JohnK - very impressive- thanks! Never quite seen it put that way. =)
How did you figure the EQ? The low end looks simple, to top end does not.
Maybe Lynn is looking for a way to do that acousticaly, instead of electricaly?
Having heard Gary's system, I can certainly say it's one of the cleanest I've ever heard in over 30 years. Quite amazing. No exotic/expensive drivers, either.
How did you figure the EQ? The low end looks simple, to top end does not.
Maybe Lynn is looking for a way to do that acousticaly, instead of electricaly?
Having heard Gary's system, I can certainly say it's one of the cleanest I've ever heard in over 30 years. Quite amazing. No exotic/expensive drivers, either.
Lynn, do you have any reliable data on that?
My experiments and own measurments - a pretty long time ago - didnt show any significant improvement for that kind of "inverse horn" shape.
Also I haven't heard of a Nautilus DIY fan community comparable to the Carlson coupler idea for example (but also haven't searched for to be honest) which for sure would exist if there is *anything* of interest on this principle (besides the cool shape).
Michael
Telstar reply.
Hi Telstar,
Thanks I have sent you the photos, enjoy!
The Acrylic cabinets were my first experiments and used good but not great bass / low mid drivers, two per side of the 8inch Seas Excel W22EX 001 run in parallel to give 4 Ohms and 91dB sensitivity.
I now use 10 or 12 inch paper coned drivers with 3 or 4 inch voice coils and 96dB plus sensitivity.
The in room -3dB was a genuine 30Hz to 33Hz using some EQ cut (not boost!) as I used the power amp level settings to boost the whole 100Hz to 350Hz band these drivers were covering and then cut down by approx 7dB a peak from about 150Hz to 250Hz if my memory is correct.
The principal being acieve desired in room response with the minimum of EQ intervention. ie rather than boost the low end, cut the top end by starting from an elevated power amp level setting.
Martin Collums (UK speaker guru and reviewer) loved them!
Cheers
Derek.
PS I will post a reply to your twin Manger question on the Manger thread you suggested as I dont want to go off topic here.
Hi Telstar,
Thanks I have sent you the photos, enjoy!
The Acrylic cabinets were my first experiments and used good but not great bass / low mid drivers, two per side of the 8inch Seas Excel W22EX 001 run in parallel to give 4 Ohms and 91dB sensitivity.
I now use 10 or 12 inch paper coned drivers with 3 or 4 inch voice coils and 96dB plus sensitivity.
The in room -3dB was a genuine 30Hz to 33Hz using some EQ cut (not boost!) as I used the power amp level settings to boost the whole 100Hz to 350Hz band these drivers were covering and then cut down by approx 7dB a peak from about 150Hz to 250Hz if my memory is correct.
The principal being acieve desired in room response with the minimum of EQ intervention. ie rather than boost the low end, cut the top end by starting from an elevated power amp level setting.
Martin Collums (UK speaker guru and reviewer) loved them!
Cheers
Derek.
PS I will post a reply to your twin Manger question on the Manger thread you suggested as I dont want to go off topic here.
john k... said:
John
I don't think that I disagree with you as much as I'd say that MP applied to a three dimensional problem is not meaningful. Here is why. Lets say that the axial response is MP and that we correct it to be flat, whatever that takes. And lets say that the off axis responses are also MP and correctable to flat. These are both points that you made that I don't disagree with. But the problem is that these two corrections are not the same. There is no single correction that can make BOTH points flat at the same time even if both points are MP individually. So in a three dimensional problem even if any given point is MP they are not MP with respect to each other. I think that this is more a failure of the concept of MP application to 3-D fields than anything else.
Sheldon DEQX question
Hi Sheldon,
Not with the DEQX, for two reasons.
(1) With the DEQX the lower in frequency any signal manipulation is implemented the more demanding it is on the DEQX software.
For example the latest HD version running at 96KHz can not cope with a 80Hz crossover point using Linear Phase filters, you have to drop to 48Khz where there is no problem.
(2) The main IO manager page where yo set the basic levels between drivers is a "blanket" DSP ie there are no more or less calculations required to set the low mid driver at +5dB than there is to set +1dB or -2dB etc.
The software has better resolution at slightly higher frequencies due to the programe priorities of Latency Vs resolution.
The DEQX is far from perfect, but most of its compromises can be worked around.
Basically you can run out of computing power if you set it up the wrong way or.... you can get more out of it by setting it up correctly. Have I just said the same thing in two different ways....!!?
Hope this helps
Cheers
Derek.
Hi Sheldon,
Not with the DEQX, for two reasons.
(1) With the DEQX the lower in frequency any signal manipulation is implemented the more demanding it is on the DEQX software.
For example the latest HD version running at 96KHz can not cope with a 80Hz crossover point using Linear Phase filters, you have to drop to 48Khz where there is no problem.
(2) The main IO manager page where yo set the basic levels between drivers is a "blanket" DSP ie there are no more or less calculations required to set the low mid driver at +5dB than there is to set +1dB or -2dB etc.
The software has better resolution at slightly higher frequencies due to the programe priorities of Latency Vs resolution.
The DEQX is far from perfect, but most of its compromises can be worked around.
Basically you can run out of computing power if you set it up the wrong way or.... you can get more out of it by setting it up correctly. Have I just said the same thing in two different ways....!!?
Hope this helps
Cheers
Derek.
Hi Earl,
I don't disagree that the MP corrections would be different under certain conditions. I just don't consider that the issue because it is more of a limitation of the design being directional. Take the post I just made above where I looked at the dipole impulse response both on and 30 degrees off axis. Both cases use the same EQ function and the impulse off axis is just about as good as on axis. It would actually be better if I lowered the crossover point. The reason it is as good as it is, and why if would improve with a lower x-o point, is because the dipole response for the bandpass I chose is almost CD. The lower x-o point would yield even better CD. The magnitude changes off axis for the dipole by a scale factor, but the shape of the response (the transfer function), which is what matters for the MP, does not. The point being that if the response is truly CD then only a single eq is required for the entire 3-d space. You of all people should appreciate that. If the system response is MP (i.e. and transient perfect x-o is used) then a single Eq function will correct the response in both amplitude and time over the entire CD window. Sort of another reason why CD should be a design goal.
I don't disagree that the MP corrections would be different under certain conditions. I just don't consider that the issue because it is more of a limitation of the design being directional. Take the post I just made above where I looked at the dipole impulse response both on and 30 degrees off axis. Both cases use the same EQ function and the impulse off axis is just about as good as on axis. It would actually be better if I lowered the crossover point. The reason it is as good as it is, and why if would improve with a lower x-o point, is because the dipole response for the bandpass I chose is almost CD. The lower x-o point would yield even better CD. The magnitude changes off axis for the dipole by a scale factor, but the shape of the response (the transfer function), which is what matters for the MP, does not. The point being that if the response is truly CD then only a single eq is required for the entire 3-d space. You of all people should appreciate that. If the system response is MP (i.e. and transient perfect x-o is used) then a single Eq function will correct the response in both amplitude and time over the entire CD window. Sort of another reason why CD should be a design goal.
panomaniac said:
The eq is designed using SoundEasy's Digital Equalizer. It allows the user to specify a target acoustic response and then given the driver's raw amplitude and phase response it computes the required filter transfer function to meet the target.
The measurements I presented are acoustic measurement. They were obtained by "playing" the SE developed digital filter (also part of SE's capability).
I uses this DE function in all my preliminary designs. Then I optimize passive or active analog circuits to emulated the digital filters as closely as possible. It's a very powerful tool.
I am sure that Gary's woofer system sounds great. So does my NaO woofer system. From Lynn's description they are pretty much the same design concept, originally due to Kalusche, 1950 and then further analyzed by Holmens (1986) and by Backman (1999) as I stated below.
By the way, Lynn's comment, "They're almost acting as if they were very short transmission lines..." In fact, that is exactly what they are, short,well damped TL's. The role of the damping is two fold: 1) damp the TL resonances and 2) act as an acoustic low pass filter.
mige0 said:
Vented and TL's are certainly MP. Horns, well I'm not an expert on horns by any means. But from what little I have read the phase velocity can be frequency dependent in a horn which would imply departure from MP at the mouth if the response at the throat is MP. How significant this is, I don't know. Perhaps Earl can address this.
Re: Sheldon DEQX question
I don't want to stray too far off the track in this thread, so I'll end with my comments here (happy to carry on in a new thread).
I have a DEQX myself and have spent the last couple of years playing from time to time. I'd say I'm half way to knowing it well. Yes, I agree that you want to set up your amps so that you are not using the dynamic range of the QX to correct for overall level from amp/driver to driver. In other words, the the relative sensitivity at the amp levels should be set up so that the amplitude range is minimized across all drivers when the output is measured. If that's all you are saying, then I agree. The only thing that matters is the total range of correction from highest amplitude to lowest that you want to flatten.
In my system, which include a CD horn, I use an analog filter in the amp to compensate the 6dB roll off about about 5kHz. I also use an analog filter to boost the low end roll off. And I adjust amp sensitivity so that the overall amplitude window is minimized. That way, my total correction span is about 6dB for the QX, and I can preserve the dynamic range for resolving the signal.
BTW, I also set it up so that no amp is allowed to clip at full volume on the QX.
Sheldon
Overkill Audio said:
I don't want to stray too far off the track in this thread, so I'll end with my comments here (happy to carry on in a new thread).
I have a DEQX myself and have spent the last couple of years playing from time to time. I'd say I'm half way to knowing it well. Yes, I agree that you want to set up your amps so that you are not using the dynamic range of the QX to correct for overall level from amp/driver to driver. In other words, the the relative sensitivity at the amp levels should be set up so that the amplitude range is minimized across all drivers when the output is measured. If that's all you are saying, then I agree. The only thing that matters is the total range of correction from highest amplitude to lowest that you want to flatten.
In my system, which include a CD horn, I use an analog filter in the amp to compensate the 6dB roll off about about 5kHz. I also use an analog filter to boost the low end roll off. And I adjust amp sensitivity so that the overall amplitude window is minimized. That way, my total correction span is about 6dB for the QX, and I can preserve the dynamic range for resolving the signal.
BTW, I also set it up so that no amp is allowed to clip at full volume on the QX.
Sheldon
The fun in audio systems is that once you really get significant improvements, the little flaws become obvious, and sometimes annoying. The good thing is that it's easier to pinpoint and vocus improvement on.panomaniac said:
I'll send you some. Pretty, it is not! But Gary is rebuilding for better looks.
Funny, I think the woofer is the weakest part of the system, but the rest is so good that tiny flaws stick out.
With the rebuild and new stuffing scheme, will be interesting to hear.
Just a day or two ago, a manufacturer was telling me they had not been able to find qualified power cables for an active speaker, and were wondering whether it made a difference if they used a different one at a lower rating. We plugged them in and just swapped them around, and they convinced themselves it made a difference. I didn't even have to argue the issue.
john k... said:
Thanks John!
To make it even more clear – we could completely substitute any dampening measures for closed, vented, TL enclosures by more sophisticated EQing – like provided by SoundEasy, Accourate, AudioLense, DEQX or even a complex analogue correction network !?!
I should have started earlier with playing around with PC XO's I guess..
Michael
gedlee said:
john k... said:Hi Earl,
I don't disagree that the MP corrections would be different under certain conditions. I just don't consider that the issue because it is more of a limitation of the design being directional. Take the post I just made above where I looked at the dipole impulse response both on and 30 degrees off axis. Both cases use the same EQ function and the impulse off axis is just about as good as on axis. It would actually be better if I lowered the crossover point. The reason it is as good as it is, and why if would improve with a lower x-o point, is because the dipole response for the bandpass I chose is almost CD. The lower x-o point would yield even better CD. The magnitude changes off axis for the dipole by a scale factor, but the shape of the response (the transfer function), which is what matters for the MP, does not. The point being that if the response is truly CD then only a single eq is required for the entire 3-d space. You of all people should appreciate that. If the system response is MP (i.e. and transient perfect x-o is used) then a single Eq function will correct the response in both amplitude and time over the entire CD window. Sort of another reason why CD should be a design goal.
A perfect on axis correction made for a min phase system that's *not* constant directivity would show frequency response irregularities when measured off axis – OK.
For the off axis measurements - would we see decay on CSD or not, as the underlying effect is basically "only" comb filter effects?
Michael
panomaniac said:
Thanks for your interest – not much more to say about than I already did.
It's not been a "scientific" research by any means, just one of my investigations in how much I'd been able to minimise resonance's *without* dampening – by trying different enclosure approaches that claim to do so.
No result showing up that's been worth to spend more time on.
Michael
soongsc said:
3 meters roughly.
Michael
Re: Re: Link for battery supply
Got the pictures last night from Derek.
Overkill Angel pictures (reduced a bit):
Side view:
Angels full setup at the London 2006 show:
Prey and Encore in another post.
Telstar said:
Got the pictures last night from Derek.
Overkill Angel pictures (reduced a bit):
An externally hosted image should be here but it was not working when we last tested it.
Side view:
An externally hosted image should be here but it was not working when we last tested it.
Angels full setup at the London 2006 show:
An externally hosted image should be here but it was not working when we last tested it.
Prey and Encore in another post.