04-20-12: Onhwy61while it might be true that older ears do not have the 20-20K respone, the music is prepared for everyone. Like the article says there is a 100 yrs worth data that shows that 20-20K is the human hearing limit. So, when preparing digital music might as well keep the audio spectrum to its max limits. Younger people certainly can hear this range & so can many other older folks.
And what's wrong with mathematically perfect response out to 16kHz? Most people, and certainly most middle age and older audiophiles' hearing doesn't go beyond 16kHz.
FM doesn't go beyond 15kHz and at its best it's pretty damn good sounding.FM has (air) spectrum bandwidth limitations that force it to curtail bandwidth. If they could help it, they would have also transmitted in the 20-20K range. Air spectrum is very expensive so this compromise seems reasonable.
FM doesn't go beyond 15kHz and at its best it's pretty damn good sounding.Like I wrote in my prev post & I'll write it again: if you start off w/ 12-b you'll end up with 9-10 bits after the mixing & mastering processes. If you start off w/ 16-b, you'll probably end up w/ 12-13 bits. The section "The dynamic range of 16 bits" explains quite well the DR of 16 bits & also how it might be possible to encode fainter signals using 16-b.
Since a lot of data already shows that sounds at absolute levels of +120dB, +130dB permanently damage ears, my understanding is that it might not be worth encoding sounds on a disk that cover the enitre 140dB dynamic range of human hearing. It appears that covering 120dB of dynamic range is sufficient. If one uses 12-b only & one attempts to encode very faint sounds my understanding is that 72dB could be a limiting factor trying to cover the entire 120 DR. 16-b & 96dB is adequate & the article shows a plot of a -105dB signal at 1KHz using clever dithering techniques.
Nyquist criteria applies only to continuous waves.nonsense! The Nyquist criteria applies to any signal that needs to be quantized. The Nyquist criteria only gives the minimum requirement; it does not say that one is forced to have only 2 samples per highest frequency.
There is no filter in the world, digital or analog, that can do that (no matter how many poles) with even group delays (or linear phase if you prefer).yeah, I know what you mean for analog filters & I agree w/ you in that respect but for digital FIR filters (linear phase) I'm not sure I totally agree with you. My understanding is that if you had a, say, 64-tap FIR you could have a very steep skirt digital filter that would have flat group delay & group delay distortion. I would have find some evidence of this before I contend this issue w/ you but for right now I'm skeptical that it cannot be done. I'll leave it that....
I see the case for upto 24/96 as it seems to alleviate most of the pressing issues such as noise creeping into the music signal during mixing/mastering, analog filters having too steep a skirt at 44.1KHz. I'm not sure that I buy the case for 24/192, etc.
If anyone is interested in looking at some signals look at the Powerpoint presentation at reference #17 in the article. SLides 20, 21, 24-28 show spectrum of instruments & spectra of music from commercial CDs. Look at the freq where the content dies off even for SACDs.