Is there a right way to measure a speaker? (1 Viewer)

[Michael R Price]

Recently there was a thread where we discussed Sound&Vision magazine's methods for reviewing speakers. This raised a few questions in my mind...

How much do measurements actually tell us about how a speaker sounds?

What kind of measurements do we need to get a good idea about a speaker's performance? Is frequency response alone enough?

Do modern hi-fi/HT magazines provide us with the right measurements for us to make informed decisions?

I think measurements can tell us a lot about a speaker. I've been able to correlate my experiences listening to my Kit281s, for example, with some of the measurements in the Audioxpress review of them. Which is very interesting. If we want to get past all the subjective commentation about a speaker, I think good measurements are very helpful.

Being a technical geek I follow the doctrine of 'more is better'. You cannot have too many measurements. On/off-axis frequency response, impulse response, spectral decay, THD, IMD, I find it all quite interesting. Whether it's practical for a magazine to do all this work, that's another story.

What kind of measurements would you like to see to go along with the (often pointless) subjective commentary we see in speaker reviews?

Well, those are just some of my thoughts. I still haven't decided if I prefer being subjective or objective. Guess I'm both.

 

[Bob McElfresh]

It's very hard to find 1 aspect to measure. Look at all the problems with calling a receiver a "High Power 120 watt receiver" -- then read the fine print to see "Stereo/2-channel mode, measured Peak-to-peak, 3% THD into a 4 ohm load".
Frequency response is one way - but that is usually measured 1 ft away in front. Some speakers do great in front, but have off-axis problems. So only a few labs/review sites measure off-axis response.
And how often do you listen to single-tone music/sound? A speaker has to produce a lot of different sounds at the same time. A good frequency response to test-tones does not show how it responds to sudden changes - common in music & movies.
And last - speakers are a matter of taste. If I took 5 different brands of Chocolate Ice cream and asked you to order them from "best" to "worst", your order would be different from mine. (for all I know, you are one of the sickos who like Butter Brickle )
We cannot currently put numbers/analytics on matters of taste. So it will be really hard to document how a speaker "tastes".

 

[John Desmond]

It would be great if there were some universal, standardized way of showing meaningful speaker tests. seems to me you'd need anechoic frequency response on axis to at least 45 degrees off. You'd also need response above and below the optimum listening height. Any you'd need some sort of "room response" . But what room? You'd also want to see distortion specs at various frequencies. lMaybe some sort of impulse response. Heck, you'd need 20-30 pages for each speaker! Hey, that's an idea, a website or mag that does REALLY complete tests of one or two speakers a month. Who's going to do that kind of work? I think the test that Don Keele used to do for Audio mag were some of the best. You could at least get a feel for what the speaker would sound like.

 

[Ted Lee]

for me, i usually want to know how much power it can handle and what the frequency response is.
i'm still trying to figure out how to read those graphs in sght - if anyone knows puh-leez explain to me.
all the other stuff (on/off axis, anechoic this, resonance that) is greek to me. i'm sure it's meaningful data, but i just can't see how any of it applies to 'real world' situations.
ultimately, i figure it's going to be myr ears that decide what sounds good.
fwiw - i love ben & jerry's super chocolate chunk fudge. yummy!

 

[Michael R Price]

IMO, an example of the kind of measurements we should see in speaker reviews:
http://www.adireaudio.com/Files/AX-KIT281Review.pdf
(BTW, I own those )

 

[Chris Tsutsui]

Patrick you just gave me an idea for testing the response for my speakers in yet another way. I'm going to overlap a crossover test from a slow sweep 10hz-10khz with a 10khz-10hz so that each speaker plays 2 types of notes at the same time mono.

I wonder if the response will be severly out of wack which it probably will be.

Let us recall what Dr. Bose said and that the only true way to judge a speaker is by listening to it. I agree with him on that with specs being a nice addition.