@blindjim
"BTW… what’s happening to those other three words from 24 to 21, if you don’t mind mentioning it?"
-----> The 21 bit resolution in the analog is an analogy. It means that the noise floor on the analog circuits sits at the equivalent level of the signal level of the 22nd LSB or approx -160 dB. I think we can hear roughly 15 dB below the noise floor for musical tones especially in the mid range so even some stuff below the noise floor might be audible. The point is that 24 bits is well worth it but any higher than that (32 or 64) really only confer advantages to complex signal processing (heavy digital filtering or complex deconvolution like in room mode DSP - often described as the number of taps or coefficients in the filter - a large number being more complex and computationally intensive but more accurate such as having less pass band ripple)
FWIW I have played around with room DSP and deconvolution filtering and I think it can improve things below 100Hz. However, even fairly complex 32 bit deconvolution like Audyssey XT32 creates audible artifacts in the mid range which to my ears are detrimental and degrading to two channel music - although for HT 7.1 and higher the benefits may outweigh the disadvantages (as speaker consistency and integration becomes more important as you increase the number channels)
"BTW… what’s happening to those other three words from 24 to 21, if you don’t mind mentioning it?"
-----> The 21 bit resolution in the analog is an analogy. It means that the noise floor on the analog circuits sits at the equivalent level of the signal level of the 22nd LSB or approx -160 dB. I think we can hear roughly 15 dB below the noise floor for musical tones especially in the mid range so even some stuff below the noise floor might be audible. The point is that 24 bits is well worth it but any higher than that (32 or 64) really only confer advantages to complex signal processing (heavy digital filtering or complex deconvolution like in room mode DSP - often described as the number of taps or coefficients in the filter - a large number being more complex and computationally intensive but more accurate such as having less pass band ripple)
FWIW I have played around with room DSP and deconvolution filtering and I think it can improve things below 100Hz. However, even fairly complex 32 bit deconvolution like Audyssey XT32 creates audible artifacts in the mid range which to my ears are detrimental and degrading to two channel music - although for HT 7.1 and higher the benefits may outweigh the disadvantages (as speaker consistency and integration becomes more important as you increase the number channels)