THE MOST FREQUENT -- and often the hardest -- question that music writers get asked is, "Which recording of 'X,' 'Y' or 'Z' is the one to buy?"

There was a day when there were so few alternative recordings of even the most popular symphonies or operas that the answers weren't hard.

Obviously, in the monaural sound period, Toscanini's "Otello" was the right one, and it still may be. And the Flagstad-Furtwa ngler "Tristan" was without compare.

Now, though, with the profusion of performances presently listed in the Schwann catalogue, such answers are no longer easy.

And, curiously, the more towering the work, the harder the choice sometimes seems to be. This is probably because there are certain compositions that nobody approaches in a slapdash way -- works that don't permit performers to get away with riding glibly on the surface, as they may, say, with Ravel's "Bolero" (a fine work, by the way). Some works are better left alone than half done -- for example, Bach's "A Musical Offering," Beethoven's C-sharp Minor Quartet, Wagner's "Tristan und Isolde," Verdi's "Falstaff."

Consider this quandary: Someone recently asked which was the best recording of Mozart's "The Magic Flute." I stumbled around and didn't really come up with an adequate answer. The best I could do was mention specific details of the available stereo recordings (plus the historic mono version that Sir Thomas Beecham made with the Berlin Philharmonic in 1937). To my knowledge, there was no best production, though there were favored performances in different sets. And now, having listened to eight of them, I think the answer is the same.

For instance, Gerhard Hu sch's superbly styled Papageno in the Beecham version or Dietrich Fischer-Dieskau's captivating Papageno on the second Karl Bo hm performance; Fritz Wunderlich's silvery Tamino with Bo hm; Ezio Pinza's matchless Sarastro (in excerpts); the clarity and strength of Otto Klemperer's direction; and the grace of Karl Bo hm. Of the digital versions, Bernard Haitink's (with Lucia Popp's superb Pamina) is especially fine.

So how does one choose one over the other? On each of these eight splendid recordings, there is something special.

And "The Magic Flute" is very close to the hard core of what music is all about. It is pure fun, in the most serious sense of purity. The gap between its surface display, as in the coloratura pyrotechnics of the Queen of the Night, and the pared-down simplicity of the music of the Three Boys is unmistakable; and that seeming incongruity is what makes this opera -- for all its occasional lack of Mozartean urbanity -- unique. What other work combines the conventions of a sometimes trite little fable with music of such profundity? Just the overture alone -- Mozart's greatest -- illustrates a remarkable dichotomy, with the tentative harmonies of the introduction contrasted with the assertive e'lan of its main section.

What are the things that matter most in a performance of "The Magic Flute"? They vary depending on one's tastes and needs. Good casting is vital. Glistening sound will be more important to some than to others. And cost may also be a factor.

The most important figure in an interpretation of "The Magic Flute" is the conductor, and in addition to the sets already mentioned, there are also versions presently listed in the catalogue under James Levine, Herbert von Karajan, and Sir Georg Solti -- as well as an original-instruments performance by the Amsterdam Baroque Orchestra under Ton Koopman.

The last is the only one that I would not recommend to the general listener, though it is valuable as a curiosity. Today's original-instrument movement is of great value, but it often operates under a handicap in opera. (Obviously, there cannot be a comparable original-voice movement.) And the major singers of today would be unlikely to attempt such an enterprise. The prestige (as well as the money) lies elsewhere.

The other doubtful recording is the Solti. It looks marvelous on paper. But the conductor yields to his occasional vice of being too driven and rigid. He conducts the Vienna Philharmonic -- and no orchestra plays Mozart better than it -- but some of its customary mellowness is lost. Sir Georg had an off day.

Martti Talvela is a distinguished Sarastro (as he is with Levine), and what an indulgence to have Fischer-Dieskau as the Speaker! Also, Cristina Deutekom, a singer little known in this country, is one of the best Queens of the Night on record. Hermann Prey, too, is a fine Papageno.

Karajan, who sometimes can be a little overbearing in his Mozart himself, fares better. And the sonority of the Berlin Philharmonic brass in this opera, where the brass is of symbolic as well as musical importance, simply could not be better.

Still, Karajan does not achieve quite the sense of give-and-take, of breathing with the music, that Bo hm gets from the same orchestra (it is, though, Karajan's orchestra, aside from the current brouhaha).

Where the Karajan version scores most is in its digital sound. The kind of sound that Deutsche Grammophon was getting in Berlin's Philharmonic Hall when this record was made in 1980 simply outclasses what could be done there when the Bo hm version was made at least a decade earlier.

Two particular standouts on Karajan's cast are Edith Mathis as Pamina and Jose' van Dam as Sarastro.

Karajan also has a formidable set of Three Ladies, Anna Tomowa-Sintow, Agnes Baltsa and Hanna Schwarz (perhaps only in the case of "Der Rosenkavalier's" Italian tenor do unnamed operatic characters command such formidable singers -- and salaries).

On Bo hm's version the Three Ladies are less celebrated, but the two armored men certainly are stars -- James King and Martti Talvela. On that set the leading women are not quite up to the men (Wunderlich and Fischer-Dieskau), but they are fine. After all, Evelyn Lear is Pamina and Roberta Peters is the Queen of the Night.

But the real glory here is the sense of ensemble that Bo hm achieved. This is one of the late conductor's finest accomplishments. He was always wonderful with Mozart, and here he was at his peak (as in his equally grand recording of "Cosi' fan tutte").

A competing version of "The Magic Flute" finds another venerable conductor at his very best -- the late Otto Klemperer. It is a different sort of reading -- more deliberate, but superbly controlled. The articulation of London's Philharmonia Orchestra in the overture is the most impressive on any of these recordings (and that it is slower doesn't necessarily mean it is any easier). In the process Klemperer infused the opera with a greater weight of sound than any other version. If Bo hm is spontaneous, Klemperer is majestic. Both are indispensable interpretations (Klemperer's sound is also a bit dated).

His whole cast is strong, but his Three Ladies sort of boggle the mind (Elisabeth Schwarzkopf, Christa Ludwig and Magda Ho ffgen).

Lucia Popp is the Queen, a fine performance though not quite the equal of her incomparable Pamina (with Haitink). Gundula Janowitz is Klemperer's fine Pamina here and Nicolai Gedda is the excellent Tamino. Walter Berry is Papageno and Gottlieb Frick is Sarastro.

Along with the conductor, it is the women and the beautiful digital sound that dominate Haitink's recent "Magic Flute." Popp is matched here by Edita Gruberova as the Queen, in a superb performance of what sometimes sounds almost like an unsingable role, so demanding are its coloratura runs. (I have heard Queens who seemed more as though they were strangling than singing.)

Siegfried Jerusalem is a fine tenor, but his voice seems a little too heavy for Tamino. Roland Bracht's bass is light for Sarastro, but he has style. Here the casting luxury is with the Two Armored Men -- Peter Hofmann and Aage Haugland.

Interpretively, Haitink, who conducts the Bavarian Radio Symphony Orchestra, steers an interesting course lying somewhere between Bo hm and Klemperer, with some of the best of both.

James Levine's recent recording of the opera with the Vienna Philharmonic is also quite fine. To my taste it is occasionally a little light and brisk, but not too much so; there is none of Solti's excessive force. Levine's cast is good, without quite matching Haitink's, except in Martti Talvela's superior Sarastro. The other leads: Ileana Cotrubas (Pamina), Eric Tappy (Tamino), Christian Boesch (Papageno), Zdislawa Donat (Queen of the Night).

The chorus of the Vienna Opera is stunning in this performance, matched perhaps only by Klemperer's Philharmonia Chorus. And Levine's is the only performance in which the spoken dialogue is not seriously cut.

Finally, there is the classic Beecham recording, a real bargain at today's prices. It was once the recorded standard for how to perform "The Magic Flute." Despite the recording's antiquated sound, the playing of that prewar Berlin Philharmonic still sounds formidable.

As mentioned, Hu sch's Papageno is a marvel. And Erna Berger's Queen of the Night is not far behind. Helge Roswaenge's Tamino is wonderfully ardent. And Wilhelm Streinz is a good Sarastro. Only Tiana Lemnitz' Pamina disappoints on rehearing.

Beecham's conducting is just what you would expect, full of wit and grace; he also stints a bit on the music's gravity.

There are other conclusions one reaches after listening to "The Magic Flute" eight times. One is that there several aspects of this extraordinary creation that may elude modern understanding -- especially the Masonic symbolism.

Mozart, of course, never wrote a score of greater clarity and imagination. But, let's face it, Emanuel Schikaneder's text cannot bear comparison with the polished books that Lorenzo da Ponte provided for Mozart in his three previous operatic masterpieces ("The Marriage of Figaro," "Don Giovanni" and "Cosi'").

Still, for all its oddity, the text of "The Magic Flute" had at least two special appeals to the composer. One was his interest -- and involvement -- in Freemasonry. And the other was his concern about the paucity of operas in German during his time (opera was an Italian invention, and in Mozart's day there were precious few in any other language).

But perhaps it's really not all that important that the symbolism of, say, the Three Ladies be clear to us. Why is it that they (the servants of the Queen of the Night) should be the ones to save Tamino, equipped with their silver javelins, emerging from their odd little temple, stabbing the serpent dead?

The British critic, William Mann, explains it this way in his book, "The Operas of Mozart": "To serious minded commentators these Ladies represent Unenlightenment or the Roman Catholic Church or even female Freemasonry." But who understands what that means?

The mystery of the Three Ladies is not, to be sure, of the universal sort that underlies a work like Mozart's Requiem, or Beethoven's "Missa Solemnis." Clearly, though, they meant a great deal to Mozart. And you can afford to take them for granted dramatically. Regard them just musically, and the message could not be clearer, as they proclaim whatever uplifting spirit that rests in the key of E-flat major, just as unforgettably as the composer did in the exultation of the 39th Symphony.