>>Manche Sänger sind allerdings ein wenig bequem wenn ein stimmhafter Laut (Vokal) folgt und singen (falsch): "sah ein Knabein Röslein stehen".
>>Das hat nichts mit faul oder falsch zu tun, sondern damit, wie der Dirigent es haben will (es soll z. B. nicht so hart klingen). Von den Noten her heißt es Kna-a-b(p) ein, das man auch Kn-a-a bein singen kann.
>>Nach meinem Eindruck ist die Auslautverhärtung im ersten Fall schwächer als in einem Wort, das wirklich auf 'p' enden würde.
I hope this important musical point won't be overlooked in all the discussion from a linguistics point of view. Perhaps other singers -- I know there are several people with choral experience -- might still comment.
Obviously I'm not a native speaker, but in practice, surely there's a big difference between indicating the standard pronunciation lightly and tastefully, vs. capitulating to it so that it destroys all musicality?
If you sing every word exactly like the phonetic symbols in the dictionary, the individual sounds will be more important than the musical line, and the overall effect can be pretty bad. Surely no one would really advocate aspirating final consonants so heavily that spit flies? (... sah ein KnaPP!!! ein...), or making a huge glottal stop so that the sound actually stops altogether in the middle of the phrase, even with totally unimportant words like 'ein' and 'es.'
Yes, sometimes large groups need to exaggerate certain sounds in loud pieces in large halls. But especially for soloists, chamber choirs, or quieter passages, any exaggeratedly correct pronunciation can easily be fatal to musical phrasing. That's why German has sometimes been said to be a barbarous language for singing compared to, say, French or Italian, and why some more old-fashioned German choirs have sounded more martial than musical (though I hasten to add that I notice that much less in recent decades, where more singers have been exposed to directors from all over Europe and the UK). If you take German pronunciation overliterally, it can lead to a choppy, harsh style, not a fluid legato line.
It seems to me that great artists like Fischer-Dieskau were great precisely because they put a lot of detailed effort into achieving the best possible compromise between those two competing demands, clear diction and legato phrasing. I think if you analyze individual lines by a really good lieder singer, you'll probably find that the final consonants or glottal stops are prominent only when the text at that point deserves extra emphasis, and that those phonetic features are much lighter when they're in the middle of a phrase where they would jar if they stuck out.
In fact, I believe that's true to a large extent even in speech. No one is going to make every glottal stop between unimportant words or spit out every T; that's why normal speech is sometimes represented by phonetic spelling, like Wasisndas? instead of Wa!SS! isTT! DD!enn DD!aSS!
To me all that could partly explain why Sachs, who is a frequent operagoer, might find exaggeratedly "correct" pronunciations distracting in the middle of legato lines. Even in a line like 'Alberich brach ihren Bunt' (#31), where the T is actually at the end of the phrase, not the middle, really spitting out the T could simply break the mood.
A good director once pointed out to me that (good) German singers' final T's are actually very delicate, partly because there is actually a micropause before the consonant, which allows them to place the consonant precisely yet quietly, without having to make it unusually loud or emphatic (aspirated to the point of spit). Maybe that's partly what many foreign singers haven't quite got the hang of -- they're over-aspirating because they know it's supposed to be aspirated, but they haven't got a feel for how much is too much in a given phrase.
I also wonder if manni's distinction between lenis and fortis could partly explain why he and Sachs might actually hear small differences in degree of articulation that people from farther north don't perceive. I don't fully understand it myself, but because they're both experienced listeners to classical music, I'm a little skeptical that they're just imagining it (e.g. #39, #57) in the way that undergraduates new to linguistics might.