Dynaudio creates new “Jupiter Room” for measuring loudspeakers. (see video)

[TemploAztlan]

It seems like it would be better for measuring towers, with several drivers stacked up, than a Klippel. Because with the Klippel, you are measuring at one height, usually between tweeter and midrange. Whereas on the Jupiter, since the mics are farther away from the speaker, it would better capture the sum of all the drivers.

 

[NTK]

And for those with a NFS I wonder if there is an option to perform a lower spatial resolution and angle limited measurement set that is achievable in minutes in order to speed up the iterative design process?

You can take a look at this Klippel paper.
https://www.klippel.de/fileadmin/klippel/Bilder/Know-How/Literature/Papers/Fast Loudspeaker Measurement in Non-Anechoic Environment.pdf

As for how big a problem the NFS measurement time is, the question is how many different design prototype designs the designers can come up with and the shop can produce a day. For the price of an anechoic chamber, you can buy several NFS systems and run them in parallel.

 

[Salt]

the entire radiation characteristics of loudspeakers can be simulated in software without having to build anything..

If, and that is the problem, the transducer's data are measured correctly ...

 

[MAB]

It seems like it would be better for measuring towers, with several drivers stacked up, than a Klippel. Because with the Klippel, you are measuring at one height, usually between tweeter and midrange. Whereas on the Jupiter, since the mics are farther away from the speaker, it would better capture the sum of all the drivers.

Klippel, and near field scanners in general, measure a 3D space around the speaker, not at one height.

Near Field Scanner System (NFS)

Near Field Scanning System, NFS

www.klippel.de

 

[Audio Science Pal]

Awesome room and loudspeaker measurement device!

 

[Pearljam5000]

Why isn't Dynaudio more popular on ASR?

 

[Salt]

Why isn't Dynaudio more popular on ASR?

Just send in some samples ... Only 15% today to pay on top ...

 

[MoreWatts]

Why isn't Dynaudio more popular on ASR?

ASR has measured 4 Dynaudios. 2 passive and 2 active, small, 2-ways. 2 weren't recommended, 2 were maybe okay, with caveats. None were super-pricey. I don't know which models were designed pre- or post- 'Jupiter Room.'

 

[Devnull]

On the contrary. Full 360° horizontal and vertical measurements of the seperate drivers in the assembled cabinet and the rest can be done from behind a computer before assembling even a basic crossover. Also, especially for companies with qualified people on board, pretty much the entire radiation characteristics of loudspeakers can be simulated in software without having to build anything..

Pretty much anything can be simulated or modeled. A well done model can be an excellent starting point, but at some point you have to enter the real world and actually manufacture and test the prototypes, simulation only goes so far.

 

[lucian]

Obviously, new Jupiter system is much faster then Klippel. And much, much pricier. And less accurate at the lowest frequencies.
But what are you meaning with "adjusting the result" for the Klippel? Are you implying that Klippel somehow is faking the measurements?

why do you expect it less accurate at lower frequencies than Klippel?

 

[NTK]

why do you expect it less accurate at lower frequencies than Klippel?

If you want a 20 Hz frequency resolution, you'll need a reflection free measurement window of 1/20 = 0.05 seconds. Sound travels 17 m in 0.05 seconds, therefore, for a cuboid room you will need the front and back walls to the speaker and mic to be at least 8.5 m (28 ft) away, and the left and right walls, ceiling and floor to be at least 9.5 m (31 ft) away.

The NFS does not have this frequency response resolution limitation because it gets its quasi-anechoic condition by using sound field separation, not time windowing.

 

[Matias]

Jupiter room is nothing new, it was built and started operating for some years now (since 2018 actually).

IMO it became obsolete when Klippel NFS was released for a fraction of the cost.

Here is what Dynaudio had to say about it.

G

Post in thread 'Dynaudio Special Forty - Review & Measurements by Erin'

Jul 24, 2023

The Jupiter room was great for the couple of years until the Klippel NFS was released and made it obsolete. Oops.

From Dynaudio

https://www.facebook.com/DynaudioGroup/photos/a.254075897969470/4201505769893110

We actually have a NFS and used it for verification and improvement of the measurement setup for Jupiter. In the daily work we prefer to use the measurement system in Jupiter due to the following reasons:

1. It takes about 6-10 hours to get a measurement with the same accuracy we can achieve in Jupiter in approximately 20 mins.

2. We have found that loudspeakers with multiple acoustic...

• GXAlan

• 

And my arguments.

Post in thread 'Dynaudio Special Forty - Review & Measurements by Erin'

Jul 24, 2023

From Dynaudio

https://www.facebook.com/DynaudioGroup/photos/a.254075897969470/4201505769893110

We actually have a NFS and used it for verification and improvement of the measurement setup for Jupiter. In the daily work we prefer to use the measurement system in Jupiter due to the following reasons:

1. It takes about 6-10 hours to get a measurement with the same accuracy we can achieve in Jupiter in approximately 20 mins.

2. We have found that loudspeakers with multiple acoustic sources (drivers and ports) require bespoke measurement setups specific to the speaker being measured. Investigating the...

• Matias

• 

 

[lucian]

If you want a 20 Hz frequency resolution, you'll need a reflection free measurement window of 1/20 = 0.05 seconds. Sound travels 17 m in 0.05 seconds, therefore, for a cuboid room you will need the front and back walls to the speaker and mic to be at least 8.5 m (28 ft) away, and the left and right walls, ceiling and floor to be at least 9.5 m (31 ft) away.

The NFS does not have this frequency response resolution limitation because it gets its quasi-anechoic condition by using sound field separation, not time windowing.

oh thx - i was assuming that when they make such an effort - they would use some more sophisticated than just windowing.

 

[NTK]

oh thx - i was assuming that when they make such an effort - they would use some more sophisticated than just windowing.

Like which more sophisticated method? "Near field" measurement (for low frequency) mean a separate measurement, which means more time and effort

 

[amirm]

An NFS scan of a typical speaker can be done in 2.5 hours with 15 minutes for computation. You could do 3 variations of the design in the same day! If they taking 4 to 6 hours each, then they are using too many samples.

My testing takes longer because I also perform a number of other tests (which would take the same amount of time regardless of what you used).

 

[amirm]

This is superior to a Klippel, this is the direct result without compensating or adjusting the result

There is nothing superior about it. They use gated measurements which Klippel also uses for high frequencies. For low frequencies, klippel uses field separation which is not a compensation but mathematics and extracts reflections. They use gating which means resolution lowers as frequencies do. Klippel has the advantage of near-field measurements which highly increases SNR and eliminated phase changes due to temperature gradients. Klippel's measurement accuracy is 1 degree which is far better than their 6 degrees.

There could be some speed advantages to their scheme but by the time they mount the speaker, lift it and measure, there is not much saving. Rotating a speaker is much trickier than rotating a mic as NFS does. Speaker needs to be very much secured which is not necessary for Klippel. They say that they measure each driver separately and then stich the results (this can create errors by itself). Each driver takes 20 minutes to measure. I don't know what this means for such things as ports.

 

[kemmler3D]

I would argue that a full “spin” is probably not necessary for much of the iterative design work.

Agree, unless what you're working on is the waveguide or a multi-driver configuration with directivity in mind?

 

[Grimoire]

So what is Dynaudio's excuse?

 

[lucian]

Like which more sophisticated method? "Near field" measurement (for low frequency) mean a separate measurement, which means more time and effort

since they measure 180° with one scan - it should be much faster. only z axis would need to be moved.

 

[Pearljam5000]

It looks much cooler than a Klippel that's for sure