What's the difference between NOS DAC and SS DAC?

[hertz]

I've been visiting Pacific Valve website (Pacific Valve & Electric Company Home Page DACS) and it seems they put DACs into three categories: NOS DAC, tube DAC and SS DAC.

For example, the Giga Lab Moon Dac uses parallel TDA1543 design while the Lite DAC uses AD1853. I know NOS means non-oversampling but do DACs always mean to be doing that for fidelity since most CD are recorded as 16bit and 44.1kHz?

Which design is better? The NOS DAC or the SS DAC? Someone please enlighten me.

 

[paara]

As you said: a NOS dac only indicate that it does not upsamle or oversample the signal during the convertion of digital to analog signal.

Both a ss dac and a tube dac can be NOS.

Tube or SS only indicate what type of "amplifier" is used after the D-->A conversion.

If you have a good source and signal a NOS dac wil be a good choice, but if the signal is less then perfect a oversampling dac can be a better choice.


(please correct me if I am mistaken)

 

[hertz]

But many manufacturers boast 24bit 192kHz chips in their DACs. If the source is the typical 16bit 44.1kHz CD, I wonder how can oversampling benefit the output?

Quote:

Originally Posted by paara /img/forum/go_quote.gif As you said: a NOS dac only indicate that it does not upsamle or oversample the signal during the convertion of digital to analog signal. Both a ss dac and a tube dac can be NOS. Tube or SS only indicate what type of "amplifier" is used after the D-->A conversion. If you have a good source and signal a NOS dac wil be a good choice, but if the signal is less then perfect a oversampling dac can be a better choice. (please correct me if I am mistaken)

 

[JaZZ]

Quote:

Originally Posted by hertz /img/forum/go_quote.gif But many manufacturers boast 24bit 192kHz chips in their DACs. If the source is the typical 16bit 44.1kHz CD, I wonder how can oversampling benefit the output?

Oversampling or upsampling is just a tool for a rough «prefiltering»* of the signal during/after D/A conversion before passing the final analogue filter, allowing the latter to be simpler, thus enable a less complex signal path (in the ideal case), which can sonically pay off. The upsampling process itself, at least if done by an uninteger multiple of the original sampling rate, may introduce (euphonic) distortion, though, so it's not unanimously accepted as sonically beneficial. In any event, it doesn't add any usable signal content compared to NOS filtering concepts or designs renouncing low-pass filtering entirely (which are often equated with the term NOS).

[size=xx-small]* in that the interpolating algorithms mimick a low-pass filter at around 22 kHz[/size]
.