Understanding How the Klippel NFS Works

[Few]

For those trying to wrap their head around the mathematics, and who took some chemistry or physics, you might find it helpful to envision the shapes of the atomic orbitals. And apologies if it turns out referring to atomic orbitals is utterly useless to all readers!

The angular features of the orbitals (not the features that depend on distance from the nucleus, but the ones that depend on direction) are precisely the same spherical harmonics used in this application. The s orbitals are spherical, the p orbitals have one planar node (there are three versions, with the nodes lying in mutually perpendicular planes), and so on. Wherever an orbital’s lobe points, the speaker has a strong response. Wherever the orbital has a node, like the pinched in waist of the p orbitals, the speaker has reduced output. The idea is that any acoustic radiation pattern can be described by summing weighted contributions from a perfectly spherical response (s orbital), a response with a single planar node (p orbitals), two planar nodes (d orbitals), and so on.

This site provides a few orbital visuals:

Again, if I’m the only person interested in audio and orbitals, then I apologize for the distraction.

Few

 

[Few]

As I wrestle with measurements of a new multi-entrant horn creation, in a too-small, reflective, and crowded environment, I find I’m eager for a rigorous process to get around these issues (without selling my soul to Klippel). So I’m tickling this thread to ask whether things are moving forward or, heaven forfend, something is inhibiting progress. I’ve had a few thoughts about how someone could minimize the hardware hassles, but I don’t want to get ahead of the computational component.
Few

 

[amirm]

As far as I know, there is no active effort to build anything. At least none that I know of.

 

[NTK]

I am also not aware of anybody pursuing it.

I recently came across another quasi-anechoic method described in these 2 related papers. It is much less satisfactory (IMO) than the NFS, but requires much less fancy hardware. As far as I understand it, the method works only for rectangular rooms.

Here is a very brief description on how it works as I understand it. You mount your measurement mic on a linear track (figs 1 and 2). As you move the mic, the change in the distance between speaker-under-test and the mic will be different from those from from the first reflection image sources, except the front wall reflection (fig 3). Therefore, with enough measurements, we should be able to separate out the direct sound from the primary reflections (except those from the front wall). I have not dug into the math to understand how exactly it is done.

Compare to the NFS, there are numerous significant disadvantages I can think of:

• It only addresses primary reflections, and cannot address the primary front wall reflection.
• Seems to only work well for rectangular rooms.
• The reflections, as you move the mic, will come from different angles from the speaker-under-test.
• Seems to require a lot of measurements per point (orientation), and a full spinorama requires 70 points. The test in the paper used 100 measurements for 1 point.

https://www.researchgate.net/public...ment_of_Loudspeakers_Using_Beamforming_Method
https://www.researchgate.net/public...oudspeakers_Using_Adaptive_Beamforming_Method

 

[carewser]

As far as I can tell the Klippel NFS eliminates room coloration thus eliminating the need for an anechoic chamber to measure speakers and as someone who owns 35 speakers i'd be interested in buying one but there are two problems, I can't find any way to purchase an NFS and even if I could I expect it would cost way more than i'd be willing to spend. Can one even be purchased?

 

[amirm]

Sure. They have representations around the world. Warkwyn is their US rep. They may handle Canada as well. They are super nice people so ask them.

Klippel is also one of the few companies that publishes its price lists. Figuring out all the options you need is hard though. Figure about $100,000 US and you will be in the ballpark. Hopefully Canada does not impose import duty on goods coming from Germany. Otherwise you have to add that fee as well. I also had to hire an import company to clear customs for me. This is a big machine by the way and requires a lot of space.

 

[ctrl]

I recently came across another quasi-anechoic method described in these 2 related papers. It is much less satisfactory (IMO) than the NFS, but requires much less fancy hardware. As far as I understand it, the method works only for rectangular rooms.

Thanks for the links!
This is exactly the case I was interested described in post#6.

What bothers me a bit about the paper is the fact that in the combined near-field far-field measurement method, no baffle-step correction was performed before merging near-field and far-field measurement, deliberately distorting the result - but then the calculated quasi-anechoic frequency response would not look so impressive anymore

The 100 measurements required for one full frequency response are not exactly few. I did not find any information in the text whether this procedure would also work with floorstanding loudspeakers on a low pedestal.

 

[carewser]

Sure. They have representations around the world. Warkwyn is their US rep. They may handle Canada as well. They are super nice people so ask them.

Klippel is also one of the few companies that publishes its price lists. Figuring out all the options you need is hard though. Figure about $100,000 US and you will be in the ballpark. Hopefully Canada does not impose import duty on goods coming from Germany. Otherwise you have to add that fee as well. I also had to hire an import company to clear customs for me. This is a big machine by the way and requires a lot of space.

I don't understand how because I not only checked out the Klippel website but also the Canadian distributor's website and couldn't find any prices or any way to order one but ~$100,000 is ~$99,000 more than I was hoping so it doesn't really matter

 

[NTK]

I don't understand how because I not only checked out the Klippel website but also the Canadian distributor's website and couldn't find any prices or any way to order one but ~$100,000 is ~$99,000 more than I was hoping so it doesn't really matter

You can find the Klippel price list here in case you change your mind
https://www.klippel.de/products/rd-system/product-brochure-price-lists.html

 

[carewser]

You can find the Klippel price list here in case you change your mind
https://www.klippel.de/products/rd-system/product-brochure-price-lists.html

Thanks for the link but i expect it would be much cheaper to just turn a vacant room into an anechoic chamber than buy one of these things

 

[pozz]

Thanks for the link but i expect it would be much cheaper to just turn a vacant room into an anechoic chamber than buy one of these things

It is way, way cheaper to buy a Klippel. To have anywhere accurate bass performace requires a full and costly rebuild, and even certified ones have limitations below 100Hz. Apart from controlling reverb you have to make sure that outside sound doesn't leak in. Plus the investment in measurement equipment and the complexity of setting it up.

 

[MarsianC#]

Try to move a room - a really large ad super heavy one. No problem with NFS, done in a day.

 

[NTK]

You can take a look at Professor Edgar Choueiri's (originator of the BACCH system) 3D3A lab at Princeton. Its '"reconfigurable" anechoic chamber is only good above 425 Hz with 8 inch wedges. From the pictures it looks like it is a bit more than a kilo-bucks investment.

https://www.princeton.edu/3D3A/Facilities.html
https://www.princeton.edu/3D3A/PhotoAlbum/3D3Aalbum1/album/index.html

 

[bennybbbx]

Is it possible to get from old klippel measure data the step response easy as in REW ?

 

[NTK]

Is it possible to get from old klippel measure data the step response easy as in REW ?

The optional Complex Data Export add-on module, which costs 6800 EUR, gives impulse response. It is should be easy enough to process the impulse response into step response (using convolution). I don't know if Amir has purchased this module (but suspect not).

 

[carewser]

It is way, way cheaper to buy a Klippel. To have anywhere accurate bass performace requires a full and costly rebuild, and even certified ones have limitations below 100Hz. Apart from controlling reverb you have to make sure that outside sound doesn't leak in. Plus the investment in measurement equipment and the complexity of setting it up.

Yeah I guess to make it soundproof would be a little costlier than just throwing up empty egg cartons on the walls and ceiling

 

[MrJoePike]

Seems a reasonable DIY
https://www.cnn.com/style/article/a...s/180326172314-microsoft-anechoic-chamber.jpg

 

[carewser]

"To achieve extreme silence, the room is designed with an onion-like structure that isolates it from the rest of the building and the outside world. It's made of six layers of concrete and steel and it is somewhat disconnected from the surrounding building, because it sits atop an array of vibration damping springs. Inside, fiberglass wedges are mounted on the floor, ceiling and walls to break up sound waves before they have a chance to bounce back into the room. The floor itself is simply a grid of sound-absorbing suspended cables."

I've already started work on mine

 

[bennybbbx]

The optional Complex Data Export add-on module, which costs 6800 EUR, gives impulse response. It is should be easy enough to process the impulse response into step response (using convolution). I don't know if Amir has purchased this module (but suspect not).

View attachment 102223

today i search for yamaha HS step response to see if there can find some for yamaha HS. here is a step response from this site.seem Amir have the module only resolution is not so good. i think the important stuff is only in first 10 ms see. so a resolution 1 line 1 ms is good. question is wy newer tests have no step response. maybe @amirm can tell something to this. in this test can find step response https://www.audiosciencereview.com/forum/index.php?threads/yamaha-hs5-powered-monitor-review.10967/

 

[BDWoody]

"To achieve extreme silence, the room is designed with an onion-like structure that isolates it from the rest of the building and the outside world. It's made of six layers of concrete and steel and it is somewhat disconnected from the surrounding building, because it sits atop an array of vibration damping springs. Inside, fiberglass wedges are mounted on the floor, ceiling and walls to break up sound waves before they have a chance to bounce back into the room. The floor itself is simply a grid of sound-absorbing suspended cables."

I've already started work on mine

Suddenly that $100k for the Klippel makes more sense.

-20dB
That's silly.