esl 63
Member
You can measure most! Another question what is most relevant and what is our brain more or less sensitive too?So then what is the magic that we can't measure but hear?
One example is the SINAD that is a complete nonsense in some respect, since it is soo far away from scientific if we look on perception psychology.
Let me explain.... Our ear is a mechanical device, it is nonlinear in its construction, and it is also changing its own distortion generator with amplitude.
Non of those parameters is taken into consideration in a SINAD measurement. Actually there is a third thing called masking effect that is alsomforgotten, and some other stuff but i will explain the first two. Fifth harmonic distortion generation is non existing in our ear and perception wise we are highly sensitive to those sounds. But not only sensitive... it makes us feel uncomfortable. If you take your cutlery, a fork as an example and scratch your dining plate you generate the most annoying sounds. In comparison you have a metal rod that generates even harmonics and it sounds like an instrument, a pleasant rich sound. Now we compare this with 2nd harmonic distortion that is actually generated by your ear with up to 10% at lower frequency and high amplitude! You are used to it and you didn't know it maybe until right at this moment! Can you compare 2nd harmonic with 5th harmonic like apple and apple? No. You can not. Not if you are interested in sound reproduction and hearing and listening experience. And if you hang around here you maybe are?? What THD figure tells you is a sum of all harmonics with NO scientific perception psychology taken in consideration. So when you see an FFT diagram you can or should neglect that second harmonic unless it is in a range of -30dB to -40dB. On the other hand if you spot odd harmonics 5th 7th 9th etc (the higher the worst) then we know that it is much more perceivable AND irritating. Then we have the noise and masking effect. You are in a restaurant or a close to a disturbing apparatus that generates with lots of noise, if you concentrate you can decide to follow a conversation even though the voice is well below the noise floor! Thats why you can not compare digital S/N figures with analogue S/N figures. A tape recorder or an LP will give you information well below the measured noise floor! But in the digital domain it is just cut off completely!
I read that we can hear information >12dB below noise floor. So the DSP you have in your brain has some attributes that is not considered at all in a SINAD measurement. So use your eyes to look at that FFT measurement and you can judge a little bit better than the "number". Another interesting thing that we can mask higher harmonics with lower harmonics. But that you can read for yourself in MP3 compression techniques.