On 20/12/2015 8:22 PM, Jim Lesurf wrote:
> In article <
ddlr68...@mid.individual.net>, Trevor Wilson
> <
tre...@SPAMBLOCKrageaudio.com.au> wrote:
>>> The problem with anechoic chambers is that what you measure is
>>> essentially irrelevant - unconnected with what a speaker in a room
>>> does. Sure it makes for consistency, but that really isn't good enough.
>
>> **Wrong. Given the fact that the characteristics of a (proper) anechoic
>> room are well known and that every listening room is different, then an
>> anechoic environment (or measurement systems which simulate such an
>> environment) are the only sane way to quantify a loudspeaker's
>> performance.
>
> I'm sure many speaker designers would *love* that belief to be correct! It
> would help them a great deal. Alas, to make it so would require something
> more than you state.
>
> To start you'd have to do the measurements around the entire theta/phi
> sphere. Then do them at different distances to assess near field effects.
> Then you'd have to find a way to assess the effects of reflections on the
> speaker itself as that may alter its coupling to its surroundings.
>
> Then you'd have to work out how to apply that to a range of listening rooma
> acoustics and listening positions, etc. In the process finding that the
> results depend on these so much that the whole process is a bit of a
> nightmare and you end up using judgement.
**Well, yes, you are correct. Defining and measuring speakers is very
challenging. I always have a little chuckle when I see speaker
measurements (even those in Stereophile), because the limitations are
huge. Many years ago, I had a copy of the specifications supplied by
Duntech for their Crown Prince model. It ran to more than 20 A4 pages.
Chock full of graphs, plots and specifications. And even that is not
enough to quantify the sound of a loudspeaker. It was a very good start.
John Dunlavy was an exceptionally talented speaker designer, who used
both an anechoic room and state of the art measurement equipment to
quantify the performance of his products. The important thing to note is
that all the tests were performed under anechoic, or simulated anechoic
conditions. The Crown Prince (and the bigger brother, the Sovereign)
stand up today is stunning sounding speakers. Even more than 25 years
after they were released. And no, I am not trying to compare the
Duntechs to the LS3/5a (that would be silly, as the Sovereign is one of
the few speakers on the planet which can provide a decent reproduction
of a full orchestral piece), but to indicate the process.
>
> Much as I am very keen on basing our understanding on *appropriate*
> measurement and analysis I am also very aware that in the end speakers are
> made and sold for people to put them into all kinds of rooms to listen to
> all kinds of music and enjoy the results. Afraid there are too many
> variables out of the developers control/awarness to make this as simple as
> your sweeping description.
**My point exactly. An anechoic test is essential, BECAUSE of the
massive variety of rooms.
>
> Anechoic measurements are very handy, and methods like gated pulses, etc,
> can also be handy. So make good sense. But they won't be enough in
> themselves. Measurements can just as easily mislead as reveal. Depends on
> the measurements *and* the judgement of the people making them and
> interpreting the results.
**Correct. Just as the opinions of those who do not bother listening to
different products is suspect.