> OK, serious question: I need to implement VU-style indicator for music
> playback device. It's just for decoration (feeping creaturism), so
> accuracy is not terribly important, but obvious, maybe exaggerated,
> bouncing LED is.

> I was planning to sample about 8Khz and run thru simple IIR filter,
the
> output works the bargraph. But the RMS thread has gotten me thinking:
> would this really do what I need? What exactly the heck is a Volume
Unit
> anyway? Is there an algorithm that makes more sense?

Volume Unit?  Nah, in my sound engineering days we called them
Vague'n'Useless.

It's *supposed* to be a scale that approximates to the non-linear
response of the ear, in other words to indicate subjective loudness.
In practice they're usually just some cheap junk with a cream dial stuck
on them.

A VU meter gives a slow, average indication of signal level so it will
be fairly accurate on an unchanging sine wave input but with real-world
audio signals will usually under-indicate the real peak signal level by
anything from from 6 to 20dB!

How about a display that mimics the broadcast-style PPM (Peak Programme
Meter), i.e. with a fast response on a rising signal or transient and a
slower decay as the signal falls. Looks nice and actually gives some
useful information about signal peaks.

Another neat looking trick (this is mainly for the look right?) is to
have the highest lit LED on the bargraph stay on for a half-second or so
after the signal drops - in other words on a rising signal it looks like
a normal bargraph but on a falling signal one LED stays lit indicating
the previous peak level.  Both of these look quite technical and
impressive.

If you don't want to code a software solution you can buy ready-made
bargraphs
with onboard driver that give exactly these kind of responses to
analogue audio inputs.

















.