[LAU] nice piece of USB hardware for 7.1 output

[Fons Adriaensen]

On Thu, Jun 06, 2013 at 01:32:08PM +0200, Jörn Nettingsmeier wrote:
 
> gigaport hd+ connected to a micstasy line in at normal Z (5.6kOhm
> unbalanced) with a crappy rca to jack cable.
> i chose 21.0dB amplification, because that should correspond to 0dBu
> at full scale. indeed, it makes a -1.0dBFS sine wave from the
> gigaport come out at roughly -14dBFS on the hdspm, which is what i'd
> expect going from consumer (-10dBV) to pro equipment (+4dBu).
> 
> jnoisemeter reports -88.4 dB in flat mode with DC filter and slow
> averaging response.

So dynamic range should be 88.4 - 13 or around 75 dB.
No really impressive...
 
> now if i feed it a white noise signal from JAAA at -1.0dBFS, the
> other end reports a flat noise level of -13.1 dB, which as a rough
> guesstimate gives us about 70dB of usable dynamic range. not great,
> but ok.

That would be a really invalid measurement. Peak / RMS ratio for
jaaa's noise generator is something like 12dB, so noise at -1 dB
RMS will be clipping as hell. That 12 dB is an implementation
limit, real gaussian noise has (in theory) an infinite peak / RMS 
ratio. So better stay at -20 dB RMS or so.

> the -1.0dbFS sine tone at 1kHz shows a first harmonic at -95dB, and
> the upper harmonics barely show in the noise, so we're fine here.

OK.
 
> however, there is a thumping DC impulse every few seconds which
> clearly shows up on the scope. it's clearly audible, not a show
> stopper, but needs looking into. it's signal-dependent: when you
> don't generate a sine-wave, it stops. complex signals either don't
> cause it or mask it completely. might be a power issue.

Very suspect. Should be interesting to record the waveform and
have a look at it (zita-scope would be handy, but not yet ready
for the real world).

> a white noise steady-state spectrum shows a slight raise of 2.5dB
> around 6khz, and another hefty boost of 3.5dB around 14khz.

Again suspect, but see above. Could you try with a lower noise
level ?

General remark about the Jaaa screenshots: to measure noise you
should really use the noise, not peak, markers.

Ciao,
 
-- 
FA

A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)