Friday, December 23, 2011

Rethinking Oppo noise measurement

Arguing over at AVGuide, I've realized I guestimated the Oppo BDP-95 noise even more incorrectly than previously thought.

My actual measurements were 39dB for player and 37dB for background.  One guy at AVGuide suggests that means the Oppo produces 2dB noise.  That's wrong.

But I was thinking something wrong also.  I was thinking that the noise was dominated by the louder source.

It turns out the loudness is dominated by the larger source only when the larger source is WAY louder.  20dB difference is often used as a criterion for accurate measurement.

If the increase is exactly 3dB, it means the two randomly correlated sources are equal in loudness.  (If they were correlated, it would be 6dB.)

So if the increase is 2dB, it means the background noise is actually louder than the Oppo, but not by much.  I don't know the exact formula, but I guess the extra dB accounts for 2dB of level decrease compared to background.  So the guestimated number for the Oppo chassis loudness playing DVD-Audio disc (measured within 6 inches btw) is around 35dB.

Of course, the randomly correlated and continuous assumptions wrt the background noise can't be trusted so the best way to do measurements is have the background at least 20dB lower.  But at least my guestimate is now somewhat better.

No comments:

Post a Comment