A perspective on audio evaluation, measurement

In the realm of audio, as any astute observer will know, there is an intense battle between the subjectivists and the objectivists. This battle, like all religious battles, will never die. Any argument about whether a particular kind of audio distortion is perceptible to the human ear will never be settled. But wait a minute. Is it about proving who is right or simply picking the part of your audio system in which you should invest your next hard-earned dollar (or euro or rupee)?

As an audio engineer who enjoys his music, I think there is every reason to take a sensible, return-on-investment approach to the benefits gained by putting a pile of money into improving one aspect or another of the audio chain. For example, would you invest $1,000 in a better DAC to improve jitter that is already at 20ps if your system's SNR would remain no better than 100dB? To get a scientific answer to this, all we need is a small, well-established formula:

SNR = 20 log(1/2πfT)

where f is the frequency of interest, and T is the jitter time. A quick calculation will confirm that you gain nothing—20ps of jitter at 20kHz corresponds to 110dB of SNR, with the figure being higher at lower frequencies. That tells you the improvement is difficult to justify.

Let's look at another way to spend your money. How about putting it into a 35bit DAC when the SNR is at 125dB? Time to fish out another simple, well-established formula for the effective number of bits (ENOB) in a digital audio system:

ENOB = (SNR-1.76)/6.02

This tells us that a 125dB SNR corresponds to an effective use of barely 20 bits. Even to get to true 24bit processing, the system must have an SNR of better than 144dB! Sorry—you can't justify the investment in a 35bit DAC.

OK, so will you invest that money in a better input stage to reduce SNR to 130dB in a living room setup? Now you need to wonder what to do with the sound of your own breathing, which is typically 10 dB-SPL as far as three meters away. Above all, will you spend even one extra dollar upgrading an opamp to improve THD to 0.0003 per cent when you know the weakest link in the system—your speaker drivers—are at best capable of a THD figure of not much better than 1 per cent?

If you've been with me so far, I think you'll understand the true role of measurement in audio system evaluation. The argument is no longer about whether measurements can capture all that is perceivable by the golden ear. The simple truth is that there are lots of deficiencies in any high-end audio system that can be measured with decent equipment—never mind arguments about those unmeasurable deficiencies the golden ears claim they can detect.

It would be far more worthwhile to invest your hard-earned money in measuring and improving measurable deficiencies and invest what you have left in buying some good music than spending lots of money on improving so-called unmeasurable deficiencies, even if you could hear the difference. (And in all likelihood, you cannot in the presence of measurable deficiencies anyway.)

What are your experiences in the world of audio?

About the author
Ram Ramaswamy earned his M.Sc in Physics from Delhi University in 1990 and his PhD in Operations Research from the Indian Institute of Management Calcutta in 1994. He built his first audio amp at the age of 12 using the popular TBA810 which, much to his delight then, instantly became a local MW receiver too. While as a pragmatist he believes that digital technologies have replaced analogue largely for the better, he rues the fact that some of the best principles of analogue design, such as the principle of minimalism which it implicitly embodies, are slowly dying away.