Accelerometer vibration measurements!

[bobkatz]

Dear John: I intend to do as scientific a study as possible of the effects of vibration isolators (e.g. the EVP Isolators from AV Room Service: http://avroomservice.com/evp-2/

I have Spectraplus, which is an analyzer that has integrated vibration measurements, calibrated in units relative to 1 G. Spectraplus supports the Digiducer, which is a digital accelerometer. Attached is a PDF describing the calibration procedure. In some ways it's like the approach that you take to read the Mini DSP USB microphone.

At this point I don't know if I can read the time delay between the stimulus (which will be the loudspeakers) and the vibration measurement in Spectraplus. Nor do I know if I can see an impulse in Spectraplus as well as I can in REW. But I do know how to display an impulse and measure time of flight in REW. It would be wonderful if we could measure vibration in REW. The units would be in G Units, which is the direct output of the accelerometer. I would display amplitude in Spectraplus in decibels relative to 1 G. It would be wonderful if we could adopt REW to do the same kind of reading and use REW for the accelerometer measurements as well as acoustic analysis.

OK, I just thought of the biggest obstacle, ASIO drivers are not available for the Digiducer. If we can conquer the calibration issue, read the calibration from the Digiducer, do you think we could use ASIO For all for this in REW for windows?

I am planning on renting a Digiducer soon. It rents for $130 for a 30 day period, not too expensive but I'd like to keep my costs down. Are you interested in expanding REW for vibration analysis?

 

[3ll3d00d]

seems to be v heavy for an accelerometer, they're usually more like a couple of g so they are then easy to fix to something and extremely light relative to the device under test.

FWIW I've written an app that runs on an rpi and can consume data from a commonly available accelerometer (https://github.com/3ll3d00d/qvibe-analyser and https://github.com/3ll3d00d/qvibe-recorder), it has a mode which lets you record the data from a REW sweep and then import it into REW for further analysis. These devices only sample at 1kHz so are only good to say 250Hz. There is also an analogue accelerometer which can be used and there's a guy on diyaudio who sells (or used to sell) a preamp for them (https://www.diyaudio.com/forums/equ...ccelerometer-testing-loudspeaker-drivers.html), these can be fed straight into REW (just plug them into an audio interface like a mic).

It would be handy to add G as a measurement unit for this use though, if you do that then ISO 1683 states the reference level for acceleration at 1 μm/s² ≡ 0 dB

 

[DanDan]

dBPod | Studio Six Digital

studiosixdigital.com

I think anything like that should be good for comparative measurements. But I would guess that REW already has all the facilities needed to measure any audio artefacts created by cabinet or tweeter vibration. I would further guess that for any decently designed, and weight, of speaker such artefacts are inaudible.

 

[John Mulcahy]

Hi Bob,

Java (at the application level) isn't great for accessing information on USB devices and tends to have particular problems with devices that are audio interfaces, since the OS tends to connect to them in a way that makes it difficult for Java apps to communicate with them (in my experience at least). I think this fits in the too difficult category, for now anyway.

 

[DanDan]

Just noticed Andrew Smith/studiosixdigital announced yet another app or suite of them called VibrationTools, Q42019. Maybe it's all there.

 

[AJ Soundfield]

Dear John: I intend to do as scientific a study as possible of the effects of vibration isolators

Hi Bob, audible effects, or measured parameters effects?

cheers,

 

[DanDan]

Ethan Winer has done some testing. http://ethanwiner.com/speaker_isolation.htm
Or take a seat and microwave some Popcorn..... https://www.gearslutz.com/board/stu...6-testing-loudspeaker-isolation-products.html
Just an odd FYI. The BBC Standard Monitor BC1, manufactured by Spendor, used a lossy cabinet. It was designed to radiate presumably in a controlled way.

A great Acoustician, my favourite, sadly and very prematurely RIP, custom made isolation for the Speakers which he also designed and built, in Studios using his unique acoustic technique. He freely shared his thoughts so I suspect you may find great wisdom with some googling. Boggy, Bogic Petrovic, Myroom Acoustics.
Also, afaik his people are continuing his work, perhaps worth a query. Given that he made everything it seems very likely that he has practical views on what really matters from speaker to ear.

 

[bobkatz]

Dear 3: Thanks for the info on the accelerometers. The Digiducer is supposed to be very light and what I like is my Spectraplus software is already equipped to read the calibration data from the USB. Plus it's good up to around 1 kHz, which may be helpful, maybe not.

Thanks for the info. I was planning on using a 1 g reference = 0 dB for acceleration, but it makes a lot more sense to use a lower reference and I can set that reference for 0 dB in Spectraplus.

 

[bobkatz]

Dear Dan Dan: it's quite difficult if not impossible to prove audibility of these kinds of vibrations by a double blind listening test. For the moment i'm sticking to measurements and informal listening tests. My readers can take or leave my subjective opinions but will have a hard time arguing over the measurements. :-)

 

[bobkatz]

Dear AJ: 99% measurements and 1% sighted observation.

 

[bobkatz]

Dear DanDan: Ethan and I are at big odds over the audibility of these phenomena. I am convinced that the EVP isolators work and work well. I've done the sighted listening tests with my 300 pound Dynaudios. There are some remarkable improvements. Remember that sound travels faster through a cement floor than it does through the air. This produces a "double image" which smears the quality of the sound. I heard that improvement and I'm not claiming I can prove that blind, I'm simply giving you my subjective observation. I also noticed the sound in the area between the back of the speakers and the front wall became eerily quiet after the addition of the isolators.

I performed some objective measurements and published them on my web site. A comparison video waterfall shows extreme synchronicity between the before/after waterfalls near 300 Hz and meaningful differences below that. I personally find Ethan to be stubborn when things don't go his way and he has never expressed any useful scientific observation that would provide an alternative explanation to the differences and similarities in my waterfall video. He goes into denial, claiming that a change of far less than an inch in height or position of the monitors or microphone invalidates the measurements. Nevertheless, the coherence near 300 Hz and the differences below 300 Hz is very strong evidence that the only variable is the isolators. The microphone is in exactly the same spot, to within 1 sample at 48 kHz sampling, using REW timing measurements and comfirmed with laser interferometer distance measurements accurate to the millimeter. The height of the speakers is as close to identical as possible, having constructed the isolators to the same height (with displacement) as the wood blocks that were formerly under the speakers.

Since Ethan and his followers are skeptics, I'm working on this new measurement series. I will document my new measurements and measurement procedure using a scientific notebook, witnesses, photographs and physical measurements. I will present the measurements in a (hopefully) peer-reviewable form. Then we will set this issue to rest, although I suspect Ethan and his followers will attempt to tear apart my new review. As long as they respond in a scientific manner using scientific method I shall not mind their objections.

Here's my blog with the video and my explanation why this is strong evidence of the efficacy of vibration isolators placed under loudspeakers: https://www.digido.com/installing-dynaudio-m5p-loudspeakers-on-custom-isolators-from-av-room/

I'll be happy to hear your reactions to my post.

 

[DanDan]

LOL, Ethan is a good pal and a cussed donkey. He is a bit tripped out on that singular myth busting stance. To the extent of misreading graphs and really falsifying tests on Speaker Controllers, e.g. Dirac Live. But.... a lot of what he points out is really quite gobsmacking and humbling even. Did you try that Converter test back in the day. I did, and despite my HD192, a Lavry, and a Teac, which I could readily identify, I preferred the frikin SoundBlaster card in his scenario........
I took a quick look, your waterfall difference is substantial, so should be really audible. Yes sound travels from the speaker cab though a solid floor much faster than through air. So a secondary path radiating LF up from below us, earlier than the sound from the woofer to ear. But going into devil's advocate mode......
These speakers look impressive. What is the base made of? I realise you needed to raise them on wooden pucks to maintain the same height as with the Isolators, but could such pucks have deteriorated the performance of the manufacturers base?
Probably not. But it then looks like the manufacturers base has no isolating effect at all. On a concrete floor, without carpet, this would be a problem IMO. At least they should warn or specify that isolation is necessary. PMC do pretty much insist on use of their own stands....
Your comments on rear radiation....... clearly the LF would be everywhere but what were you hearing? If there were other frequencies, then this cabinet is behaving very unusually I reckon. Most high end cabs are really brick like and there is no audible or other tactile radiation.
That said, not many are floor standers. I reckon pretty much everybody uses some form of isolation either at top or bottom of the stands, or at both ends even.
Overall I think Dynaudio should be made aware of this anomolous situation. The BBC BC1 Monitor had a lossy cabinet, which everybody knew about.