Hey David!!
Love your program.
I've been also been playing with AKAKBAK a little off and on over the past few years.
They seem to relate to each other well once AKABAKS quirks are figured out.
There is one thing that I find all speaker designing tools and real world design approaches lacking.
I don't know if its trivial to simulate along with everything else, but If I had something to add to my speaker design wish list is, to predict acoustic and impedance decay over time.
You can get a sense for where acoustic resonances are by looking at the impedance charts, but we don't get a sense of how well or poorly damped those resonances can be, and other periodic weirdness.
I've always wondered what a speaker with minimal decay would sound like or at least "critically damped".
David do you make your source code available?
Love your program.
I've been also been playing with AKAKBAK a little off and on over the past few years.
They seem to relate to each other well once AKABAKS quirks are figured out.
There is one thing that I find all speaker designing tools and real world design approaches lacking.
I don't know if its trivial to simulate along with everything else, but If I had something to add to my speaker design wish list is, to predict acoustic and impedance decay over time.
You can get a sense for where acoustic resonances are by looking at the impedance charts, but we don't get a sense of how well or poorly damped those resonances can be, and other periodic weirdness.
I've always wondered what a speaker with minimal decay would sound like or at least "critically damped".
David do you make your source code available?
If you can get the simulation impulse response, then probably it can be processed like any measured impulse response.
Yes impulse response is good for determining quite a few things.
I don't know if I've ever seen a simulated impulse response.
I've measured a few with a TEF.
I've played with some of the MLS stuff as well, gives some nice spectral decay graphs.
A lot of manufacturers don't measure spectral decay either. I remember some of the old Data Sheets for the original SPLTD1's had some spectral decay grahps. That was back when Servodrive/SPL had a really simple web site. Not that there is much of a SPL Servodrive site anymore...
I don't know if I've ever seen a simulated impulse response.
I've measured a few with a TEF.
I've played with some of the MLS stuff as well, gives some nice spectral decay graphs.
A lot of manufacturers don't measure spectral decay either. I remember some of the old Data Sheets for the original SPLTD1's had some spectral decay grahps. That was back when Servodrive/SPL had a really simple web site. Not that there is much of a SPL Servodrive site anymore...
Hi sumsound,
Not even to my mother 🙂.
In Hornresp, select the Tools > Impulse Response menu command from the SPL Response window to calculate and display the impulse response. The impulse data can then also be exported as a *.wav file if required.
Kind regards,
David
Not even to my mother 🙂.
In Hornresp, select the Tools > Impulse Response menu command from the SPL Response window to calculate and display the impulse response. The impulse data can then also be exported as a *.wav file if required.
Kind regards,
David
Last edited:
Awesome! Thanks for the slider updates.
sumsound: you could export the record to AkAbak and then do an iFFT on both cone movement and impedance.
Not sure how useful it would be. Looks pretty though.
Built in IR for SPL in hornresp is easier to use than AkAbak.
sumsound: you could export the record to AkAbak and then do an iFFT on both cone movement and impedance.
Not sure how useful it would be. Looks pretty though.
Built in IR for SPL in hornresp is easier to use than AkAbak.
Hello,
If there is some interest for that, I can eventually adapt my quasi-wavelets Matlab routine to be used in Hornresp.
You can see an example of quasi wavelets graph at
http://www.diyaudio.com/forums/attachments/multi-way/167266d1271433163-horn-honk-wanted-zilch_a.gif
from my post:
http://www.diyaudio.com/forums/multi-way/161627-horn-honk-wanted-66.html#post2155843
Best regards from Paris, France
Jean-Michel Le Cléac'h
If there is some interest for that, I can eventually adapt my quasi-wavelets Matlab routine to be used in Hornresp.
You can see an example of quasi wavelets graph at
http://www.diyaudio.com/forums/attachments/multi-way/167266d1271433163-horn-honk-wanted-zilch_a.gif
from my post:
http://www.diyaudio.com/forums/multi-way/161627-horn-honk-wanted-66.html#post2155843
Best regards from Paris, France
Jean-Michel Le Cléac'h
Interest? Of course there is interest!
Although if an external wavelet app could be made to analyze any IR, that would probably make more sense. Could be launched from inside hornresp.
Although if an external wavelet app could be made to analyze any IR, that would probably make more sense. Could be launched from inside hornresp.
Its interesting listening to some of those impulse waves.
I listened to the impulse from a Lab12v2 it seemed very deep with a bit of a snap to it.
Listening to the impulse of the model of a TH-mini sounds a bit like a bongo...
I listened to the impulse from a Lab12v2 it seemed very deep with a bit of a snap to it.
Listening to the impulse of the model of a TH-mini sounds a bit like a bongo...
One way to find a good driver, is to use the "System Design from specifications" tool. You are asked to specify a small number of parameters, among them the rear chamber volume, and then Hornresp calculates both the required driver parameters and the horn parameters. You can then search the data sheets from various driver manufacturers to find the driver that most closely matches the one Hornresp has calculated. When you have found one that looks good, enter its parameters in Hornresp, and check if the design still works well.
Good luck!
Bjørn
Hornresp Version 26.40
Hi Everyone,
Hornresp Version 26.40 has just been released. Changes are:
* The Loudspeaker Wizard S1 to S6 sliders for other than tapped horns, and the Vtc slider for all horns, now have value-dependent multipliers. See my earlier post #1079 for details. My thanks to David_Web for suggesting the enhancement.
* Ap1 and Lpt can now be used with conventional and compound horns to specify a conical adaptor between the throat chamber and the horn throat. The adaptor can be considered as part of a compression driver, or alternatively, as an additional conical segment to form a 5 segment horn.
Ap1 = conical adaptor entry or 'throat'
Lpt = conical adaptor length
S1 = conical adaptor 'mouth'
Note that the Horn Segment Wizard works with the adaptor, and that Ap1 and Lpt can still be used to specify a throat chamber port in an offset driver or tapped horn.
* A bug which under some circumstances caused the Loudspeaker Wizard combined response baseline to be calculated incorrectly, has been fixed.
Kind regards,
David
Hi Everyone,
Hornresp Version 26.40 has just been released. Changes are:
* The Loudspeaker Wizard S1 to S6 sliders for other than tapped horns, and the Vtc slider for all horns, now have value-dependent multipliers. See my earlier post #1079 for details. My thanks to David_Web for suggesting the enhancement.
* Ap1 and Lpt can now be used with conventional and compound horns to specify a conical adaptor between the throat chamber and the horn throat. The adaptor can be considered as part of a compression driver, or alternatively, as an additional conical segment to form a 5 segment horn.
Ap1 = conical adaptor entry or 'throat'
Lpt = conical adaptor length
S1 = conical adaptor 'mouth'
Note that the Horn Segment Wizard works with the adaptor, and that Ap1 and Lpt can still be used to specify a throat chamber port in an offset driver or tapped horn.
* A bug which under some circumstances caused the Loudspeaker Wizard combined response baseline to be calculated incorrectly, has been fixed.
Kind regards,
David
Ahh it keeps getting better and better.
David McBean FTW!
Is it possible to tell it to use combined response before Calculations?
Make it easier to compare.
Also still down for "memory" slots to be able to compare multiple designs at once.
David McBean FTW!
Is it possible to tell it to use combined response before Calculations?
Make it easier to compare.
Also still down for "memory" slots to be able to compare multiple designs at once.
Hi FlipC,
The way that Hornresp has evolved, unfortunately the work now required to make such a change would be considerable. It is unlikely to happen, as from my perspective the benefit to be gained is not particularly great. There would also be the added complication of having to specify a path length difference before calculating any results at all.
As far as comparing combined responses is concerned, one set of results can be captured by pressing Ctrl+C and then compared to a second set by pressing F4. It works for me 🙂.
Sorry, but the memory slots now provided with the Loudspeaker Wizard are about as far as I am prepared to go 🙂.
While there are certainly lots more improvements that could be made to Hornresp, the program has now reached the point where the addition of any 'major' new features would probably require a complete re-write of the code to implement properly. It is not something that I am in a position to undertake. Life is short, and I am not getting any younger... 🙂.
Kind regards,
David
The way that Hornresp has evolved, unfortunately the work now required to make such a change would be considerable. It is unlikely to happen, as from my perspective the benefit to be gained is not particularly great. There would also be the added complication of having to specify a path length difference before calculating any results at all.
As far as comparing combined responses is concerned, one set of results can be captured by pressing Ctrl+C and then compared to a second set by pressing F4. It works for me 🙂.
Sorry, but the memory slots now provided with the Loudspeaker Wizard are about as far as I am prepared to go 🙂.
While there are certainly lots more improvements that could be made to Hornresp, the program has now reached the point where the addition of any 'major' new features would probably require a complete re-write of the code to implement properly. It is not something that I am in a position to undertake. Life is short, and I am not getting any younger... 🙂.
Kind regards,
David
Hi FlipC,
One other thing that I forgot to mention - if the Loudspeaker Wizard is used, then either the Default, Output 2 or Combined option can be set for all SPL response calculations. It is not necessary to display the default response before the combined response can be calculated.
Kind regards,
David
One other thing that I forgot to mention - if the Loudspeaker Wizard is used, then either the Default, Output 2 or Combined option can be set for all SPL response calculations. It is not necessary to display the default response before the combined response can be calculated.
Kind regards,
David
Hey David, BFM gave you major props on his site!
BillFitzmaurice.info - View topic - These guys seem to have an 18" loaded horn that works...
Ron K wrote: according to Tom Danley, Akabak models quite differently then Horn Response.
Bill Fitzmaurice wrote: It used to, but not anymore. Where response is concerned HornResp is now just about as accurate as Akabak, and is infinitely easier to use. Akabak offers some more advanced features, but McBean issues improved HornResp versions on average every six months, and I expect he'll eventually meet or exceed Akabak."
BillFitzmaurice.info - View topic - These guys seem to have an 18" loaded horn that works...
Ron K wrote: according to Tom Danley, Akabak models quite differently then Horn Response.
Bill Fitzmaurice wrote: It used to, but not anymore. Where response is concerned HornResp is now just about as accurate as Akabak, and is infinitely easier to use. Akabak offers some more advanced features, but McBean issues improved HornResp versions on average every six months, and I expect he'll eventually meet or exceed Akabak."
Depends on how you use it.
I was able to get my results almost identical using akabak, before I realized that hornresp exports directly to akabak format!!
Hi BP1Fanatic,
Thanks for making me aware of Bill's comments regarding Hornresp. I think it would be fair to say that he started off as being somewhat sceptical of classical horn theory early on, preferring to use his own intuition and experience to develop new "ground-breaking" designs that seemingly defied conventional wisdom and theory. It appears though, that he has perhaps been "converted" somewhere along the line 🙂.
Hornresp and AkAbak use the same fundamental lumped-element loudspeaker model, as documented in the classic texts by Olson and Beranek - which means that results produced by the two basic models will be identical.
Differences do however arise in some situations because Hornresp uses either a plane or isophase wavefront horn model, depending upon the circumstances. Akabak uses a plane wavefront model in all cases, meaning that results can become less accurate once Cir exceeds 1. Also, Hornresp uses a more sophisticated directivity model than AkAbak, meaning that pressure response results will be different.
Power response results for simple horns in 2 Pi space having a value of Cir < 1 will be identical for the two programs, provided that the input parameter values are the same, and that the Hornresp results are calculated with resonances not masked.
AkAbak can simulate conical, exponential, and hyperbolic-exponential horns, whereas Hornresp can also simulate Le Cléac'h horns, oblate spheroidal waveguides, spherical wave horns and tractrix horns. Comparison with actual measured results has shown the Hornresp predictions to be reasonably accurate - they are certainly very useful for establishing an optimum design for a given loudspeaker configuration.
However, Hornresp will never be able to offer the absolute design specification flexibility that AkAbak has...
Kind regards,
David
Thanks for making me aware of Bill's comments regarding Hornresp. I think it would be fair to say that he started off as being somewhat sceptical of classical horn theory early on, preferring to use his own intuition and experience to develop new "ground-breaking" designs that seemingly defied conventional wisdom and theory. It appears though, that he has perhaps been "converted" somewhere along the line 🙂.
Hornresp and AkAbak use the same fundamental lumped-element loudspeaker model, as documented in the classic texts by Olson and Beranek - which means that results produced by the two basic models will be identical.
Differences do however arise in some situations because Hornresp uses either a plane or isophase wavefront horn model, depending upon the circumstances. Akabak uses a plane wavefront model in all cases, meaning that results can become less accurate once Cir exceeds 1. Also, Hornresp uses a more sophisticated directivity model than AkAbak, meaning that pressure response results will be different.
Power response results for simple horns in 2 Pi space having a value of Cir < 1 will be identical for the two programs, provided that the input parameter values are the same, and that the Hornresp results are calculated with resonances not masked.
AkAbak can simulate conical, exponential, and hyperbolic-exponential horns, whereas Hornresp can also simulate Le Cléac'h horns, oblate spheroidal waveguides, spherical wave horns and tractrix horns. Comparison with actual measured results has shown the Hornresp predictions to be reasonably accurate - they are certainly very useful for establishing an optimum design for a given loudspeaker configuration.
However, Hornresp will never be able to offer the absolute design specification flexibility that AkAbak has...
Kind regards,
David
Last edited:
No problem guy! You deserve it seeing that most people on the different audio forums use your amazing product
