Quote:
Originally Posted by Fabian
...Is this a valid test?
|
Well, it does prove that there is an LED, and that it lights up.
Quote:
|
Is it meaningful? What does it tell us?
|
It tells us that the manufacturer has provided, as a courtesy, a little red bulb that lights up to warn us that we may be pushing the design limits of the system. It's like "redline" on a car tachometer-- it's not really a scientific measure of anything. The car doesn't blow up if you go over the redline, nor does instantly and constantly accelerating to one RPM below the redline necessarily prolong its life beyond ordinary sensible driving that might occasionally include "redlining" the engine.
In short, if we have a sort of autistic need for everything to indicate a precise boundary and meaning, then no. The light doesn't have any meaning. If, on the other hand, we are better-equipped to make use of vague and "messy" information, then it could be very useful visual cue to help alert you when the singer is succumbing to "volume creep".
Quote:
|
Does it mean that 0dBu is around -14dBFS?
|
I don't think that has been conclusively proven by your test. There are a lot of ways that a "clip" LED might work, and they don't all relate precisely to individual sample values-- suppose the clip LED takes 1ms to respond, which is a perfectly reasonable time to detect audible analog clipping: that means that, at an 88.1k sample rate, 88 samples will have gone through the converter before the clip LED detects "analog" clipping.
So if you have an OCD need to know exactly what everything "means", then you're going to need much better test equipment. Otherwise, you could just record lower levels and not worry so much about what it "means."