There's a fairly extensive chorale for brass in the last movement of Gustav Mahler's Symphony No. 2 in c minor ("Resurrection") (1894). This chorale (beginning at rehearsal 10 in the Dover edition) starts softly, in the trombones, tuba, and contrabassoon. After three soft four-measure phrases, the trumpets and horns join in for the last, seven-measure phrase. A gradual but inexorable crescendo pushes through the first five measures, culminating in a fortissimo (ff, very loud) accent followed by an immediate reduction in volume to piano (p, soft). The final two measures combine another crescendo (with bass drum and timpani rolls) with movement towards a harmonic resolution and the beginning of a new section.
It's a very challenging passage because of the control needed to maintain the soft dynamics and intonation at the beginning of the passage, the control needed to pace the increasing dynamics and not peak too soon, and finally the control needed to keep the tone from "spreading" at the high volume levels. Though Mahler marks both of the loudest points "ff" it is normal to leave a little room at the top of the first crescendo for the second, climactic one, to be louder still. Naturally, the addition of the percussion will help with that, but the brass should be louder, too.
I heard a performance of this work by the North Carolina Symphony in the early 70s. I believe Gerhardt Zimmermann was the conductor, but I could be mistaken. At any rate, it had been a well-played, well-paced performance. When the height of the first crescendo was reached the sound was full, rich, and centered, and as loud as I'd ever heard anything. I thought they hadn't left any room.
Not so. They passed the height of the first crescendo with a full two or three beats left, and kept going. As they continued to get louder there was no loss of tone quality. In fact, the sound got ever richer. The climax was electrifying and set the stage for an extraordinary finale.
The players walked a tightrope throughout that passage. The risks they took (had the sound or intonation gone off, the spell would have been broken, probably irretrievably) took the entire performance to a new level. Not only because they succeeded--the attempt itself added tension and excitement to the sense of expectation the chorale embodies. This atmosphere of risk, of chances taken, can't be captured in a recording, no matter how expert and expressive the performance. The reason is simple--you know they are going to "make it". If there was a take where the brass had lost it, it wouldn't have been included. Besides that, once you've heard the recording you know how it's going to come out. This tightrope effect is an important part of any live performance, in any art.
I'll go a step farther. It seems to me that this tightrope effect is actually part of the expressive content of music. Not all of it, of course, but in passages like this one and much of the virtuoso repertoire the difficulty itself is expressive, especially when the limits are pushed, as in the Mahler performance I described.
PS: It's also why using microphones is always wrong in opera. The projecting of character through the unaided voice is integral to opera performance. The struggles of the character are mirrored in the achievement of the singer, or in some cases, his or her pride and/or arrogance.