Well you are not listening - they are the most important, but not "all that are important".
I thought I did answer the question and I did mean the spider not the suurond.
I said "all 15" speakers that I have measuired". It can be fixed, I showed B&C how, but alas, they did not change because "no one else has complained". They once admitted that I find flaws in there products that no one else ever comments on. They always agree that they are there, they just never do anything about it. So why not use other drivers? because no one else is any better or any different. Some are just plain arrogant about small manufacturers like me. Its a real prblem in my situation.
Thanks Pano, again, for moving us forward with some actual experimenting. More data, less wank (Zilch).
I completely agree with your above statements. There is definitely something wrong when the whole spectrum is not panned together. The >700 Hz panning is easier to localize but you can still clearly hear that something is wrong.
Okay,
I accept that statement. I am a little surprised though I haven't seen or heard your speakers that you say you have little output below 50hz with your speakers. Is this intentional or just due to a limited enclosure size with a large 15" speaker? I think I read you were using a 15" speaker. I know many people are perfectly happy with 50hz as a lower cutoff frequency and most are happy with 60hz in reality.
I accept that statement. I am a little surprised though I haven't seen or heard your speakers that you say you have little output below 50hz with your speakers. Is this intentional or just due to a limited enclosure size with a large 15" speaker? I think I read you were using a 15" speaker. I know many people are perfectly happy with 50hz as a lower cutoff frequency and most are happy with 60hz in reality.
700hz
Actually lord Rayleigh already gave us model 100-years ago not perfect one, but essentially under 700hz sounds are localized via phase difference as interaural level differences are too small due ears being too close vs wavelength.
lots of people like vinyl records that have quite bad crosstalk issues in lower frequencies but nobody is complaining vinyl images badly because of it.... area where cd offers technically huge superiority but our ears arent up for the task to notice night&day difference.
Panning itself can be bit tricky and make things sound "wrong" known as long as stereosound
Actually lord Rayleigh already gave us model 100-years ago not perfect one, but essentially under 700hz sounds are localized via phase difference as interaural level differences are too small due ears being too close vs wavelength.
lots of people like vinyl records that have quite bad crosstalk issues in lower frequencies but nobody is complaining vinyl images badly because of it.... area where cd offers technically huge superiority but our ears arent up for the task to notice night&day difference.
Panning itself can be bit tricky and make things sound "wrong" known as long as stereosound
The thing I don't get on the Keele presentation is his CBT is presented as getting twice as big due to floor reflection but he doesn't present the same examples with the straight line array and the power tapered line array. (page 63)
His example is presented as a 60 degree curved array (while the speaker is only half of that) and it would be only fair if the sims showed a 4 m long or 6m long or even infinite long (counting ceiling reflection as well) sim for both types of the straight arrays. It would also be interesting to see what a floor reflection does to the JBL type array's. as they seem to be power and time tapered toward the top and bottom side.
This is just a way of presenting data to make your own version shine I.M.H.O.
I like the data as presented but basically he's just mimicking a point source due to the natural time delay of the curved array and the power tapering. But to have the highest output at floor level seems strange to me.
Last edited:
Yes, though the prototype is extremely similar to the PE build.
Yup, page 102.
http://www.linkwitzlab.com/Keele - Introduction to CBT Loudspeaker Arrays.pdf
I used 15" in the Summa, but I don't sell Summas anymore. I ported them at first and they would do 40 Hz without much trouble. But years after the first design I concluded my research into the modal aspects of small rooms and concluded that multiple subs was more than just a good idea, I believe that it is essential.
I started recommending that people not use my speakers without subs. Hence it just became rational to delete the ports and sell them as closed box, which end up mating with subs better and they EQ more easily. I now recommend everything be close box. It is just so much easier to deal with.
So it was kind of something that I backed into, but now I think that it makes perfect sense. Everyone who has done this approach is very happy with it.
Abbeys have always been a closed box 12" and are about 3 dB down at 60 Hz. I know people who use them without subs - I wouldn't. They could be ported for another 10 Hz lower cutoff, but then they must not be EQ'd anywhere near their cutoff.
Pano,
perhaps you are the only one of us who has done this shifting test with noise AND with the wolf. I read your comments as if you heard a (qualitative?) difference betweeen shifting noise and shifting speech. Could you explain what was the essence of that difference for you? And would you dare to offer an explanation for that?
Rudolf
perhaps you are the only one of us who has done this shifting test with noise AND with the wolf. I read your comments as if you heard a (qualitative?) difference betweeen shifting noise and shifting speech. Could you explain what was the essence of that difference for you? And would you dare to offer an explanation for that?
Rudolf
Pages 58-60 mention the conditions for those sim.s. Page 59 Ex. 1-6, page 63 Ex. 1-6, appear to be the same to me.
Those sim.s (63 & 64) are not presented reflecting to anything - it's totally anechoic (and at 1 kHz only).
Pages 63 & 64 are just about combing and dispersion at 1 kHz with the number of small sources/drivers simulated. Other designs that utilize actual line sources won't behave in the same manner, nor will those that utilize a waveguide that properly "load" down to 1 kHz or lower.
He mentions a bounded/reflective condition on page 57 (right side) and again on page 69. Other than that, all the previous pages discuss un-bounded/non-reflective conditions.
http://www.linkwitzlab.com/Keele - Introduction to CBT Loudspeaker Arrays.pdf
In other words the sim should have been about the CBT, both cases, the curved from Keele and the JBL time aligned one. So it should have been half the arc shown in that sim.
Both show a 60 deg arc but the speakers are only half of that. They need the floor reflection to make it like the one in the sim. In that case the floor reflection should be added to the others as well to make it a fair comparison.
Without the floor reflection the CBT is not the speaker simulated on those pictures.
Little output below 50Hz ? Sounds compatible to these.An externally hosted image should be here but it was not working when we last tested it.
I'm still kind of
here.. Those sim.s are only about very similar arrays with the differences presented on page 59. "JBL time aligned one"?
Conditions 3, 4, and 5 are 60 degree "arcs" - and the subsequent sim.s presumably conform to that condition. No reflective condition is required, nor is it presented.
Is there perhaps another page (or group of pages), that you are referring to?
You were the one who got all starry eyed when this research come up. Most of us here found it rather hard to believe, or that it applies much to loudspeakers. That's the extrapolation I object to.
If someone is looking for a justification of lack of control under 700Hz, then it makes perfect sense.
I don't claim to have done the research that Griesinger has, but my experience and 30+ years of work in audio make me question the application of his research to music reproduction. I've done a few simple tests that easily put the whole "location is dominate above 700Hz" thing into question. I know that you will never believe such tests, that's OK. I just hope to show that how easy it is, with a few simple tests, to poke holes in the idea.
I've never followed the common belief that bass is omni-directional, now we want to push that up to 700Hz?
I can do tests until the end of time, some people will simply never give them any consideration or tests things for themselves. I accept that. But I do not take what someone else has published at face value. If it run counter to what I know, then I'll test it.
I'm still kind of
Those sim.s are only about very similar arrays with the differences presented on page 59. "JBL time aligned one"?
Conditions 3, 4, and 5 are 60 degree "arcs" - and the subsequent sim.s presumably conform to that condition. No reflective condition is required, nor is it presented.
Is there perhaps another page (or group of pages), that you are referring to?
The number 5 is representing the JBL CBT:
JBL :: Product Family
a straight array with digital time alignment and power tapering to make it seem like the curved array from Keele. But on page 140 he shows a delay curved array and that one is curved one way, half of what is in the sim. No curving toward the bottom half. Fot that it needs a reflective floor.
So numbers 3 and 4 are presenting the Keele CBT array. One without (3) and one with shading (4).
As you know the Keele CBT looks like this:
An externally hosted image should be here but it was not working when we last tested it.
That's only half the array used in the simulation. So why show an array double the size in your simulation? If you put the array from the simulation on a reflecting floor the response isn't going to be like the sim.
So why simulate the 60 deg array? That one needs to hang somewhere between the floor and ceiling to work well.
Here's more from his website:
So why focus the sim on the 60 deg arc if what you're selling isn't that model.
To do the straight line array justice you should make it infinite long as it was intended just as he presented the CBT as intended.
I still don't see the point of the floor standing CBT though. If I was going to use a speaker like that I would hang it from the ceiling. Why would you want the most output be on the floor level?
He's somewhat creating a point source but would you want the acoustic center on the floor?
Attachments
Last edited:
Bravo! Sir John Gielgud it is.
I did find that speech was a little more susceptible to being pulled to one side by shifting >700Hz than was noise. But I simply don't find the range above 700Hz to completely dominate. Its effect is stronger, but not by a huge amount. If forced to point at where the sound is coming from, I'd point to the part >700Hz, but would ask why I hear part of it coming from somewhere else. It's confusing.
I listened to the tracks again today and have to admit they are somewhat laborious and too subtle. Listening on headphones and on speakers isn't the same, either. I'll redo them and post again in a better format.
Ah.. now I understand.
(and yup, #5's not an arc - my bad. )The short answer is: that's not the only CBT.
See pages 114 on..
The smaller ground CBT (the one you are showing in your pic.) was only designed to be compared to the Revel and similar "standard" designs.
-He *starts* with the conceptual no-ground array, and then moves onto the ground/bounded/reflected type at page 69.
It isn't until page 114 that he starts back into the no-ground array again.
Last edited:
I know, I saw the other pictures. But my point is the CBT that is for sale isn't the one in that simulation. The simulation is his best case scenario against a straight line array that's missing something to present it's case at it's best. Seems a bit unfair to me to compare it like that.
Like I said above, can't see the point of the ground plane CBT. I'd rather have the output as presented in the simulation.
Like I said above, can't see the point of the ground plane CBT. I'd rather have the output as presented in the simulation.