MLS is narrow "pulsed" signal that can be easier to achieve a result more similar to an anechoic result. At lower freq.s though you need more time for each pulse and you need a low-noise environment.
Basically freq. response shaping/correction on the low-end to deal with room and proximity issues.
Ex. (blue is corrected result)
This kind of "Klippel" testing is the kind that gives BL(x) Kms(x) Le(x) etc. and is done with a laser scanner and different software, that while expensive is nowhere near the cost of the NFS. The NFS is a very different setup for measuring completed speaker designs.
It's likely going to cost more than $500 to get a circle approximately 2 meters+ in diameter with more than 12" thick paneling for walls, ceiling, and floor. I'm probably going to go with wire mesh as the support rather than hanging, and using magnets to attach the framing to the ceiling (with plastered-over screws in the ceiling for those magnets).
It is on my "to-do" list though, but recently I've been "mired" with running electrical cable for a different project (getting 220v 50amp to my garage for a welding machine and plasma torch). 50 ft of 6 gauge is a B!tch to work with.
I don't think it's "inaccurate", but I do think it's over-emphasized from at least 150 Hz to perhaps as high as 1 kHz when compared to a far-field (2m) result.
Note: I'm also not stating that the horn is non-diffractive, in fact I think it's exceedingly diffractive (more than 130 degrees of "sharp" bend) with added resonant issues (owing to the size and depth)
Last edited:
This old BBC paper has some good information on the performance of different types of wedges for anechoic chambers
https://www.pearl-hifi.com/03_Prod_Serv/PR2/Refs/Anechoic_Chamber_Design_and_Construction.pdf
That doesn't match my understanding.
My understanding is MLS has the same time constant properties as any stimulus signal, as they all have to deal with the time vs freq resolution tradeoff..
and I think MLS has proven to be inferior to swept-sine and periodic noise for most acoustic measurements.
(I'm NOT a super guru on this...but feel pretty confident about those points.)
Is that measured acoustic response ? Looks like it's just an electrical high pass filter tacked on, if blue is acoustic measured (which looks doubtful given how linear the rolloff is)
Can't see what a hpf has to do with acoustic measurements down low...other than maybe eliminating response (or truncating data)
I continue to think, there is simply no substitute for reflection free measurements without gating, for low freq work.
At low frequencies the wavelength of sound is large compared to the cabinet and any reasonable driver is moving in a smooth and predictable manner. There is no need to measure with a microphone in order to accurately determine the sound being radiated. Manufacturers may want to do so for non-engineering reasons and checking is always wise but there is no primary engineering reason to do it since at these frequencies even simple predictions can be expected to be more accurate than measurements. At higher frequencies measurements of the radiated sound from a speaker with a microphone can provide useful information and in a room at low frequencies a microphone certainly provides useful engineering information but for speaker radiation...
Unless you are a manufacturer operating on a large scale spending large sums of money to get reasonably believable low frequency microphone measurements just isn't cost effective in order to engineer speakers. To make youtube videos or support audiophile forums perhaps but that is a different activity to designing and making speakers in a home environment.
As others have suggested in the thread, for those with the inclination learning a bit more about the physics and engineering is likely to be a more cost effective way forward than stumping up substantial sums of money to get access to expensive measurement tools. Obviously this depends on ones view of scientific knowledge and how much "school" work one wishes to include in a hobby activity but it is a viable alternative these days given the modest cost of computation.
Here's a thought:
Look at the manufacturers and DIY'rs who produce data for their designs. Then look at the NFS results.
I'd say many, many of them are going to be off by a fair bit. Most notably what I find is baffle step influence (not just baffle step but the influence combined with other things like port resonances). Manufacturers such as Kef and Revel/JBL don't tend to have this issue. But they're unfortunately the exception, based on my direct experience.
Statistically speaking, DIYr's don't have a great track record thus far. DIYSG designs is one shining example. Overnight Sensations, Parts Express Samba... there's plenty to count. I've had some DIY "big wigs" try to accuse me of creating bad data for controversy (clicks) only to have them recant after me proving multiple ways that my results are indeed legitimate and that their DIY ways simply aren't sufficient. Especially below about 1kHz. And almost certainly below about 300Hz. The strong enclosure/port resonances aren't apparent with gated response and the bass results are often off due to not understanding how the standard GP methods can lead to different results simply based on orientation. Especially with large (tower) speakers.
So, if you're one saying that the NFS isn't necessary, it might be fun to test your design on the NFS and then compare the results. You might be right. Maybe your methods are good enough. But odds are, you're going to find you're wrong. Not being snarky. I'm speaking purely from experience without any malice.
- Erin
Look at the manufacturers and DIY'rs who produce data for their designs. Then look at the NFS results.
I'd say many, many of them are going to be off by a fair bit. Most notably what I find is baffle step influence (not just baffle step but the influence combined with other things like port resonances). Manufacturers such as Kef and Revel/JBL don't tend to have this issue. But they're unfortunately the exception, based on my direct experience.
Statistically speaking, DIYr's don't have a great track record thus far. DIYSG designs is one shining example. Overnight Sensations, Parts Express Samba... there's plenty to count. I've had some DIY "big wigs" try to accuse me of creating bad data for controversy (clicks) only to have them recant after me proving multiple ways that my results are indeed legitimate and that their DIY ways simply aren't sufficient. Especially below about 1kHz. And almost certainly below about 300Hz. The strong enclosure/port resonances aren't apparent with gated response and the bass results are often off due to not understanding how the standard GP methods can lead to different results simply based on orientation. Especially with large (tower) speakers.
So, if you're one saying that the NFS isn't necessary, it might be fun to test your design on the NFS and then compare the results. You might be right. Maybe your methods are good enough. But odds are, you're going to find you're wrong. Not being snarky. I'm speaking purely from experience without any malice.
- Erin
Thanks for the firsthand experience Erin. Unfortunately, I can't even get someone to discuss this live where much more progress and clarity over the other's position could be made. I can guarantee they will not be subjecting their designs to scientific scrutiny, no matter how much they say they are citing The Science (TM).
The practical reality you lay out has been laid out before, and it's like some of the critics don't even read it. They just respond with their own irrelevant nit-pick.
queue "if you were a certified professional like me you would know the answers," and "I'm baffled by all these stupid humans, maybe I can help by insulting them but at the same time providing no concrete method myself" in 3, 2, 1...
This is a bit hand-waving unless you explicitly state the frequency range you mean by "low frequencies". Under 50Hz? Under 100Hz? Under 500Hz?
IMHO one of the biggest loudspeaker measurement challenges is in the range of 200-800Hz. Here you can have cabinet and driver resonances, but it is difficult to get very high resolution measurements without resorting to ground plane or elevating the speaker far above the ground, or using a large anechoic chamber.
Below 100-200Hz things get easier again.
And I apologize for getting a bit angry about this, but at least two people have shown they are not good-faith participants. They seem to have no self-awareness of how condescending their comments are, and myself and others are tired of it.
For my situation there are practical issues like traffic and wind that are hard to escape. And building a test rig I would trust to not topple over on my head. I think the open space idea is great in principle, but the fact that we see someone post of doing it that day, happens, what, once a year? I think that tells us most of us just can't pull that off. Something that can be done in one's house would be ideal.
I've managed subwoofer measurements in my local park with a battery powered amp. The measurements are very clean compared to anything else I have managed to get:
https://www.diyaudio.com/community/threads/isobaric-sub-based-on-pyle-plpw15d.345807/#post-7049973
I often take measurements for system tuning at festivals and find with sufficient averaging you can measure in pretty noisy environments.
I'm also building a big turntable to get polars, on my turntable I'm fitting a subwoofer distance pole with hand crank so I should be able to get pole mount speakers a bit higher than the average DIY'er. For very large speakers like my midbass horns though I have to resort to ground plane measurments and then matching the groundplane with BEM (BEM of groundplane) then assuming I have matching using BEM of the freespace to get polars.
If your willing to spend a few thousand you can get a line array lift tower and make a large turntable for it or a cherry picker. Again all depends on land value!
https://www.diyaudio.com/community/threads/isobaric-sub-based-on-pyle-plpw15d.345807/#post-7049973
I often take measurements for system tuning at festivals and find with sufficient averaging you can measure in pretty noisy environments.
I'm also building a big turntable to get polars, on my turntable I'm fitting a subwoofer distance pole with hand crank so I should be able to get pole mount speakers a bit higher than the average DIY'er. For very large speakers like my midbass horns though I have to resort to ground plane measurments and then matching the groundplane with BEM (BEM of groundplane) then assuming I have matching using BEM of the freespace to get polars.
If your willing to spend a few thousand you can get a line array lift tower and make a large turntable for it or a cherry picker. Again all depends on land value!
Hello Erin, No disagreement for my side: the shortcomings of quick and dirty ARTA, REW etc measurements between 200 and 1000 Hz are known.
The issue, which as been adressed in the Klippel on shoestring thread, is whether there are cheaper ways to skin the cat. That has also been the topic on ASR. Member Aslepekis seems to have gotten fairly close to the Klippel measurements in Shoestring thread.
The issue, which as been adressed in the Klippel on shoestring thread, is whether there are cheaper ways to skin the cat. That has also been the topic on ASR. Member Aslepekis seems to have gotten fairly close to the Klippel measurements in Shoestring thread.
kipman725> have you ever measured a full spectrum speaker using ground plane? Have you compared it to other methods? Ground plane is ideal for low frequencies, but so much higher ones from what I've seen over the years on these forums.
The other issue with groundplane is that if you want to make a measurement on the listening axis you have to tip the speaker forward so that that axis intersects the microphone. Otherwise you are just measuring the "on the floor" response. Who listens there???
But that is just for the "on axis" position. How do you tilt the loudspeaker for other axes???
Also, you still need a relatively reflection free environment for gruondplane measurements, meaning away from large objects. So it is not solving all the problems out there.
I haven't yet but this thread has good results: https://www.diyaudio.com/community/...how-close-is-close-enough-to-anechoic.353347/
for polars it should be fine to rotate the speaker.
For reflection free enviroment I recomend an empty car park and a battery powered amp, should be posibile almost everywhere.