Propeller Head Plaza

Posted by Soundmind (A) on June 01, 2006 at 15:06:10

In Reply to: I agree that such data exists or can be established, but that is one of several required steps. posted by rditmars on June 1, 2006 at 14:37:16:

Suppose you had to design a lens/ccd to obtain the sharpest possible image of a known scene at a given distance. Suppose that you determined that 98% of people could not distinguish two separate lines if they were closer than 1 mm. Then you might design your system to say a resolution of 1/2 mm. You might test it at 3/4 mm to verify that your design is correct. If you prove that you can take photographs this sharp with it and it therefore exceeds the human eye's ability to resolve, it would be pointless to design the system to a resolution of say 1/4 mm because only 2% or or fewer could perceive the added sharpness and it might cost 10 times as much.

Supppose you had to design a digital circuit containing an A/D and D/A converter which would reproduce music. How much resolution meaning individual loudness levels would you need? Suppose you determined through testing that fewer than 1/10% of the population including a famous editor who makes recordings could hear sounds different in volume of less than 1/10 db. You know that the quietest room for performing music is 27 dba (AIA standard for concert halls) and the loudest sound at the threshold of pain is 120 db. That means you need a range of 93 db and 930 individual levels to encompass all possible loudness levels as a minimum design criteria. If you designed a system which could logarithmically evenly space the levels of a 14 digit binary code signal over say a 100 db range, you would have 2 to the 14 loudness levels or 16,000 different loudness levels within a 100 db range, far better than you would ever likely need.

That's how I would approach it.

As for comparing two cameras or two amplifiers, that is not necessarily a good test because it does not use an external standard, it merely judges one item which may or may not meet the required design criteria with another which may or may not. This is why judging two pieces of audio equipment against each other is not necessarily a good idea. Far better would be to judge them individually as elements of a system against live music to see how close they come. Isn't this after all their real ultimate test? It takes the element of preference out of the equation. This is also a difficult test which needs to be carefully thought out and conducted. Few manufacturers or hobbists even try it.