Should we (I) get into speaker testing & measurement

[Fluffy]

Who cares? Why would it matter? It is like asking "how would my car perform without power steering and automatic traction control?". Perhaps interesting to a car suspension/handling design engineer, but not to an user.

I've seen how DSP headphones like the ones from Bose or Sony display unacceptable amounts of THD, as their mediocre driver is forced into submission by the brutal DSP equalizations.

The discussion on this thread and these measurements indeed point out to some high distortions in the bass area with the Dutch & Dutch 8c. If there are DSP speakers that don't display those levels of distortion, then that is something worth considering.

 

[RayDunzl]

Dabbling is painfully resource intensive.

I guess you aren't a dabbler.

dabbler: someone who takes a slight and not very serious interest in a subject, or tries a particular activity for a short period

 

[amirm]

I guess you aren't a dabbler.

dabbler: someone who takes a slight and not very serious interest in a subject, or tries a particular activity for a short period

I know that is what it means. I assumed you meant to measure a few speakers the manual way versus a lot more using the automated system.

What are you proposing outside of that?

 

[q3cpma]

I've seen how DSP headphones like the ones from Bose or Sony display unacceptable amounts of THD, as their mediocre driver is forced into submission by the brutal DSP equalizations.

The discussion on this thread and these measurements indeed point out to some high distortions in the bass area with the Dutch & Dutch 8c. If there are DSP speakers that don't display those levels of distortion, then that is something worth considering.

Unlike what some may say, it's impossible or very hard to replace woofer size with anything else when going for full range. Look at the K&H o500c (http://www.neumann-kh-line.com/neum...g_discontinued-monitors_studio-products_O500C) if you want to see DSP complementing an already incredible speaker, and not being used as glorified damage control.

 

[RayDunzl]

I listen in a room.

Everybody listens in a room.

I could accept measurements that are taken across a range of speakers operating under the same conditions - i.e., in the same position in a particular room with the same excitation.

After a few measures you might be able to tease some of the room out of the data, show a raw and "adjusted" measure..

Advantages: Low cost, easy to try, maybe learn something new, discover distortion levels, compression onset, frequency limits, truth in specifications, matching between a pair, etc.

Cons: Doesn't make you a speaker measuring god.

 

[617]

I listen in a room.

Everybody listens in a room.

I could accept measurements that are taken across a range of speakers operating under the same conditions - i.e., in the same position in a particular room with the same excitation.

After a few measures you might be able to tease some of the room out of the data, show a raw and "adjusted" measure..

Advantages: Low cost, easy to try, maybe learn something new, discover distortion levels, compression onset, frequency limits, truth in specifications, matching between a pair, etc.

Cons: Doesn't make you a speaker measuring god.

The goal of serious home speaker designers these days is speakers which are relatively room independent. Some speakers, like big horns, will sound relatively similar in different rooms. Bass will always sound different based on measuring location and speaker location. Conventional box speakers will sound pretty different based on how they are set up. Dipoles are somewhere in between.

Also, Amir, there are professional measurement products which cost far less than the one you are looking at. Ground plane measurement is not really required for bass, nearfield measurement works really well and provides great detail.

Here are some links to platforms which work with Clio, arta, etc:
https://outline.it/outline-products/measurement-systems/et-250-3d/
https://www.audioxpress.com/news/he...es-high-precision-hrt-i-measurement-turntable

 

[amirm]

I have looked at all those options. Still a lot of money (thousands for the turntable alone) and lots of manual work. People who do that put out a review every other month. My goal is to be as speedy as my electronics measurement. The Klippel system measures a speaker in as little as 20 minutes with little intervention. I don't know that we can move the mark in the industry by measuring 4 or 5 speakers a year.

 

[Juhazi]

Distortion of headphoes and speakers is not linear - it gets higher with higher spl. The spl capacity without audible distortion is pretty much dependant on driver surface area, cone excursion capacity comes second with the expense of distortion.

D&D 8c has acoustical cardioid response for the midrange, which means poor efficiency and makes it prone to distortion, or limiting it's usable spl. It is however ok for most domestic hifi listening scenarios.

I have been asking for publification of distortion measurement, but Amir doesn't see it necessary. I still think that some kind of stress test is needed, spl/dispersion doesn't tell enough. It is easy to measure distortion profile in a standardized way without anechoic chamber. Room effects are same and can easily be normalized. Many web sites do this, but it is very difficult to compare different tests to each other.

From Sounstge/NRC, here is D&D 88c and a small two-way and a large 3-way. All measurements here https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=16&Itemid=140

 

[617]

I have looked at all those options. Still a lot of money (thousands for the turntable alone) and lots of manual work. People who do that put out a review every other month. My goal is to be as speedy as my electronics measurement. The Klippel system measures a speaker in as little as 20 minutes with little intervention. I don't know that we can move the mark in the industry by measuring 4 or 5 speakers a year.

Assume the klippel system measures distortion in some comparable way? How does it measure bass? Nearfield I assume.

I see where you're coming from. I measure my own designs and it probably takes a 40 minutes to get polars, one axis.

How long does it take you to measure a DAC?

Also have you considered offering to test speakers for manufacturers? Might help offset the cost. I know Kali uses a lab in the Midwest to do the measurements they publish in clf format.

 

[edechamps]

I've seen how DSP headphones like the ones from Bose or Sony display unacceptable amounts of THD, as their mediocre driver is forced into submission by the brutal DSP equalizations.

Even then, the important part is the THD itself, not how it's produced. The fact that it uses DSP is irrelevant; it's the measurement of the end-to-end system that matters. As long as I have access to all the relevant measurements, I don't care if the speaker uses DSP, waveguides, horns, crossovers, coaxial drivers, reflex ports, acoustic transmission lines, pixie dust, automatic traction control, or whatever else. For the purposes of choosing a speaker as an end user, all that matters is the end result, not how it's achieved.

(Corollary: most of what's written in marketing materials - or, god forbid, audiophile magazines - is of little value as they seldom convey information that pertains to the actual end-to-end performance of the entire unit. Debates about internals and mechanisms are not the best way to determine which product is best - measurements are.)

 

[Fluffy]

@edechamps
I agree, the end result is what matters. That's why I think DSP is just another solution among many to the regular problems that plague speakers. It's all about implementation and outcomes, and DSP surely isn't a perfect cure-for-all, otherwise everyone would use it.

The exception is of course pixie dust. I only buy speakers that has at least 5% TPD (total pixie dust).

 

[edechamps]

DSP surely isn't a perfect cure-for-all, otherwise everyone would use it

Of course. For example DSP cannot fix directivity issues, so even if you had perfect DSP it would not be sufficient to make a perfect speaker. It can help, though.

 

[fredoamigo]

I have decided that if this needs to be done, I need to pay for it and then hope members over time help offset the cost. Waiting for it to be funded could take too long, if practical at all.

The challenge now is how I can justify spending the cost of 2 to 3 new cars on this in my retirement days. Or how I can justify it based on how many speakers I can test.

On the positive side, I am very interested in doing it.

the audio industry doesn't like measurements, especially if they are made by an independent... don't you fear some kind of "coalition" against you?

 

[watchnerd]

I don't know that we can move the mark in the industry by measuring 4 or 5 speakers a year.

But you might move it more than measuring no speakers at all, which is the current state.

I don't think you need to boil the ocean on your first attempt.

Start with some training wheels with in room measurements. That's still vastly more informative than most subjective-only review. Once you get better at using a basic set up quickly, then upgrade the test harness.

 

[Juhazi]

Here is Klippel's papers about loudspeaker distortion. Linear and non-linear types well explained.

https://www.diyaudio.com/forums/att...rtion-caused-design_part-1_klippel_werner-pdf

https://www.diyaudio.com/forums/att...-distortion-defects_part-2-klippel_werner-pdf

 

[Vovgan]

You have to lift speakers to 9 to 10 feet or more outside. Rotate them 360 degrees. Move the mic up and down and do that again. Then for bass, put them on the ground away from any barrier for hundreds of feet. Perform "ground level" measurements then. Then stitch the two curves together, knowing that the latter measurement includes a gain that cannot be backed out.

Speakers measurement does seem to be a heavy cross to bear (both financially and physically), but unless you or somebody else earnestly take it up, we will fetishize ourselves to death here with the narcissism of small differences between Matrix/Topping/Chord/Benchmarks of the world, while letting time and chance happen to the most critical part of music reproduction:

 

[Bjorn]

Have you considered sending the speakers to i.e. NWAA Labs with the help of donations?
https://nwaalabs.ipower.com/our_services.html

 

[Juhazi]

nwaalabs charge $1600 per standard speaker test...

This Loudsoft system might be cheaper than Klippel. It presents data differently, but perhaps there is option to make H&K spinoramas
https://www.loudsoft.com/fine-rd/

Let Peter Larsen tell himself:

 

[watchnerd]

Speakers measurement does seem to be a heavy cross to bear (both financially and physically), but unless you or somebody else earnestly take it up, we will fetishize ourselves to death here with the narcissism of small differences between Matrix/Topping/Chord/Benchmarks of the world, while letting time and chance happen to the most critical part of music reproduction:

View attachment 37533

I was calling this measurement fetishism, but I like "narcissism of small differences", too.

And, I concur.

It's somewhat ironic for a site called "Audio Science Review" to spend so much energy measuring things that make such little difference to what we actually hear......at least according to psychoacoustic science.

 

[RayDunzl]

It's somewhat ironic for a site called "Audio Science Review" to spend so much energy measuring things that make such little difference to what we actually hear......at least according to psychoacoustic science.

Life is filled with little ironies.

Audio Buddy had a fetish for Chrysler Slant 6, and had a nice(ish) '63 or so Valiant Station Wagon. He lovingly rebuilt the engine and Push-Button Automated Transmission, and installed an engine oil pressure gauge. It may have replaced or bypassed the idiot light.

Too soon after, a fitting expired on the gauge setup, blew out the oil, and fried the motor on the interstate.