From the very opening bars you know you are in for something special: spacious, ample, long-breathed, no sense of rushing or urgency but one of letting the lyrical lines fully blossom, and yet biting, accenting the summit and resolution of the crescendo which Schubert's marks forte but with no accent mark. And that very beginning sets the seal of the whole interpretation: both spacious, extremely lyrical, truly singing, AND biting, vehement, to the point of violence even. There have been urgent and dramatic readings of Schubert's C-Major Quintet (from the Hollywood Quartet in 1951 and Stern-Casals and friends in Prades in 1952, to Kagan-Gutman and friends live in 1989, Archibudelli in 1990 and the Borodin Quartet in 1994, via Heifetz in 1961 and the Taneyev Quartet in 1963 already with Rostropovich, see links in the comments section) and there have been spacious and lyrical ones (from the Vienna Konzerthaus Quartet in 1951 to the Marlboro Quintet in 1986 and the Foné Quartet in 1992, by way of the Aeolian Quartet in 1966), and there have been plenty that were somewhere in between. But the Melos Quartet and Rostropovich in 1977 achieved a unique and unprecedented reconciliation of apparently incompatible extremes.
Incidentally, they were also one of the first ensembles to play the first movement exposition repeat and fit it on an LP-side with the Adagio, preempted only by the Tatrai Quartet and the Alberni Quartet, both recorded two years before (Franz Schubert - String Quintet in C Major (Hungaroton / White Label) and Schubert: String Quintet In C Major- Op 163-D956). Prior to that, the time constraints of an LP side didn't permit it, and other than the Taneyev Quartet nobody had recorded the first-movement repeat; and on the Taneyev Quartet's original Melodiya LP the slow movement was divided between two sides (I think the later reissues for Western consumption repaired that, Schubert: String Quintet Op. 163 Taneyev Quartet With Mistislav Rostropovich). Early CD versions, the 1983 Alban Berg Quartet or the 1984 Orford Quartet, still didn't play it. Melos also conforms to the pronouncements of modern Schubert musicology by not playing the big chord before the repeat bar the second time, before seguing into the development section (and they weren't the first to do so, even among versions that didn't play the repeat), an option which I find both highly disputable and jarring, as it is not based on any original sources (the manuscript is lost) but only on the musicologists' surmise of Schubert's intentions, and I can't imagine that Schubert did NOT intend the "echo" effect deriving from the repetition of the same chord in another key three measures after the repeat bar. Fortunately Melos does not conform to the other pronouncements of said musicology, which shortens the Scherzo's second repeat (I've found only four recordings that did, starting with the 1986 Juilliard Quartet) and requires the Quintet's final chord to be strongly accented rather than played diminuendo - intended by Schubert or not, but a masterful conclusive touch.
And the second movement! Other than the Taneyev Quartet (but is was only belatedly and parcimoniously distributed in the West), nobody had ever taken its outer sections so slowly before, and many thought then and still think now that Melos-Rostropovich were too slow. I don't, I think it elicits a wonderful "time suspended" mood. In 1985 the Lindsay Quartet went even further, and by a margin, also extending (like the Taneyev Quartet had done) the slow approach even to the middle, agitated section (Schubert: String Quintet in C, D. 956). Some also have taken exception with the sonorous, "vulgar" pizzicatti of Rostropovich's second cello (which, indeed, Schubert marks pp, and even ppp on their return at the end). But it is obvious, it makes total sense: not only is it meant to dialogue with the ensuing first violin pizzicatti: it is a death knell, the ticking of time announcing the implacable coming of the Grim Reaper, and Melos' breathless middle section is a violent cry of revolt. Certainly this is not "Urtext", not Schubert's intentions, he wrote no tempo change between the first, "slow" section, which he evidently meant to be paced much faster than what Melos-Rostropovich do, and the middle one, which should be called "agitated" rather than "fast" (and very few versions even attempt anyway to play both sections at similar tempi). But who cares about "Urtext" with a reading of such depth and feeling? This is no "realization" of Schubert's music, this is a re-creation, a reinvention of it.
Melos' tempi in the Scherzo and Finale are not out of line with the interpretive traditions existing then and now. The Scherzo is not particularly fast but played with a kind of despaired vigor which makes it sound "presto" enough, and the same mood is present in the finale, where, unlike many, Melos does not take the more lyrical moments as an excuse to linger and coy. They end with a dashing coda and, as already mentioned, the long decrescendo on the final chord which modern musicology now disputes (Schubert's mark is now understood as an emphatic accent), but which I will ever find a master stroke, and no matter whether Schubert intended it or not.
I've been recommending this version in my other reviews of Schubert's Quintet, although I hadn't heard it in quite a while: I relied on memory. Since I return to it only occasionally, the risk is that, because in the meanwhile I've heard many different versions and opened my ears and mind to many other interpretive approaches, I'll now hear flaws in it that I hadn't perceived before. And indeed, there are a few. The sonics are very detailed - the litmus test being how clearly the viola can be heard - but also somewhat distant and resonant, giving a "symphonic" perspective but depriving the music of some of its impact, as with the violent shouts of the first violin after the first movement repeat bar at 10:57, especially in comparison with subsequent versions, like the excellent Chilingirian Quartet in 1981 (String Quintet in C, a version strikinly close in its overall approach to Melos) or the famed 1983 Alban Berg Quartet (Schubert: Quintet in C, d. 956). For the same reason the Scherzo and Finale sound powerful, but somewhat thick and muddled. Instrumentally, there is a small lack of lyrical bloom and traces of frailty in first violin's Wilhelm Melcher's tone in the long lyrical lines of the first movement, after the repeat bar, especially given the richness of the cello section. I don't hear anything in first cello Peter Buck that seems inferior to second cello Mstislav Rostropovich, but, surprisingly, I've heard crisper phrasing of Schubert's staccato than Rostropovich's in the first movement: he tends to legatoize it, which deprives the music of some of its bite and tends to thicken the textures.
But what still tips the scales for me in favor of this version, even over some that may be very close in approach and better sonically and instrumentally (starting with the Chilingirian Quartet - other than the fact that Chilingirian doens't play the first-movement repeat), is Melos' unique Adagio. Taneyev and Lindsay went even further, but in particular in their very slow middle section they can easily be felt to be way over the top. Rarely, if ever, do I recommend just ONE version of a great composition, because I feel that no one interpretation can say everything about a masterpiece and highlight all its facets, and this one shouldn't be the only one in the collection of any serious lover of Schubert's String Quintet. Ideally it would be complemented by two "extremes", a very urgent version like the Borodin Quartet and a very spacious one like the Marlboro Ensemble or the Foné Quartet. And because of its Adagio I wouldn't exactly call Melos a typical "middle-of-the-road" version either. Still, if you have only one version in your collection, this can be the one, because it achieves such a great combination of the opposites. And for those who'd find Melos-Rostropovich's Adagio over-the-top expressively, the Chilingirian Quartet's version offers a great if unexpected "B-plan" (but without first-movement repeat). Other possible options (and with first-movement repeat) are the 1980 Brandis Quartet (Schubert: Qnt D956), the 1990 Guarneri Quartet (Franz Schubert: String Quintet in C Major, D. 956 (Op. 163)) and the Melos Quartet's remake from 1993, with second cellist Wolfgang Boettcher (Schubert: Quintett in C-dur). And if you are not afraid of a really extreme Adagio, don't miss the Lindsay Quartet (link above). See my reviews for more details.