How much can be measured -- and how much cannot?

I second Mapman's comments. As I see it the main usefulness of measurements, and also specifications, is that they allow one to identify and RULE OUT candidates for purchase that would be poor matches to either the rest of the system (e.g., impedance incompatibilities, gain and sensitivity mismatches, etc.) or to the user's requirements (e.g., maximum volume capability, deep bass extension, etc.).

Also, they can facilitate diagnosis of issues that may exist with components that have already been purchased, but that may not be optimally matched to the rest of the system.

The number of times that I and many others have referred to John Atkinson's measurements in helping Audiogon members to make purchase decisions or to diagnose problems is practically countless.

A secondary benefit is that measurements and specifications can help to provide a better understanding of the design concepts and approaches that underlie a product, which is something that I always prefer to have.

Regards,
-- Al

03-19-12: Ahendler
Actually I have never found measurements to help me in any audio decision to by a certain product. Measurements are helpful in the design process but do not tell you anything how a component will sound.

They won't tell you that the component will sound good, but they will often enable you to predict that the component will sound bad in a particular application, as a result of the kinds of mismatches that have been cited.

Basically, they allow you to narrow the field of potential candidates for purchase, to minimize trial and error, and to reduce the likelihood of expensive mistakes.

Regards,
-- Al

03-19-12: Ahendler
Actually I have never found measurements to help me in any audio decision to by a certain product. Measurements are helpful in the design process but do not tell you anything how a component will sound.

Let me supplement my previous response with some specific examples, that illustrate what I, Mapman, Roscoe and others have been alluding to:

1)Someone has read good things about the sound quality that SET amplifiers can provide. He or she is considering purchasing a SET amp and using it in conjunction with a speaker having a specified impedance of 8 ohms. JA's measurements reveal that the impedance has wide swings over the frequency range, including dips to low values at highly capacitive phase angles. Purchasing the SET amp would be an expensive and/or time consuming mistake.

2)Someone is considering adding a powered sub having only line-level inputs to his or her system, and driving it from a second set of output jacks that are provided on the preamp. As is commonly the case the two sets of output jacks are not separately buffered. The power amp is solid state and has an input impedance of 20K, the sub has an input impedance of 10K, and the preamp is tube-based and has a specified output impedance of 400 ohms. JA's measurements reveal that the output impedance rises to 3K at 20 Hz, which is not uncommon. Purchasing the sub would be an expensive and/or time consuming mistake.

3)Someone is considering purchasing a tube amp to use in conjunction with a speaker having a specified impedance of 6 ohms. JA's measurements reveal that the impedance is close to 4 ohms at low frequencies, and close to 8 ohms at high frequencies. That may work OK in some cases, but it definitely suggests the possibility that the speaker was designed with the expectation that it would be used with a solid state amp. The result stands a good chance of being an expensive and/or time consuming mistake.

Countless other comparable examples could be cited.

Disclaimer: I have no affiliation with JA, Stereophile, or anyone else associated with the audio industry.

Regards,
-- Al