demetri_j wrote: was a little disappointed by the following but I don't really understand the reasoning: ...
As for the 24-bit claim, the resolution is 24-bit, BUT signal/noise ratio is slightly over 100dB, effectively limiting the dynamic range to the equivalent of 17-bit. "
Due to thermal noise and at the signal levels commensurate with consumer (and pro) systems it is physically
impossible to build an ADC or DAC with true 24 bit resolution and dynamic range.
You can make a chip with a 24 bit architecture, meaning that all the parts are in place that would make it 24 bit in an ideal world, but it will always perform to a lower standard, again, due to that innate noise.
For instance I have a Tascam DV-RA1000 recorder. It sports a BB PCM1804 ADC, which was the best single-chip ADC available at the time. Its noise-based performance is around 18 bit.
I also have a Beis standalone ADC, with a next-generation chip and moderately-freaky analogue input stages, and it performs at 19 bit.
dynamic range to the equivalent of 17-bit. Due to dither, the psychoacoustic dynamic range will approach 19 bits.
This is wrong. Dither slightly reduces the dynamic range, but even this notion is moot since it is outright stupid (illegal?) to operate a digital audio channel without dither. What he means is dither combined with noise shaping, which can indeed bend the noise to lower levels in the midrange, compensating with higher levels in the treble. But this only with a source (ADC-sourced) that has better resolution/SNR to begin with, so it won't help an ADC that is thermally limited at 17 bit.
Such texts are par for the course in this world with internet-educated 'experts'.
Michell Orbe/SME IV/Lyra Delos/Michell Iso & Orca & Alecto Stereo/Quad ESL-63/Denon DVD-2930/Apogee MiniDAC/Pioneer N-50 streamer