I've been listening to M7 a lot lately, and heard Haitink's very strong performance with the CSO a couple weeks ago. The concert got off to a rocky start when the Tenorhorn player cracked some notes, including his first. And, in M7, many of the effects depend on subtle orchestral colors, some of which vanish in the fog and glare of Orchestra Hall. Despite that, I think Haitink did an excellent job, leading a performance that was vigorous and straighforward.
However, he has yet to come to terms with the second-last note in the brass (in the Finale), which starts loud and is supposed to fade to piano. In his 1969 RCO recording, Haitink makes a crescendo instead, and to my recollection, he did the same in his BPO recording, which I no longer own. At the concert, he held the loudness steady. It's a wonderful Mahlerian joke if played as written, hinting at the onset of yet another false ending or digression; why ruin it?
As I've listened to numerous recordings, here are some quick observations: I've decided that slow tempos in the Nachtmusik movements really don't work. Movement II is a march and doesn't take kindly to gratuitous ritards, which break the momentum. I think movement IV should sound lighthearted, but some of that character is lost when the tempo drags.
A few notable recordings: Haitink 1969 (except for that crescendo) and Kubelik (studio) are both superb. Inner voices register clearly, particularly so in Haitink, with the RCO in top form. Levine/CSO is a very strong performance in good sound, though his movement IV is a bit slow for my taste. (Barry, note the magnificently atmospheric tamtam in the soft strokes in movement II.) As both a music lover and a Chicagoan, I'm pleased to note more than a whiff of Fritz Reiner's CSO still remained then.
Kondrashin/Leningrad is a nicely characterized and unmannered performance, in sound that is a little dim but quite listenable.
In Inbal/Frankfurt, a fine performance hides beneath homogenized sound.
Stenz is a likable and effective performance, with some peculiarities. E.g., at cue 4 in the first movement, Mahler instructs "Von hier an (unmerklich) drängend" ("From here on (imperceptibly) accelerating"), yet Stenz abruptly shifts to a faster tempo.
Gielen, to my ears, does just about everything right, and has a fine orchestra captured in good sound. However, the interlude in the first movement is unusually slow, and I find that questionable. On the other hand, if you want to hear every note in the mandolin and guitar in IV, played with great character, this is your recording.
And the performance that impressed me the most: Jansons/BRSO. Everything comes together: Compelling conducting, first-class orchestral execution, and excellent sound with both detail and depth, a rarity in live recordings. An example: In the first movement, two bars after cue 64, the bass drum plays two 32nd notes followed by a quarter note; at the same time, the snare drum plays a triplet of 32nd notes followed by the quarter. That cross rhythm is clearly audible, nearly a miracle in any circumstances, amazing in a live recording. Jansons is reportedly working from a revised edition, and I'll note one oddity: In the Finale, at cue 254, the first two thuds on the bass drum are missing, yet are present when they recur two bars later. I don't know if the player was inattentive or if this is another of Mahler's jokes, reflected in the new edition. Regardless, this is a high water mark in M7s.
Finally, in a review of Haitink's CSO performance, John von Rhein of the Chicago Tribune mentions that Erich Leinsdorf wrote that while the first movement is "is at least bearable, one certainly cannot say the same about the finale." Yes, this movement is episodic, often manically so, but I wonder if it wasn't Mahler's way of suggesting that the Austrian symphonic tradition was cracking apart. I'm also tempted to suspect that the tempo indications in the Finale are a sort of joke independent of the music; they're fun to read on their own, and suggest the tantalizing prospect that he was poking fun at himself.
