Its been my experience that no digital plug-in can really match the sound of something like analog tape or a particular tube amp or guitar distortion or classic analog compressor any closer than "sort of". The variables are just far, far too complex in the real world to model all the possible scenarios. For instance with tape; what kind of tape (they all act differently to some extent, and they don't all sound the same), what bias level is being simulated as being used (or abused) for that particular tape? what tape speed? how is the recording and playback EQ set for the "simulated" tape machine (alignment can change everything) is the tape machine half-track? full-track? quarter track? cassette? what type of magnetic heads (their faces, their impedances) are being simulated-again this can change the sound character, especially at low frequencies. And what type of electronics are following those heads? tube? transistor? ICs? What about a tape machine which doesn't have perfectly clean speed, i.e. wow and flutter? How about scrape flutter? Is modulation noise which happens with real tape addressed (never in my experience)?
Really, just about anything that exists in the real world analog realm cannot be -exactly- simulated with the present state of the art in computing, especially a plug-in which is limited by its host computer and sound card DSP having to share resources with all the other processes taking place in a DAW.
For some of my uses, plug-ins might get "close enough" for something buried in an overall mix, and depending on the music, but unfortunately, the real world is just too complex to model exactly (real-world concert hall reverb is another example that I've never heard being exactly duplicated, especially if for instance a violin soloist is moving somewhat, thus projecting the directional sound from the instrument in subtle ways in different directions - artificial reverb would totally miss this interaction with the environment and real air).
