I am sorry if Nick feels that my comments are unjustified - I am merely trying to correct his slight errors in applying theoretical physics to a practical subject.  Acoustics is not my profession, but electrical engineering is - and it would be hard to find any field more closely related to the practical design of acoustic systems.  I did attend one of prof. Holmes lectures, took Amar Bose's acoustics class, and also did my senior project on an electronic implementation of a trumpet-like resonator.  But most of what follows is based on the practical experience of applying system theory to real devices which I gained in electromagnetic fields courses.

Quoted verbatim from his web site:

 What in fact happens is this: as the mouthpiece is removed from the
 horn while you are still playing the note, the feedback is indeed
 removed, thereby removing the support of resonance. What this
 physically means is that now you have no "slots" or notes that will
 pop out discretely. You can play a continuum of pitches for a rather
 wide range. The mouthpiece, by itself, acts as a tiny "megaphone" of
 sorts. There is no reason whatsoever that you shouldn't be able to
 sustain a tone while doing this.

 It might feel a little funny for some players to feel the feedback
 suddenly disappear, and they may unconciously stop the lips from
 vibrating by opening the aperature between the lips too much in the
 process. This will stop the note, but it is the player stopping the
 note.

What Nick is missing is the fact that the trumpet resonances don't just make it hard to play non-notes, they also make it substantially easier to play notes.  Thus buzzing without these resonances requires substantially more work - because it is "all you" with no help from the non-existent instrument.  Think for a moment of the analogy of pushing a swing - it takes very little effort, exerted at the right time, to keep the swing oscillating with a fairly high amplitude. This is because the mass and the rope length combine to form a resonant system, which helps us out.  Now imagine manually trying to lift the person and the swing up and down at the same rate.  Even if you don't try to lift them as high (buzz softer than you play the trumpet) it is still much more work.

Sitting balanced between the steady air pressure and resonance of your oral cavity and the steady back pressure and resonance of the trumpet, it is very easy for the lips to vibrate.  Remove almost all of the external side leaving only the mouthpiece, and it is much harder for the lips to vibrate due to the inbalance.  Thus unless the player compensates by changing the aperature (something trumpet players do automatically when changing register or volume, without realizing it) the tone will most likely stop.  Buzzing is IMHO usefull, but it is
somewhat different than playing the trumpet.

[For those who really care about theory, if one had a trumpet that was perfectly resonant at all frequency components of the sound spectrum desired then it would be a purely resistive load, requiring no compensating reactance from the lips.  One would think you could remove the trumpet and have the lips operate identically into a different magnitude of resitance at a reduced amplitude, but this assumes that they operate linearly across this rather wide range of resitances, which they decidedly do not - ever wonder why you can't play infinitely pianissimo?].

another quote:

 In physical analysis of the spectrum of resonance produced by trumpets
 a device is used to energize the horn electromechanically. The device
 is called a salpignometer. They were used by the likes of Arthur
 Benade and William Cardwell, both expert acousticians and physicists
 on brasswinds. In the best device the horn is energized by a
 transducer that places a membrane over the mouthpiece sealing it
 (closed end resonance). The membrane is exposed to a chamber that is
 driven by a small speaker. The sound pressure level is controlled by a
 small feedback loop and voltage controlled amplifer. If the mouthpiece
 is removed from the horn while this device is playing it, it doesn't
 suddenly stop producing a tone. It simply stops resonating. The tone
 continues, but it is only a "mouthpiece" tone.

 The horn is a resonator. It doesn't control the input. The player does that

This last statement is Nick's crucial error.  The trumpet does its best to control the input - meaning that the player is usually forced to stay fairly near one of the pitches the trumpet will permit.  The salpignometer is a device carefully engineered to produce the same volume of sound at any frequency and into any reasonably load, so that it may be used to analyze the trumpet alone without regard to the player half of the system.  (This consistency across volume and frequency is the purpose of the electrical feedback system).  In contrast to this "ideal source" the player's lips constitute a non-ideal source, which is very, very heavily influenced by the trumpet - in short, the trumpet does its best to control the input.  I can force an almost recognizeable first-octave major scale out of a trumpet without using the valves - but it takes a very, very skilled player to really be able to do this musically.  The reason this takes so much skill is that the player is forced to contort his lips into some very unnatural positions to fight the trumpet for control. Benade's salpignometer by contrast would get hot - but do the job fine - - so the comparison is invalid.

And a final quote:

 I have a great deal of respect for the musicians who have occasion to
 make such mistakes. What bothers me is that they are influencing
 students erroneously. If they are indeed true academics and
 intellectuals, they should be willing to accept correction and modify
 their positions. They should be glad to actually know that they can
 make statements based on correct information giving credence to their
 academic positions. In short, there is nothing to be lost by getting
 it right.

Nick is 80% percent right... I'm just trying to make a 10% correction, and will readily admit there is another 5% of established knowledge I don't understand, plus the 5% no one has yet figured out.

Chris Stratton

- -------
Christopher C. Stratton, stratton@mit.edu
http://web.mit.edu/afs/athena/user/s/t/stratton/www/stratton.html