[time-nuts] Oh dear

Magnus Danielson magnus at rubidium.dyndns.org
Mon May 7 18:50:41 UTC 2012


On 05/07/2012 08:15 PM, Poul-Henning Kamp wrote:
> In message<4FA80913.7000703 at medesign.ro>, MailLists writes:
>
>> That was a big problem with the dynamic range of tape recorders, which
>> had to be solved with noise reduction circuits. Even good 16 bit ADCs
>> have a higher DR than the SNR of most instruments in quiet recording
>> studios.
>
> Not so fast there...
>
> Yes, in theory your ADC could digitize a signal 14*6 = 84 dB below
> reference level, but it would do so with 50% distortion, because
> there would only be three distinct levels: {-1, 0, +1}
>
> This is a much overlooked issue, in particular with classical music
> where dynamics in the music can account for way more dB than people
> realize.
>
> We must start out by defining the acceptable level of total distortion,
> if we choose 0.5% then we need 200 digital levels, roughly 8 of
> your 16 bits for the signal.
>
> That gives you a headroom of 7 bits (leaving one for the sign) and
> that gives you 42 dB of S/N.
>
> That isn't very much, headroom, 42dB, when the conductor waves the
> entire philharmonic AND the full opera choir in, for for that wonderful
> "Dies Ira" of Verdis.  Or Carmina Burana.  Or any of the many
> other 'shock-effects' classical composers have enjoyed.
>
> With digital, you get most distortion at weak signals, where your
> ears are much better at detecting it, with vinyl you get more
> distortion on strong signals, just like your ears, meaning the
> level becomes unbearable sooner.
>
> That is why, in plain and simple terms, classical struggles with
> digital:  High distortion in weak passages.
>
> It is also why the CD media has changed rythmic music, which
> went from a love of distortion to a love of pure tones when
> the CD media made it possible to play loud pure tones.
>

The late Julian Dunn has covered this in AES papers and pre-prints.

It relates to side-band, modulation frequency and masking-effects. He 
came up with a sinusoidal modulation mask.

Look up his work!

Cheers,
Magnus



More information about the time-nuts mailing list