Technobabble jargon inbound! Please strap yourself in!
To store an analog thing (wave) digitally it is sampled and converted to a defined bitstream which represented the waveform of the analog. A wave, lookup a picture of a sine wave, holds data like amplitude, phase, period, etc. In audio, the goal is to sample the audible frequency of the wave and then reproduce that wave somewhere else; covert the digital back to analog.
There are two predominate encoding methods for analog waves.
Linear Pulse Coded Modulation (LPCM, or just commonly PCM)
Pulse Density Modulation (PDM, or the method DSD uses at 1 bit)
LPCM uses an absolute set of values to represent the amplitude of the wave; the bit-depth.
DSD, being 1 bit, only represents two possible values of the wave; high (1) or low (0). It does this at such a high resolution these transitions from high to low in the analog reconstruct allow for analog wave when run through a D/A convertor (after high pass flier). That’s the “density” in PDM. The bits are packed at a very high density.
This difference is debatable as how audibly different DSD vs. PCM is at equivalent resolutions. However, to most people, they do sound different and therefore one is often preferred over another.
A few of us debated this and went into details here:
Yes, because capturing the DSD on the SACD is the only way to preserve the PDM encoded audio. FLAC is a container for PCM. You need to store the DSD into something that preserves the DSD (like a DSF).
The other “problem”, as others have pointed out, is that some SACD’s (called hybrid-SACD) is a multilayer technology allowing for both a traditional CD (or “redbook”) layer as well as the SACD layer (containing the DSD). If you rip the CD layer you’re not capturing the SACD part (DSD) you’re only capturing the CD at 16/44.1 PCM. DSD also supports multiple channels so a multichannel SACD may have both a 2 channel and multichannel version of the album. You’d need to rip both of these if you want to capture them.