Since for loudspeaker measurements we are only interested in direct line-of-sight response between microphone and DUT (speaker), this can in theory also be done with a good cardioid microphone in reverse vs pointed directly at the DUT.
(or a dipole mic at 90 degrees)
And/or can be drastically enhanced even more by using a array to make this beam even much smaller.
Which I think was used/tried before in this topic? (or another topic, can't remember)
Point is that using a omnidirectional microphone makes zero sense to begin with for loudspeaker measurements!
However, the part that is still totally unclear to me, are we talking about a fixed setup or a (semi) mobile setup?
Because in a complete fixed setup we can even build a baby anechoic chamber between microphone and DUT to mitigate any additional reflection.
Why?
To make all the heavy lifting and trickery later much easier!
Resulting in already much better results and therefor not having to deal with so many additional issues to begin with.
Last edited:
This is part of a fundamental disagreement between your position and mine. Earlier you said this is about designing a speaker. I disagree with that too. To me this is about measuring/evaluating/characterising a speaker. The information may be useful in designing but that is not it's only purpose.
I don't want "a good indication" of polar response. I want it to be as accurate as I can make it. If there is a way to avoid an interpolation error I would prefer to have the option to use that method.
It appears that you do not see the value in that because it won't help you design a better speaker. I mostly agree with you on that point.
There is value to me (and others it seems) in being able to measure a speaker or device in as much detail as I can. I am not suggesting that this level of detail is necessary in every application. This thread at it's core is about improving the ability of DIY'ers to measure beyond what is currently available to them without very high cost or significant inconvenience.
To be able to measure as well as can be done. We literally may exceed Klippel resolution if we attack this correctly. At the least we will equal it. And for someone like me, and I think there are other people that make a living from this in this thread, it is of great value.
So if I may summarize in my own words to see if I understand you well?
You're basically building a complicated car that's going to take a lot of time and effort, that's capable of driving 300km/h (190 miles per hour), but that's practically never capable of doing so?
I don't agree with you what the core is of this thread. The way I have been reading this all along, is to mitigate and solve the problems we have with just gated measurements.
Since gated measurements are limited in what they can offer.
Something a Klippel system provides a solution for.
It's only the last couple of pages where this tool all of a sudden became a goal on its own and even beyond that.
It feels very much that you want to measure a voltage with 20 significant digits, while the DUT is fine with just two.
Although it makes zero sense for me just to go for a goal that has no practical sense, I don't have a problem with it an respect it.
But the reason why it's important to talk about it, is because it matters an awful lot to approach the entire project all together.
It's important to make sure everyone is on the same page with similar expectations.
Otherwise you can never solve a complicated puzzle together.
@b_force here is a reminder of what the core of this thread is:
So when all the the posts by others are contributions to modal analysis, sound field separation, hardware or software design, or a novel measurement method, but your posts in response to them is:
I love Gedlee's suggestion on this project:
In other words, primarily to investigating modal analysis and sound field separation, and secondarily to share other novel measurement methods. You and I both know that intent behind what is written can be misread, but as the person who started this, I can say with authority that the primary purpose is developing a tool, and secondarily to mitigate and solve the problems we have with just gated measurements.
So when all the the posts by others are contributions to modal analysis, sound field separation, hardware or software design, or a novel measurement method, but your posts in response to them is:
It feels like you are, not only on a different page as the rest of us, but are actively trying to force the rest of us to the page that you are on. Granted intent is hard to read in text, so that may not be your intent, but you should be aware that it really feels like it is your intent.
I love Gedlee's suggestion on this project:
If you, b_force, find one of the stops of this train is interesting to you, that's great! Please feel free to contribute, you have already made what I feel are excellent suggestions on the topic of lessening room reflections. But if the intended end of its journey seems illogical to you, sounds like nonsense, or whatever it is, is it possible to just accept that and let it be?
This is a good example of misreading.
Because that's not my intention at all.
The intention is to make sure that the goal is what people want.
Reason why asking, is because if I read between the lines, I can very clearly read that not everyone is on the same page.
And that is most definitely not just me.
Secondary is it's important to define these nuances clearly.
You can quote everything you want, but I can assure you that I have already read most of it multiple times.
But there is a huge difference going for perfection for the sake of it, or go for just a usable tool.
I say that without any offense, because both are great.
But the approach and way of thinking is very different.
I explained this is my previous post that I am already fine with that and respect it?
Anyway, at this point I agree with @mbrennwa to move on.
Boundaries and confusion has been cleared up.
Klippelus Maximus? Klippelus not Needucus? Oh no, Monty Python latin is haunting me now. Oh, and Bugs Bunny Latin too.
I don't know if it has dawned on others as it has now dawned on me, but we have two incompatible approaches going on here.
My approach does not make sense if one wants to rotate the mic as the moving mic approach in not optimum for use in a small room, which was the issue that I was stuck with from the beginning. I can explain this further if need-be, but it should be obvious if one just thinks about it.
Weinrich's approach requires a rotating mic and as such is not compatible with my approach until one comes to the very end - a completely automated rotating mic analysis system. Basically there are no train stops along the way if one takes this approach.
One has to decide which approach they will take before getting on the train to that destination.
Are we talking hardware or the software/math behind it?
And to be clear, you're talking about a moving mic vs rotating mic if I understand correctly?
What do you define as a small room, since Klippel's approach (= moving mic) is already for "small" rooms?
I see some possibilities.
Both approaches share a lot: automation (for those who appreciate that), rotation, control, data management, fit spherical harmonics, evaluate those at a denser grid, view results, calculate CAE-2034 plot, ...
The hardware will be quite different, but I see quite some shared building blocks.
Both approaches share a lot: automation (for those who appreciate that), rotation, control, data management, fit spherical harmonics, evaluate those at a denser grid, view results, calculate CAE-2034 plot, ...
The hardware will be quite different, but I see quite some shared building blocks.
Btw, very small angles can be easily done with using microstepping with stepper motors.
Up to 256 microsteps is totally possible which will give a 0.007 degree angle with a standard 1.8 degree stepper motor.
There are 0.9 degree stepper motors available as well.
The downside is that you will loose a bit of torque, but I don't see that there is much torque needed to begin with.
Nema 17 will be plenty here, for moving just a mic probably even smaller.
Although Nema 17 is so extremely common that I think they will be much cheaper.
Up to 256 microsteps is totally possible which will give a 0.007 degree angle with a standard 1.8 degree stepper motor.
There are 0.9 degree stepper motors available as well.
The downside is that you will loose a bit of torque, but I don't see that there is much torque needed to begin with.
Nema 17 will be plenty here, for moving just a mic probably even smaller.
Although Nema 17 is so extremely common that I think they will be much cheaper.
Both
No, I am talking about rotating the mic versus rotating the speaker.
Any room in a home is "small". Weinreich's approach works in a small room only when it is completed, i.e. full 3D. Prior to that point it doesn't offer anything.
Check!
Rotating mic as in totally free, or just as a circle around the speaker? (+ maybe changing the diameter of that circle).
Both methods can do more or less the same (in sense of coordinates), the last one is easier to do mechanically speaking.
If it's fully automated it doesn't really matter in what order the measurements are being done, as long as each measurement gets its own unique reference name (coordinates)
With such very small angles the mounting and fixture of the microphone needs to be very rigid.
Not difficult to do, but definitely not something to forget.
No, if that is what you took from my words I'm not sure how trying to explain it would help.
Very true.
It's already established that the mic must rotate to allow the reflection subtraction algorithm to function correctly.
Very definitely. As soon as you cross 500 hertz there are roughly 4 foot wavelengths already. Even 1kilohertz you have roughly 2 foot wavelengths.
No. The entire premise was to duplicate Klippel. That required a stationary DUT and a moving mic along a controlled arc, and a variable height and depth towards and away from the DUT. Anything else will not allow the same capabilities. The only true differences will be manual for those willing to take a day or two for a full measurement, versus full automation.
That really cuts out the true DIY doesn't it? It certainly places it at a very high level of commitment.
The very vast majority of time of such project is gonna be in software development.
Mostly the user interface, as well as making sure that all measurements sync up correctly with the right angles etc.
I have done and managed many of similar projects (non audio/acoustics related) in the past and that process always takes very long and is extremely tedious.
The difficulty of it lays in just the sheer bulk of it.
(It reminds me a lot of professional coffee machines I used to design the hardware for)
It does make me wonder why Klippel didn't choice for multiple mic measurements at the same time?
Even two would already cut down the measuring time drastically.
At the same time it doesn't really make it much more complicated and on a 100k price tag the additional costs are also negligible.
Speaking of time, for a multiple speaker system it would also be very wise to switch between drivers (sources) at the same coordinate.
Otherwise you have to do it all over again.
Which takes A LOT more time, but also introduces (potential) errors in coordinates.
Not very smart thing to do.
Why? If the collective brain here, and the collective experience put together both a manual and an automatic system, then why does it exclude anyone. Those that have stuck to this thread I believe are capable of making or cobbling together what is required.
If you were determined you could make all of the required measurements with a string to mark a consistent arc and a mic stand with a reasonably long boom arm. Probably what many of us have done for many years. Having a central stand and a movable mic that has predetermined stops is the next step up. If you want motorized and computer controlled then go that route.
From reading your comments I see that you have done similar measurements in the past. And that it has now occurred to you that this is totally different. It is indeed! totally different from what I have been doing for a long time as well. But I see the utility in fleshing out a measurement system that will allow me to do better work with less hassle. I have a country setting, a large yard. Tall ladders, room for groundplane measurements. But I also have winter! And rain. Here winter is nearly 5 months of the year. So this measurement idea is important to me.
If anyone else finds it valueable then they will be here. If I can lend a hand on the mechanical front, Just ask. I keep trying to think of the easiest methods to make this stuff. By the end of next week I'll have my son's kitchen installed and be able to devote shop time to this endeavour. Pics and videos as well.