Re: Digital Speakers
On Sat, Apr 19, 2003 at 01:17:14PM -0700, Deryk Barker wrote:
> Thus spake Paul Johnson (baloo@ursine.dyndns.org):
> >
> > On Fri, Apr 18, 2003 at 01:14:00PM -0400, Bijan Soleymani wrote:
> >
> > > Of course a professional solution would use a D/A converter seperate
> > > from the speakers.
> >
> > Isn't that the whole idea behind a sound card to begin with? Rip out
> > the DAC and you have a nice PCI to l-plug adapter. I can't imagine
> > this being cheaper, since it requires additional external hardware to
> > deal with that...it's like one step down from a winmodem. (Winmodem:
> > Sound card with the wrong outputs).
>
>
> No doubt it would cost more, but it will probably sound better. Most
> sound cards are very, very far from Hi Fi quality.
>
> I notice that HK also have a computer-to-stereo box which can take
> either raw digital bitstreams or MP3s and feed the analogue output to
> your stereo system.
>
> My HK CD recorder make far better sounding recordings than either of
> the sound cards I've used.
It's for this sort of reason that I've built a 24-bit outboard ADC that
feeds my sound card's S/PDIF input.
In this endeavour I have found that sound cards with digital (S/PDIF) i/o
are few and far between, and mostly overpriced. The only cheap ones are
those with the CMI8738 chip, like my Trust SoundExpert, which cost about 30
quid. There are other cards with the same chip that ONLY have the S/PDIF i/o, no
analogue i/o at all; they cost about 130 quid. The phrase "rip-off" springs
to mind.
The common Soundblaster is no use for this; the more expensive ones have
S/PDIF, but they resample everything to 48kHz, which is unhelpful where CDs
are involved.
Software support is patchy too; to get 24-bit input to work in Windoze
involves loading drivers from three different manufacturers on top of each
other; in Linux I've had to hack the kernel driver module, and use the
latest 2.4.21-pre kernel. Result: S/PDIF now works, but nothing else does...
more work needed!
"Digital speakers" where the D-A conversion is performed by the
sound-producing element have been experimented with. One approach uses a
"bitstream" signal to modulate a powerful IR source of a frequency which is
strongly absorbed by air. Presumably a correction is applied for the
variation of the speed of sound with temperature, to avoid severe second
harmonic distortion. Another approach involves an electrostatic speaker
element with a capacitor whose value can be digitally modulated by means of
gas discharge elements.
Pigeon
Reply to: