Wow, that's one "over the top" review!
After looking at the review and the manufacturer's "theory" link, I think the reviewer may have missed a very important design decision that WILL affect the overall performance of this speaker :scratch: . This speaker company is VERY convinced that 4th order Linkwiz Riley crossovers are the best choice bar none. :shake: Unfortunately that flies in the face of most other high end speaker designer's findings. Those type of crossovers have many unwanted side effects d ...
Hello everbody!
My name is Bob Smith and I'm the president and chief engineer of SP Technology. I just stopped in to see what all the fuss was about. I guess Bill Roberts' review of our product created some controversy. I suppose I'm sticking my neck out on the chopping block here but if it helps dispel some partially incorrect notions then it's worth it.
First off, I'm not here to plug our products - don't have to. That's what reviews are (hopefully) for. Our products stand in their own defense. I am only concerned here with the underlying physics of design. I would hate to see a potential customer turn away from us or any other manufacturer's product because he or she was led to believe a partial truth. Reference the above quote. I offer the following for the more technically minded individuals out there that do consider the science behind the design in their decision making. I'm certainly not trying to impress anyone with my technical knowledge either.
Now first off Mr. Ehider, I'm not here to slam you or suggest that you don't know what you are talking about. Quite the contrary, you are correct, the fact is well established that higher order filters introduce larger amounts of phase shift than do lower order designs. But the complete analysis of a design has to take into consideration a multitude of variables, some of which will alter a simple "textbook" interpretation of the facts. The universe is full of complex phenomena that tend to represent "apparent" paradoxes, refusing to yield their secrets to simplistic linear analysis.
In crossover design, one is concerned with the relative phase between drivers more than with total phase rotation - or at least he/she should be. The need here is obvious. If the drivers are not in-phase at crossover then they will not sum to a maximum amplitude. This does not mean that the speaker will not exhibit a relatively flat response on some pre-designated vertical axis. If they do though, it will be due to the fact that one or both drivers have a peaked response at crossover to begin with. This is how many designers manage to get a flat response out of inferior or problematic drivers. The problem with this approach is that as you move off vertical axis, the response becomes very non-flat and is different depending upon if you move up or down from that point.
Also, it is a well known fact that the first order (6dB/Oct.) crossover theoretically sums to perfectly flat magnitude AND phase response. The problems with this crossover are many though. The frequency and power demands on the drivers chosen for this type crossover are extreme and for full range reproduction, at least one midrange driver is virtually unavoidable. For arguments sake, if we throw money at the power and frequency response problems, we can make them go away. What we can't get rid of though is that third driver or the extreme vertical off-axis errors that result.
Combining the output of two drivers employing a true first order response (which many claim to but don't) presents enough of a challenge to the designer in his/her attempt to minimize the severe off-axis colorations in the vertical plane. Combining three drivers in this manner virtually guarantees that the "sweet spot" will be very narrow vertically unless the listener is seated a great distance away from the speaker.
One high-end manufacturer has offered what appears to be a solution: make use of coincident driver technology. If you design the tweeter to co-exist at the center of the midrange driver then you will have eliminated the source of vertical (or horizontal) lobbing errors. This concept is by no means new except for the fact that, in this instance, it has been applied with a first order crossover design. This would seem to be an ideal solution to the first order design’s inherent lobbing error issue except that, upon closer inspection we find a whole new level of design complexity and the resulting compromises inherent in co-incident driver design.
Seeing such technology is, for the most part, public domain and quite mature, virtually all loudspeakers would be built using co-incident drivers if they truly offered a final solution to this problem without introducing even worse ones. For further reading on the subject an engineering white paper will soon be available on the SP Technology website.
As I have said, lower order networks often (
but not always) exhibit many problems that I am certain many enthusiasts and certainly most professionals would prefer to avoid. I could go on further into the effects of vertical lobbing errors, the resulting non-flat reverberant field produced in the listening space, the systems unnecessarily high sensitivity to room placement and the severely limited listening positions that also result, but that is another post.
Absolute total phase rotation together with frequency amplitude response will reveal the total "group delay" through the system. This, I believe, is what Mr. Ehider is referring to. And I strongly agree, group delay has been proven to be audible under certain conditions. Its most profound effect is upon transient response. The percussive attacks of musical instruments sound constrained and lack realism when their waveforms are "ripped apart" and smeared by being passed through filters exhibiting large amounts of group delay. We could conjecture about 3-D information and soundstage being degraded but I am convinced this is more a function of diffraction errors, higher distortion levels, poor resolution and inferior dispersion than the minor delay effects in the 0 - 2 millisecond range that crossovers produce.
Mr. Ehider, what you and many others may not realize is that much research by others has gone into the study of the Linkwitz-Riley 4th order crossover with regards to its effects on transient response. One study I am aware of (can't recall the names right now) actually showed that it exhibited the best overall response compared to all others except the true first order type. It does this while also offering all of its many other benefits with regards to driver protection, reduced distortion and idealized vertical lobbing response.
What you may also not be aware of is that to a certain degree, the delay it does introduce can be somewhat corrected for by simply setting the high frequency driver back further form the plane of the lower frequency driver. This approach is not a cure-all but it does help improve time alignment considerably. Another peculiar fact; many do not know that under certain conditions, a highly linear phase response can be obtained by combining a 4th order low-pass network (to the low frequency driver) with a high-pass first order network (to the high frequency driver). I know this to be fact as I have produced such results in my own laboratory. This completely flies in the face of conventional thinking.
Again Mr. Ehider states "
It's one HELL of allot easier to design a great measuring speaker (frequency response wise) when you choose 4th order!" I must ask you, have you ever tried to design AND implement into a massed produced speaker product a 4th order network? All of the designers out there reading this know the answer to that question. A fourth order network implemented with the full compliment of componenets is anything but easy to produce. It is even harder to reproduce on a consistant basis due to component sensitivity. Try running a Monte Carlo analysis of the component tolerance variations and see what you get. I've heard of more than one designer stating that they'll never attempt a passive implementation of a fourth order design again! You did get the "HELL" part right though.
So you see, things are often not as simple as they may seem. It is wisest if one does not get caught up in the pet theories of others and pass judgment based on an overly simplistic view of the facts. A little knowledge is not the same as knowing.
To address one last small matter: Mr. Ehider made the comment that, "
This speaker company is VERY convinced that 4th order Linkwitz-Riley crossovers are the best choice bar none." This is not true. In fact, we went out of our way on our website to state that, "
Our position is to use what ever works the best in a given application. While it is true that we prefer to use the Linkwitz-Riley alignments due to their superior phase properties, we will not be restricted by academics in our quest for excellence. From time to time we may choose to use other alignments as well -- form always follows function. Natural Law will not permit any other approach and we ARE bound by it both philosophically and practically."
I sincerely hope the above information has been informative and useful. It is not to be interpreted as condemning or given in the spirit of "one-upmanship." I sincerely thank every individual that has posted here concerning our product, especially Mr. Ehider. I too struggle a bit with a tinge of cynicism from time to time. With all of the misleading advertising and unbelievable claims being made in the loudspeaker market, it's a wonder we're not all hopeless cynics.
Take Care All,
- Bob
