[time-nuts] from Sputnik to CD
phk at phk.freebsd.dk
Sat Oct 6 12:48:02 EDT 2007
I have paid my dues in the HiFi world, but I have always insisted that
there be a solid scientific basis for any product I bought.
When faced with claims that picosecond jitter can be perceived by
the human ear, my flim-flam-o-meter blows a fuse.
A 1 nsec phasemodulation of a 44100 kHz clock signal will, worst
case, come out to an noise component 87 dB below the intended signal:
1/44100 [Hz] = 22.675 [uS]
22.675 [uS] / 1 [nS] = 22675 [ratio]
20 log(22675) = 87 dB
And that is even assuming that the two relevant samples have
opposite signs, an impossible situation according to the Nyquist
I find such hifi-nut claims particularly ridiculous, considering
that every single loudspeaker on the market is designed in defiance
of how electromagnetism works.
All amplifiers and loudspeakers on the market are designed assuming
that the task is to hold the voltage across the speaker terminals
proportional to the desired sound signal, despite the fact that
the magnetic field induced in the speakers coil is proportional
to the current, not the voltage.
Who in their right mind will claim that they can hear a distortion
of -87dB, on speakers designed contrary to Maxwells equations ?
And do I need to mention that very few rooms sport a S/N better
than about 50 dB these days, and that requires you to turn the music
up so loud that you are waaaay out of your ears most sensitive
I'm totally with Randi.org on this one.
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
More information about the time-nuts