assessment of loudspeakers

[pdan]

Inspired by the 'logic' of the 'first watt', I wondered if a similar 'logic' could be applied to the assessment of loudspeakers.

What I've come up with, as a suggestion, is the notion of the 'first test' which the loudspeaker MUST pass. Briefly put: Is the loudspeaker able to reproduce speech in a effortless natural manner. If yes then the loudspeaker has passed the first test and the loudspeaker design has merit. If not, then its back to the drawing board.

In other words, we do extensive listening of speech ( Audio books are great for this ) before listening to music. Since we are so directly familiar with speech any false 'notes' are easily heard, not only that but also, the fr ( about 100hz - 10000 ) takes us right to the heart of the matter.

So, just as the 'first watt' is the heart of a great amp, I suggest that a passed 'first test' is the heart of a great loudspeaker.

pdan

[Cal Weldon]

Originally posted by pdan
What I've come up with

Wow, I'm glad to finally meet you after all these years. I was thinking you must be dead by now.

What you have "come up with" my friend is probably the best, tried and true method for designing a loudspeaker. Start with the mid, see what it's capable of and work your way out from there.

Welcome to the world of diyAudio. What a long strange trip it's been.

[Tenson]

I have to disagree. Our hearing is very tolerant of poor speech conditions since we are 'trained' to understand the meaning, not assess the tonal quality and subtle noises. Human hearing is great at picking out speech from below a noise floor or really bad acoustics but in most cases we don't pay attention to the voice itself.

Thats not to say you can’t determine anything from doing the test you describe. I find male voice can often tells you a lot about the 100Hz-500Hz range but it won't tell you if the rest of the frequency response is particularly flat.

Personally I find distorted electric guitar tells a great deal, as it is close to wide-band noise... I suppose playing in a band helps make that a good test for me too.

[Cal Weldon]

Originally posted by Tenson
I have to disagree.

Uh-oh, now we're in trouble.

Actually I can't say listening to voice is the best place to start but I do like the idea of starting at the mid. Perhaps I should have been more clear in my analysis of his post.

Now go to bed, you're up too late!

[Hartono]

Hi Pdan,

"Inspired by the 'logic' of the 'first watt', I wondered if a similar 'logic' could be applied to the assessment of loudspeakers.

What I've come up with, as a suggestion, is the notion of the 'first test' which the loudspeaker MUST pass. Briefly put: Is the loudspeaker able to reproduce speech in a effortless natural manner. If yes then the loudspeaker has passed the first test and the loudspeaker design has merit. If not, then its back to the drawing board."

you mean we should ***** the speaker when driven with current ? this might be a good idea.

I've particularly interested in using current drive for amplifier.


Hartono

[Saturnus]

I partly agree with Tension. The tolerance of noise in vocal sounds is the main reason why phones, and especially mobile phone works. They have absolutely terrible sound quality and bandwidth but are none-the-less great for communication.

I only partly agree because the trick is to listen to voices in a language you don't understand at all. Then your brain will stop trying to understand the speech and you can more easily listen to tonal quality.

[Cal Weldon]

Saturnus,

I'm having trouble with your comparison. The tolerance of noise in vocal sounds is just that, a tolerance. I think the midrange must be the least coloured part of the system as we are most able to distinguish good from bad in that area.

Interesting idea of listening to foriegn voices but how do you know what that person is supposed to sound like when you don't ever listen to them because you can't understand them?

[Tenson]

How are you supposed to know what any other person is supposed to sound like unless you have talked in person with them?

Maybe make a recording of someone’s voice that you know well and use that.

[Cal Weldon]

I use pieces of music myself. No woofer, no tweeter, just me, the midrange, a Jimmy Buffett album and a fine pale ale.

[soongsc]

The voices of people that we really know are quite good for first test. If we hear something wrong, how do we identify whether it's the mic or the speaker?