In fact, I think what Rachmaninoff's recordings show—what the advent of recorded music shows—is that our perception of a musical past is, in many ways, an illusion. In general terms—
The perceived difference between the musical past and the musical present is a symptom of the limitations of information technology.As information technology improves, the distance between past and present shrinks. You can make a taxonomy of musical examples. Nearly the entire corpus of surviving Ancient Greek music fits on a single web page, with tempo and tuning largely educated guesses. Chant and medieval music exists in more complete sources, along with more detailed instructions. Western classical music is documented in standardized notation and fuller contemporary accounts. Music written since 1900 has increasingly—an with increasing fidelity—been recorded by or, at least, under the guidance of the composer.
That brief rundown might be considered coincidental—the steady progress of information technology, directional over time, corresponding to how "old" a musical epoch seems to us. But much musical scholarship of the past century has been devoted to, in a way, improving the backwards compatibility of more rudimentary information technology. If we know more about how Renaissance music, or Medieval music, or Ancient Greek music is supposed to sound, it's because scholarship has filled in gaps that make performers better able to assemble the existing instructions—notation and treatise—into realized music that exists in the present.
In other words, all music—regardless of age—is ideally immediate and timeless, and as the technology of reproducing music—regardless of age—improves, the "pastness" of music falls away.
I already see the two implicational poles of this with the digital reproduction and distribution of music—old music ceases to gain an advantage from the imprimatur of age, but also ceases to suffer in comparison with music of greater chronological novelty. (This is, I think, why concerted efforts to bring classical music into alignment with pop music, to make it "appeal to new generations," always come off as tinny to my ear: they're trying to solve a "problem" that's fading away on its own, and, ironically, dating themselves—the technology advances faster than the solution.) As the reproduction converges on the fidelity and immediacy of live performance, and as those reproductions become ever-more immediately available, the "dead white guy" factor of classical music diminishes. It's like Faulkner famously said: "The past is never dead. It's not even past."
Here's another card in this file, one I will almost certainly be returning to in the next few months. Brian Wilson finished his unfinished masterpiece, the aborted "SMiLE" album, in 2004, ending over thirty years of fan speculation via bootlegs and homegrown reconstructions. Nevertheless, additional bootlegs and reconstructions of "SMiLE" have continued to surface since 2004. Why? Because of information technology—the raw materials of the original "SMiLE" sessions, once released into the digital world, have proliferated such that anyone with Internet access has near-instant access to them. Fragments of music originally thought lost have made their own way, to the point where they're constantly re-introduced into the present, refusing to be pinned down to 1967 or even 2004. Mahler's piano rolls, Rachmaninoff's 78s, studio tapes from the 1960s—they're all here and now. And, as the present continually renews itself, only more so.