Watch for confusion in terminology here:
In older analog times, "signal to noise ratio" referred to the difference b=
etween the noise floor and the average program signal (0 on a real mechanic=
al VU meter with standardized ballistics). It was presumed that there would=
be additional range called "headroom" above the average program level for =
louder peaks. S/N + headroom was referred to as Dynamic Range.
In the digital world, the terms "signal to noise ratio" and "dynamic range"=
are often used interchangeably, with confusing results. When a noise spec =
is listed in a digital system, it's typically relating the noise floor to t=
he maximum peak value of 0 dBFS (aka "full scale"). Makes the numbers bigge=
r...makes the ad writers happier. There is no headroom above full scale (le=
aving aside floating point systems).
Average program level in a digital system is variously defined by different=
devices as anywhere from -20 dBFS to -12 dBFS. So if you choose -20 dBFS a=
s your target for average program level, then your -138 dB noise spec/dynam=
ic range translates into 118 dB of signal to noise ratio plus 20 dB of head=
room. Insert picture of glum ad writer here.
--Flawn
On Jul 22, 2012, at 1:23 AM, wrote:
> 1b. Re: artifact in spectrogram analysis
> Posted by: "sainganrob" sainganrob
> Date: Sat Jul 21, 2012 1:22 pm ((PDT))
>
>
>
> --- In "Avocet" <> wrote:
>>
>
>> I haven't looked at 24 bit recordings on my affordable Tascams, but I
>> suspect that the least bits are noisy. For a start, no analogue input
>> is likely to have a signal to noise ratio of 138dB if only because
>> that takes it beyond thermal noise.
>>
>
> Not quite, remember, signals go above reference levels as well as below :=
)
>
> A 150R, 20KHz BW, 20 degC source at a nominal 'Line Level' peak of +8dB r=
e (0.775) is 139dB SNR.
>
> A typical clip point for a Line Input stage is +20dB and so would equate =
to a 151dB SNR.
|