....It was rumored that many in the business found his regular numerous returns of out of acceptable spec parts annoying.
|
Before I mislead you, he didn't say that exactly. Though he might(?!) have been implying it. Remember John Dunlavy was much about wave form and ergo time fidelity. In an article in Stereophile he confided that step response was his go to measurement, (I'm paraphrasing here) when the step response is good everything else will fall in place. So, hmmm?, I really don't know. |
Unsound,
That's a really interesting perspective. As a DIY speaker designer, I never bothered to match drivers, and had to trust the manufacturer's provided high quality, and very consistent parts. I cannot afford to buy 100 midranges to sort.
The idea of not just using DSP for crossover and speaker design, but actually digitally compensating for minute individual differences in driver response is pretty forward thinking.
Now I have a new project. :)
Best,
E
|
^While I agree, there is added interest in that the mod was done by an ex DAL employee. In one of my last conversations with John Dunlavy he said he was excited by the potential of going forward with digital active cross-overs and individual Class D driver amplification. He did say that the he still waiting for better chips to work with and that the initial offerings would necessitate high development cost that would have to be passed on to the consumer. On the plus side; once that development occurred that it would eventually and progressively cut down on the labor costs. Remember each speaker that Dunlavy made at that time was made with complex cabinets, selected and sorted parts, and each and every speaker hand tuned to model specs. He expected to be able to eventually offer better products at much lower prices. |
erik, I reply not as an engineer but as someone who owned Duntech speakers for 19 years, auditioned several Duntech and DAL models, and had the pleasure of personal conversations with John on two occasions.
John Dunlavy was one of our most respected speaker designers. He believed strongly in the benefits of time/phase coherent speaker designs and so all his models were based upon that. To achieve that time/phase coherency with his multi-driver speakers required a rather complex crossover.
My advice would be to leave the stock crossover as is, or else find some other brand if you aren't satisfied with the Dunlavy and do your experimenting with that. I believe you will only mess up your Dunlavy and likely ruin them for resale.
|
The original Dunlavy's didn't appear require that much complexity. As for me, I find that at I consistently prefer speakers that are time coherent to those that aren't. YMMV!
|
@unsound
That's kind of what I am asking. From a technical perspective, recreating a square wave, or near perfect impulse response with purely passive, multi-way speaker is really difficult.
Worthwhile?
|
Perhaps I'm missing something? Is this just supposed to be an academic exercise? It would appear to me that at great extra complexity and cost one has made what might be marginal improvements over the simpler, less expensive alternative? |
Linear Phase is preferable except when high Q filters (aggressive filter) are used.
High Q filters can result in audible pre-ringing and a minimum phase filter will have no pre-ringing - so a snare hit will sound more natural with a minimum phase high Q filter.
However, I think high Q filters should be avoided period!
So if you are dealing with low Q filters then linear phase is always the best. The reason linear phase is best is because it preserves the relative phase information in the audio. The timbre of a musical sound that covers many octaves or percussive instruments that have a wide spectrum response will be preserved faithfully by a linear phase filter. Minimum phase changes the relationship between various frequencies and can really mess up correct timbre.
In a speaker crossover with a low Q filter (gentle filter) I would recommend always linear phase (so as to preserve timbre especially in higher frequencies) |
For people who don’t know, Bill invented OmniMic and XSim. He knows his stuff, so questions below are for my own education. I ask all of this not really knowing if ideal phase is all that beneficial. From what I have heard, I could not tell a difference.
I’m not sure if you are arguing room acoustics or off-angle driver phase matching.
Room/Power Response
============================
You are, of course, right in the sense of the overall power response, but isn’t the goal of good speaker placement and room acoustics to have a nice delay between the initial signal and the first reflections?
Off-Angle
===========================-
Take something like the Dunlavy. Assuming your ears stayed at tweeter level wouldn’t being off-axis be fairly benign? And isn’t it better to start from "ideal" than a normal 3-way alignment?
Of course, I can see a true coaxial being ideal here, like Thiel or Kef, where the alignment stays consistent over a broad range.
Best,
E
|
Here’s the (an) issue with that. They may do those nice square waves at one point in space, or maybe on a line from the speaker to the listening seat. But what about what comes out of it off-axis? All it takes is to face a speaker away from you to know that the sound level really doesn’t drop down all that much. IOW, you hear the wavefronts that leaves the speaker off-axis -- not just the one going out perpendicular to the baffle. Linear or minimum phase only makes sense with a coaxial or point source speaker, otherwise it’s just a techno-game (IMO).
|
