredients, nor by using a shoddy "this doesn't matter much, nobody will notice" approach to doing the task. Imagine how poorly a musician would play if they took that approach to practice.
P.S. Here's a wonderful thread by Jack Bouska if the good readers here haven't already seen it. I have a somewhat different set of design criteria (more emphasis on rapid time decay and a preference for DHT-triode amplification), but there's plenty of good info here.
This single-point correction unfortunately degrades everything outside the measurement point. At low frequencies, we're in good shape - the wavelengths are long enough to cover much of the room, and all of the listening area. This is the below-300 Hz correction I recommend - with a four-foot wavelength, the correction area is large enough to encompass most of the listening area.
But at a frequency ten times higher, things aren't as pretty - wavelengths are now inches long, and the ear reaches peak sensitivity to lateral differences in phase and amplitude in the 3 kHz region. Attempting to correct for faults in driver resonances, internal reflections from abrupt boundaries in a horn, diffraction from the horn-mouth and cabinet edges, or inter-driver phasing errors (from poor crossover design) is asking for trouble. Problems must be solved in the domain where they occur, not after the fact.
Time domain problems must not be solved in the frequency domain - they have to be solved at the source, which will be reflection and diffraction as the wavelength expands from the diaphragm, as well as standing-wave modes on the diaphragm and surround. Frequency domain problems result from resonances in diaphragms and acoustical standing waves in cabinets and horns. Nonuniformities in dispersion (narrow spikes) are caused by reflections and diffractions from sharp boundaries as the wavefront expands from the diaphragm. The real power of MLS, FFT, and TDS measurement techniques is being able to find these problems and resolve them (at the source).
It is foolish and counterproductive to attempt to solve these problems "after the fact" - very much akin to disguising the taste of spoiled food with artificial flavors. It may sell, it may be profitable, but it's not ethical, nor is it a quality product.
High-quality audio, like high-quality food, is the result of good ingredients skillfully put together in a pleasing manner. You're not going to get to beauty by not using good ingredients, nor by using a shoddy "this doesn't matter much, nobody will notice" approach to doing the task. Imagine how poorly a musician would play if they took that approach to practice.
P.S. Here's a wonderful thread by Jack Bouska if the good readers here haven't already seen it. I have a somewhat different set of design criteria (more emphasis on rapid time decay and a preference for DHT-triode amplification), but there's plenty of good info here.
Rapid decay is one of my focus on speakers. Here is a link to where I posted some results.
http://www.diyaudio.com/forums/showthread.php?postid=1410441#post1410441
Slow decay is like noise to music. Maintaining the same onset transient and dynamic capabilities, the faster the decay, the cleaner and more focused the music. This is also one important criteria that effects the audibility of absolute polarity in music.
http://www.diyaudio.com/forums/showthread.php?postid=1410441#post1410441
Slow decay is like noise to music. Maintaining the same onset transient and dynamic capabilities, the faster the decay, the cleaner and more focused the music. This is also one important criteria that effects the audibility of absolute polarity in music.
The Bouska thread is fascinating if only for the insight it provides into an audio-alphamales mind: What jolly fun the chaps had listening to the Chesky. He left out the bit when they were beside themselves with happiness listening to the arrival and departure of steam trains.
Another beer chaps?
Another beer chaps?