|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
80.66.130.220
In Reply to: Re: er... posted by Aleksunder on September 15, 2006 at 04:19:28:
No, this is about band-limiting, thus prior to the sampling ADC.
In practice the filter I propose (really not me, but Craven, Lagadec, et al.) has to be inserted in the digital domain, after an oversampling ADC. I have Matlab prototypes that take in a normally-sampled 88.2kHz stream and convert it to 44.1kHz for CD mastering.
Follow Ups:
Ok, that sounds intersting.However, I still think that it ultimately boils down to simple lack of 'time-domain resolution' (which isn't the same thing as band-width or frequency response) and that this is intractable.
IMV, a 44.1Hz sample rate doesn't actually 'resolve' above about 8KHz (if that), whatever the various threorums say about 'reconstructing' frequencies above this with anti-aliasing et al.
Anyway, in the D/A process (which I would maintain is where the real conumdrum lies) you mentioned 'ringing' on signal edges with complex over-sampling filters, there is also the issue of phase-shifting of harmonic content (loss of 'time domain' coherence). Both of these profoundly effect perceived timbre, despite measuring as tiny numbers as 'THD'.
"time-domain resolution' (which isn't the same thing as band-width or frequency response) "It is exactly the same thing.
Time and frequency are two sides of one coin.
Time and frequency are two sides of one coin.I agree - two factors. A musical instrument can produce an immensely complex sound and a commensurately complex waveform, which is composed of many simultaneous harmonic components. Tiny changes, almost at the threshold of 'measurability', in the relationship of these harmonics can (and do, in 16/44) change timbre quite radically.
This post is made possible by the generous support of people like you and our sponsors: