These types of threads always get me a little riled up because of the amount of misinformation and pure bulls**t that people spew. So that you know where I'm coming from, I'm a full-time recording engineer who uses top-flight digital and analog gear on a daily basis.
To begin with, the difference between a 16-bit, 44.1kHz recording and a 24-bit, 96kHz recording is pretty subtle. Multi-track recording to 24-bit is extremely important because it results in a significantly easier mixing and mastering process and a drastically reduced noise floor, but the difference between a 24-bit stereo file and a 16-bit stereo file is very slight. The S/N ratio of most recording and reproduction gear (including most DACs and especially microphones) barely exceeds that of a redbook CD anyhow, and I challenge anyone to find any recording with more than 96dB of dynamic range (which would be a recording who's loudest passages have an amplitude 2,000,000,000 times greater than the softest). As for higher sample rates, the audible advantage to getting the sample rate above 44.1K is getting the filter (a low-pass filter is involved with all A/D conversion) out of the audible frequency range. Most converters (especially in the pro world) oversample and get the filter out of the audible range anyhow.
The main differences that you are hearing in your SACDs and DVD-As is in the mixing and mastering. Most redbook CDs are compressed to hell (just import a track from a CD into any audio editing program and look at the waveform), meaning limited (read no) dynamic range, not compressed as in MP3s, especially compared to the old analog releases. This is because people expect to put a CD in their car or stereo and have it be as loud as the rest of their CDs. It's also an attempt at having the loudest track on the radio. Most SACDs and DVD-As are mixed and/or mastered with audiophiles in mind, meaning enhanced dynamic range, and a more natural presentation.
As for the poster who said that extreme equalization was needed to make digital recordings sound natural, you actually have it backwards. RIAA equalization was already mentioned as it pertains to LPs, but you might be interested to know that significant equalization is also applied to multitrack analog tape to even out the frequency response. Digital requires no such EQ, and is usually as perfect and natural a representation of the original event as is possible.
Even most engineers that prefer analog tape as a recording medium will admit that the aspects that they like about tape are tape saturation (resulting in natural compression as the tape is driven with a hot signal) and harmonic distortion, two things which make the recording LESS natural.
In short, there's really nothing wrong with redbook as a medium. SACD offers some improvement through DSD, and DVD-A offers slight improvements through higher bit-depths and sample rates (although they are very subtle), but incredible sound is possible via redbook. The problem with most bad sounding recordings is in the mastering (due to *gasp* PUBLIC DEMAND), and somewhat in mixing. Part of what many of you consider the problem to be with most commercial recordings is that realistic and natural reproduction of an acoustic event is NOT the typical goal.
These are just the opinions of someone that works with analog and digital audio of all types all day, every day, and who produces CDs for a living.