I would define 4th order as par, not steep, not shallow. 24db/octave still requires keeping better part of two octaves away from strong resonance/break up in avoiding IMD.
As seen in #347:
An externally hosted image should be here but it was not working when we last tested it.
Above crossover at 900Hz is accomplished with 1024 point FIR. This is steep, about 200dB/octave, and typical of what I listen to. With drivers flat for phase and frequency, and about 1/4 wave apart at crossover no ringing is heard anywhere in room. Driver corrections are calculated from measured IR and then convolved with crossover filters to build playback filters.
twest820:
This is conflicted. Subjectively claimed benefits? That don't agree with your own subjective preferences? A driver's directivity is physical property of driver, and what it is mounted to.
Regards,
Andrew
You don't, though experience with live sound and the ability to play an instrument provide a good reference---unplugged can be particularly useful as it eliminates a lot of variables. It's also my experience most characteristics that are supposed to be good for one genre or another aren't actually genre specific. Maybe you notice them more or less with certain mixes but there's not usually a major amount of masking. Harmon-Kardan found about 80% of a speaker's perceived sound quality can be explained by factors other than the actual sound quality but, once those are controlled for, there's a clear, statistically significant, subjective preference for speakers which do well in objective measurements. So you may want to measure some polars.
The overall optimum in a multidimensional tradeoff is rarely ideal for any particular dimension. I have tried 100-200dB/octave slopes via ReaFIR and, while interesting initially, didn't find them satisfying in the long term and eventually relaxed to around 40dB/octave. Not much difference between that and an LR6. It could be interesting to revisit steeper filters with tools like rePhase and an OpenDRC/miniSHARC but it's not currently a priority for me. I'd also like to try a coaxial magnetostat to minimize motion of the acoustic center over the tweeter to mid XO and see if that produces better results with steeper crosses. But nobody makes one and DIYing one is also low priority for now.
An externally hosted image should be here but it was not working when we last tested it.
Linkwitz Pluto type speaker with SEAS Prestige L16RN-SL and Peerless 830970 allows driver centers to be closer together than when on flat baffle.
Top firing woofer only sees reflection from far end of pipe, after sound travels through more than 70" of stuffing. Very little sound re-radiation. No panel resonance.
Forward firing woofer in box tower, distance to back panel is short. Back, sides, and top form incredibly good retroreflector, sending energy back at driver. Little distance for stuffing to absorb back wave unless packed too tight for good low end. This is further aggravated with supper light, acoustically transparent diaphragms found in full range drivers. Box sound.
Regards,
Andrew
YEs, one way to get the drivers close.
It is a tradeoff... with the small diameter pipe you have, there are going to be signficant side-wall reflections.
dave
It is a tradeoff... with the small diameter pipe you have, there are going to be signficant side-wall reflections.
dave
At what point of pipe are sidewalls reflecting energy back at driver?
Regards,
Andrew
Does not happen in fact (in my experience). Maybe if you tried to run it full-range you'd see something, but I doubt even that, since the back wave is symmetric all around the "enclosure" and any reflections that don't go "down the pipe" tend to cancel. You'll see spider and frame issues before any pipe reflections (but neither, of course, crossed at 1000 Hz.).
I'd sure like to see the evidence of that in an end-driven pipe. I don't doubt that it happens if you put the driver on the side of a long thin box and try to run it full-range, though.
FR is what i mostly do, but the audiable issues are in voices, which have pretty much petered out by 1 kHz.
Here is an example (with a 2" driver in a 4" pipe it wasn't a big issue, it starts impinging with the 3" scan now in the pipe, and that will have more clearance than the pipe pictured above.)
dave
Those pipes look rather short to avoid longitudinal reflection/resonances (which is what I suspect you are hearing). But in any case frequency response curves (which should show the ripple, wherever it is) would tell a lot more. Not to be too glib about it, but I don't see any evidence of lateral reflections in the pictures . . .
Damped till aperiodic. FR will not show the problems these have. Typically called cuppiness.
dave
Hard to do in that short a pipe.
I haven't heard "cuppiness" in any of the PLUTOs or PLUTO clones I've heard (including my own). Can't figure why your implementation as shown would have lateral reflections while none of the others do.
Magic, same design in one case incurs significant trade off, and in other design works just fine.
Aperiodic tuning must be complete only after "Cuppiness" is tweaked out by ear?
What is good test setup for demonstrating "Cuppiness" that my ears may learn to hear something that can't be measured?
Regards,
Andrew
Aperiodic tuning must be complete only after "Cuppiness" is tweaked out by ear?
What is good test setup for demonstrating "Cuppiness" that my ears may learn to hear something that can't be measured?
Regards,
Andrew
There is an awful lot of stuff that can be heard that cannot be measured.
It is well accepted that 10,000 hrs of serious training are required to become really good at any endeavour.
dave
That you would say that says alot...
dave
OK, please show me a setup where one variable changes that produces an audible sound difference which cannot be measured with a microphone.
Anybody.
Regards,
Andrew
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.