At least, that test taught me that i now can use a much cheaper midrange for my 3 way project. And that, my dear friends, is an excellent news.
(sorry for ressurecting that thread, don't reply! haha)
(sorry for ressurecting that thread, don't reply! haha)
Want to know what that midrange might be ?
Probably not. But i'll tell you anyway.
Visaton BG 17
or
Visaton BG 13 P
Do you only now how GOOD is the quality/value of these German-designed-but-made-in-China drivers is ?
They're built for elevators, but man, onced EQd, they would smack the yellow out of the B&Ws!!
'don't believe me? refer to post numero uno.
Probably not. But i'll tell you anyway.
Visaton BG 17
or
Visaton BG 13 P
Do you only now how GOOD is the quality/value of these German-designed-but-made-in-China drivers is ?
They're built for elevators, but man, onced EQd, they would smack the yellow out of the B&Ws!!
'don't believe me? refer to post numero uno.
Basically: noboby smacks nobody.
What counts is the mechanical limits and the sensitivity.
So, yeah, the Visaton with their 92-93db 1w/1m are on their way for the big prize. And they look sooooo cheap! Stamped steel, tiny magnet, as sexy as an old rusty Beetle.
What counts is the mechanical limits and the sensitivity.
So, yeah, the Visaton with their 92-93db 1w/1m are on their way for the big prize. And they look sooooo cheap! Stamped steel, tiny magnet, as sexy as an old rusty Beetle.
I designed and built speakers that use equalizer APO as the crossover. I can adjust each driver to fractions of a decibel.
I'm not surprised by the results, though they do seem to be stated in way that jabs at audiophiles.
But we know that what you pay for with better drivers is flatter frequency response over a wider range, higher excursion (bass), better materials, ability to play louder etc. Sometimes its just the driver you want isn't mass produced so you have to pay more.
So yea equalization can go a long way, and harmonic distortion isn't as much of an issue with modern reasonably well made drivers. And you could still tell a difference anyways.
But theres also issues of box noise, baffle diffraction, and dispersion and room acoustics which all matter a lot, and so just flattening some drivers on axis in a box won't make a really nice speaker.
I don't know of any 1500$ mid drivers, but I'd pay an extra 15$ for a dayton reference over something cheaper if it fit the application.
Like a lot of things theres a tendency to associate audiophiles with the worst examples of it. If anything audiophiles are critical of the nuttiness that certainly exists more than anything.
I'm not surprised by the results, though they do seem to be stated in way that jabs at audiophiles.
But we know that what you pay for with better drivers is flatter frequency response over a wider range, higher excursion (bass), better materials, ability to play louder etc. Sometimes its just the driver you want isn't mass produced so you have to pay more.
So yea equalization can go a long way, and harmonic distortion isn't as much of an issue with modern reasonably well made drivers. And you could still tell a difference anyways.
But theres also issues of box noise, baffle diffraction, and dispersion and room acoustics which all matter a lot, and so just flattening some drivers on axis in a box won't make a really nice speaker.
I don't know of any 1500$ mid drivers, but I'd pay an extra 15$ for a dayton reference over something cheaper if it fit the application.
Like a lot of things theres a tendency to associate audiophiles with the worst examples of it. If anything audiophiles are critical of the nuttiness that certainly exists more than anything.
1. Auditory capacities of humans are massively overestimated by audiophiles
This is a subjective claim, but nevertheless I would tend to agree. I would also agree with the more specific claim that audiophiles (i.e, old men) overestimate their own auditory capacities. I am lucky enough to have a teenage daughter who has proven herself able to hear subtle issues (component issues in crossovers, imaging, etc) that I cannot.
2. Frequency Response is King.
This is common knowledge insofar as it refers to on-axis response. Flat on-axis FR is known (from the work of Toole, Olive and others) to be the best predictor of listener preference. Because of the availability of good microphones and modeling software, a competent DIY designer can create, using arbitrary drivers, a speaker with quite flat on-axis FR. However, this is a necessary but not sufficient condition for a truly good design. The smoothness of the off-axis (polar) response is also critical, and separates a good from an excellent design. It is also common knowledge that some very expensive loudspeakers are poorly-designed, and some inexpensive loudspeakers are well-designed.
3. Once EQ'd, a 10$ midrange can mimic a 1500$ midrange, if within mechanical/electrical limits.
This is mostly correct. To lowest order, you can model a cone/dome radiator as an acoustic point source (this is what Vanderkooy did in 1986 to understand the impact of non-coincident drivers AES Journal Forum >> Power Response of Loudspeakers with Noncoincident Drivers-The Influence of Crossover Design). Much insight can be obtained with this approximation, even though it neglects the high-frequency directivity characteristics of real drivers. To next order, you can retain much of the directivity features of the driver by considering a flat, circular piston rather than a point source. This is a very good approximation even though it cannot distinguish different types of drivers of the same radius. Thus, the fundamentals of acoustics will virtually guarantee that two drivers of the same nominal radius will sound nearly identical. Of course, (1) the real cone profile (concave, convex), (2) cone breakup, (3) motor nonlinearities/distortion still have some effect, but among modern drivers with similar motors and cone profiles, there just isn't any room for different sound. That said, who on earth would pay $1500 for a midrange? Arguably, the best midrange in the world costs about $300.
4. DSP/EQ/in-room measure tools might be the best investment an audiophile can make in our era.
While you can use EQ to address the on-axis FR of a poorly-designed speaker, you cannot address geometric problems (like directivity mismatch). I would claim that well-designed 3-way speakers and room treatment are the key factors, and this is enough to achieve excellence. Conversely, no amount of DSP is going to save you from a bad design.
5. Others will have to continue spending hundreds and thousands for a natural uncorrected FR.
I don't really follow this. The Infinity Primus P363 is a very inexpensive speaker that does not require any EQ to sound good (Infinity Primus P363 Loudspeaker | The Absolute Sound). You cannot make it meaningfully better with EQ.
This is a subjective claim, but nevertheless I would tend to agree. I would also agree with the more specific claim that audiophiles (i.e, old men) overestimate their own auditory capacities. I am lucky enough to have a teenage daughter who has proven herself able to hear subtle issues (component issues in crossovers, imaging, etc) that I cannot.
2. Frequency Response is King.
This is common knowledge insofar as it refers to on-axis response. Flat on-axis FR is known (from the work of Toole, Olive and others) to be the best predictor of listener preference. Because of the availability of good microphones and modeling software, a competent DIY designer can create, using arbitrary drivers, a speaker with quite flat on-axis FR. However, this is a necessary but not sufficient condition for a truly good design. The smoothness of the off-axis (polar) response is also critical, and separates a good from an excellent design. It is also common knowledge that some very expensive loudspeakers are poorly-designed, and some inexpensive loudspeakers are well-designed.
3. Once EQ'd, a 10$ midrange can mimic a 1500$ midrange, if within mechanical/electrical limits.
This is mostly correct. To lowest order, you can model a cone/dome radiator as an acoustic point source (this is what Vanderkooy did in 1986 to understand the impact of non-coincident drivers AES Journal Forum >> Power Response of Loudspeakers with Noncoincident Drivers-The Influence of Crossover Design). Much insight can be obtained with this approximation, even though it neglects the high-frequency directivity characteristics of real drivers. To next order, you can retain much of the directivity features of the driver by considering a flat, circular piston rather than a point source. This is a very good approximation even though it cannot distinguish different types of drivers of the same radius. Thus, the fundamentals of acoustics will virtually guarantee that two drivers of the same nominal radius will sound nearly identical. Of course, (1) the real cone profile (concave, convex), (2) cone breakup, (3) motor nonlinearities/distortion still have some effect, but among modern drivers with similar motors and cone profiles, there just isn't any room for different sound. That said, who on earth would pay $1500 for a midrange? Arguably, the best midrange in the world costs about $300.
4. DSP/EQ/in-room measure tools might be the best investment an audiophile can make in our era.
While you can use EQ to address the on-axis FR of a poorly-designed speaker, you cannot address geometric problems (like directivity mismatch). I would claim that well-designed 3-way speakers and room treatment are the key factors, and this is enough to achieve excellence. Conversely, no amount of DSP is going to save you from a bad design.
5. Others will have to continue spending hundreds and thousands for a natural uncorrected FR.
I don't really follow this. The Infinity Primus P363 is a very inexpensive speaker that does not require any EQ to sound good (Infinity Primus P363 Loudspeaker | The Absolute Sound). You cannot make it meaningfully better with EQ.
I think its obvious that drivers tend to sound like what they're made of, especially the cone. I've never equalized two different drivers well enough to say that they sound the same. I suspect that its possible, but doing this in a home environment is tough.
Was this accomplished in the experiment?
Was this accomplished in the experiment?
Driver designers work very hard to have some useful bandwidth where their drivers do not 'sound like what they're made of'. Jon has limited his tests to this comfort zone, so it's similar to testing a bunch of amps rated from 40-400 watts at 10 watts. Barring real problems they should sound the same through the same speakers(if adequate for the 10 watts). When you want to crank it up, drop it 3 ft, or use it fifty years, or fix it, quality makes a difference.
The point is that how you use it makes a difference.
His fullranges as midranges are loafing, and can do an excellent job when you narrow the parameters of usage beyond what they are being marketed as. They are not abuse-proof, but I assume they will stay dry, warm, and safe, and should last.
Many fine speakers have had better results with cheaper drivers well implemented, and many poor speakers have had wonderful drivers poorly used.
The point is that how you use it makes a difference.
His fullranges as midranges are loafing, and can do an excellent job when you narrow the parameters of usage beyond what they are being marketed as. They are not abuse-proof, but I assume they will stay dry, warm, and safe, and should last.
Many fine speakers have had better results with cheaper drivers well implemented, and many poor speakers have had wonderful drivers poorly used.
Good point. A cone will definitely sound like "what its made of" when the cone break-up is significant. In this regime, it does not radiate as a solid, and the radiation pattern will become erratic and not meaningfully correctable with DSP. On the other hand, below the break-up regime, the cone is effectively rigid and thus the radiation pattern is independent of the cone material. When considering the classical ideal radiators (point sources, pistons, hemispheres, spheres) infinite rigidity is assumed. Below the breakup regime, however, the radiation pattern will still differ due to cone geometry. For example, an ideal concave cone tend to develop an SPL peak at ka = 1, whereas a convex cone develops a hole there (Suzuki and Tichy, JAES, 1981).
Thank you!
Jon, I just want to thank you for taking the time for taking on this very daunting task. I am in the same camp as you not expecting the outcome you got.
1. Auditory capacities of humans are massively overestimated by audiophiles (and probably by most humans as well)
2. Frequency Response is King.
3. Once EQ'd, a 10$ midrange can mimic a 1500$ midrange, if within mechanical/electrical limits.
4. DSP/EQ/in-room measure tools might be the best investment an audiophile can make in our era.
5. Others will have to continue spending hundreds and thousands for a natural uncorrected FR.
Obviously item #3 is a big deal and a game changer for many of us. I had a project where I used decent but cheap drivers with amazing results. I use active/DSP so that certainly helps, but I still figured it must have been dumb luck.
Now I think dumb luck may have less to do with it than previously thought.
Thanks again for the diligent work!
Jon, I just want to thank you for taking the time for taking on this very daunting task. I am in the same camp as you not expecting the outcome you got.
1. Auditory capacities of humans are massively overestimated by audiophiles (and probably by most humans as well)
2. Frequency Response is King.
3. Once EQ'd, a 10$ midrange can mimic a 1500$ midrange, if within mechanical/electrical limits.
4. DSP/EQ/in-room measure tools might be the best investment an audiophile can make in our era.
5. Others will have to continue spending hundreds and thousands for a natural uncorrected FR.
Obviously item #3 is a big deal and a game changer for many of us. I had a project where I used decent but cheap drivers with amazing results. I use active/DSP so that certainly helps, but I still figured it must have been dumb luck.
Now I think dumb luck may have less to do with it than previously thought.
Thanks again for the diligent work!
Talking about full range drivers for midrange units. A couple of years ago I was talking with Aleks Radisavljevic(RAAL owner/designer) and he tells me he's real big on Supravox fullrange drivers for use as midranges, says nothing beats them.
He didn't specify which series he prefers but I'm sure if you sent him an email he would tell you. We talked about the 165LB, 165 GMF and the 215RTF models.
Part of it is just the OCD nature of audiophilia; dsp is seen almost as cheating so its shunned. The full range people are the most extreme in this regard, some won't use a sub for the bottom two octaves even though it would make a huge difference.
Am I correct that its the narrowing of dispersion that tends to cause break up, not the other way around? This is why some drivers can decouple and and maintain wider dispersion (then a piston), because the driver material itself wants to flex and behave as two different radiating surfaces.