Correlation between technical data and sound quality

Posted by: Arye_Gur on 01 May 2002

We talked about this subject several times at the past, and few of us (including me) thought that because of the nature of our hearing system, we can hear things in a way that can’t be measured.

I remember that Ross said that he read an interview with the first engineer of Exposure and he told that they built the amps by the physical and electrical principles.

I think that Naim are doing the same – and all the other manufacturers too.

So how it comes that there is no correlation between the data sheet and the quality of the sound? If the manufacturers are following the technical data, how it comes that a system with better data sounds worse than a system with less quality technical data.

Now I think that I know the answer -

First, there is a correlation between technical data and sound quality.

But the question is WHICH DATA – and that is the big secret!
The technical data that published by the manufacturers is only a part of what the manufacturer is taking care of. This part is published because of historical reasons that were common in the Hi Fi papers many years ago, and it is common nowadays too, but it says nothing.

Each manufacturer has a batch of data he thinks is the most important to achieve the right goal with it. Each manufacturer works hard to bring the data he thinks is the right data to high and accurate level. And as each manufacturer emphasizes what he thinks is the important data – it makes the differences between the various components.

So it is not that we can hear unmeasured phenomenon – we simply don’t know
what are the technical points that the manufacturer is taking care of.
And it doesn’t change the main point – we can’t trust the technical data sheet,
We must trust our ears!

Arye
Posted on: 01 May 2002 by Dr. Exotica
quote:
Each manufacturer has a batch of data he thinks is the most important to achieve the right goal with it. Each manufacturer works hard to bring the data he thinks is the right data to high and accurate level. And as each manufacturer emphasizes what he thinks is the important data – it makes the differences between the various components.

Ayre raises an interesting question/point here. Are there quantifiable performance observations that manufacturers make which they focus their design energies upon. It may be something a simple as tuning the output waveform on an oscilliscope when probing an output transistor with a particular input signal. Or perhaps it is minimizing propagation delay (or signal attenuation) in some critical piece of hardware. It is undoubtedly all of this plus a whole lot more. Who knows (I am obviously talking out of my hat here not being a EE)...

Does anyone have any additional insight here (e.g., some of our notable DIYers)? It is an interesting question.

Erik
Posted on: 01 May 2002 by Martin Payne
quote:
Originally posted by Arye_Gur:
But the question is WHICH DATA – and that is the big secret!



Arye,

yes, it does appear to be a commercial secret, and quite rightly, too!

JV used to call them the 'Naim numbers', but he never actually shared them.

cheers, Martin
Posted on: 01 May 2002 by JeremyB
Arye,

I think you hit the nail on the head.

I built a power supply once (affectionately called "the biscuit tin" because that's what it looked like (actually die cast aluminium in case you wondered). I made the best effort possible and of course the circuit topology is famously simple. It sounded way better than the built-in amp power supply so I was pleased. Having read somewhere that every DIY supply Naim had tested was inferior to a Hicap, I did a quick check with an AC voltmeter, which flickered between 0 and 1mV (the minimum resolution) although it should be in 100s of uV of course which I couldn't measure, still the result was certainly acceptable given the measurement method.

When I finally got around to buying a real Hicap, I was not really surprised when it sounded even better than the biscuit tin. Sure enough, when I used the same voltmeter the reading was solidly at 0 even under load! Of course, I was curious why this was and could make some guesses but of course would never reveal them!

This experience greatly increases ones confidence in Naim's integrity, making the products incredible value for money IMO (sort of priceless).

Jeremy
Posted on: 02 May 2002 by bam
Ayre,
I agree with all you conclude. Another justification is simplty that all electronic parts, even with the same nominal values and batch codes, vary in their parameters. Semiconductors can vary enormously - 50% or more. In order to make thousands of NAP250s, for instance, with all of them achieving a very similar sound quality (and similar sound quality between L and R channels) requires that many parts are individually screened and selected. The latter requires specific measurment criteria and limits. These are the secrets of the manufacturing process and are guarded closely by the mfrs.
BAM
Posted on: 04 May 2002 by graphoman
We all know several good companies that build their products on a basis what we call electro-acoustics. Their equipments, however, have allways sounded VERY different. It was a common sense explanation that none of their products were perfect.

During the decades a lot of these companies have learned how to build better and better products. This process might have resulted in equipments more similar than ever. Though as we all know the auditive difference did not decrease but rather increase instead. The characteristic sound pictures of the good audio companies’ products are more pronounced and more different from each other than ever.

Therefore, where is the “correlation between the data sheet and the quality of the sound?”

graphoman