[LAD] [ANN] Invada Studio LV2 Plugins 1.2.0

[Fons Adriaensen]

On Mon, Aug 24, 2009 at 01:38:37AM +1000, Fraser wrote:

> > What's the point of using a meter if you adjust it to the signal ?
> 
> You are not changing the signal, you are changing the amount of headroom the VU
> meter has. http://www.reference.com/browse/wiki/Headroom

Yes, I know that. Still same question: what's the point of
using a meter if you adjust it to the signal rather than 
adjusting the signal until the meter says it's OK ?
Of course a slow meter must be calibrated to show 0 well
below full scale digital. But that should be fixed.

> > Anyway the meter plugin freezes my machine if the signal is muted
> > or removed. Probably due to denormals.
> 
> I can't reproduce that.
> A clue may help. (distro, arch, kernel, soundcard, host etc).

I added some simple code to avoid denormals in the
bandpass filter and in the 'envelope' calculation,
and that removes the problem (just adding a small
DC offset, 1e-20).

> > The VU is not a VU,
> 
> Correct since this plugin isn't in the analogue domain.

So what ? If you call it VU and it looks like one, it
should behave as a VU and have VU ballistics. This one
surely doesn't.

> > the spectrum doesn't use valid 1/3 octave filters by any standard, 
> 
> please advise where this standard is.

For example IEC61260, or the BS/EN with the same number.

> > and the phase meter doesn't indicate anything useful.
>
> Apart from the relative phase difference between the left and right channels.

It measures an average of atan ((l-r)/(l+r))), which maybe
related to phase for a sine wave with equal amplitude in
both channels, but there it ends. Panning a sine wave L to R 
produces all values between 0 and 45. If it measures phase
that should be 0. 

Anyway phase is not the thing to measure, correlation
is. And this is certainly not a correlation meter, just
compare it with a real one.

Ciao,

-- 
FA

Io lo dico sempre: l'Italia è troppo stretta e lunga.