@cloudforest: I never claimed greater "fidelity" for vinyl across the frequency range - I pointed out that vinyl had a greater frequency range than CD, which is true. You seemed to introduce various reasonings about high frequencies included (or not) in vinyl recordings that can either be mitigated against or don't really exist anyway. Spectral analysis between vinyl and CD output shows the greater frequency range that records can bring to a reproduction. Vinyl frequency reproduction is also reasonably flat to 30KHz anyway and experimental pressings have been made up to 120KHz.
Your point about mass consumption is lost on me since we are talking about a niche market for audio here and the possibilities (as Peter Walker from Quad opined) for the "closest approach to the original sound". Design philosophies that can easily and faithfully reproduce all original frequencies interest me more than design philosophies that would sharply filter out frequencies above the old "red book" standard because the designs themselves would introduce their own additional harmonic distortion. "Mass consumption" means "mp3", which is objectively lower quality than CD! Your point about the "majority" of playback systems only underlines my earlier points about the exacting nature of vinyl (or any faithful sonic) reproduction.
Recent developments in audio with SACD and higher bit-rates and sampling frequencies becoming standard for audio recording and mastering indicate that the old "red book" standard is not sufficient and that better quality (with high end systems) can be attained with improved technologies. These improvements would not have become the new standard for digital if 16 bit 41KHz was considered sufficient for "perfect" audio.
DAC designs have improved, but oversampling techniques have also been used in the early CD players to maximise the sample rate so that frequencies approaching the Nyquist limit have sufficient resolution and have reduced artifacts for brick wall filtering. The additional headroom of higher sample rates makes it easier for filters to work above the audio range and reduce phase and aliasing errors. Although perfect on a mathematical level, the actual waveform produced near the nyquist limit has always needed additional processing for the human ear.