Well, you CAUT's have touched on one of my pet subjects (not quite a pet peeve.): how hard to work on shift work, and to what end.
I must agree with Kent Webb when he asserts that the one thing that is completely consistent across pretty much all artists on all occasions is that they would like more. This is his comment from Steinway's long history of working with artists.
So what, in reference to the shift preparation, would provide more? And of what? I will refer to what I think of a shift that is very well set-up:
Well, IMHO, the "first shift position"--between strings and striking all 3 strings--provides a quieter and more cottony (wooly) sound. A small but definite extension of the tone range.
The "second position"--missing the left string--provides a noticeably different timbre, AND a significantly longer "after sound". Because the left string is not struck, it robs a bit of power at the start and returns it later when it comes into sync with the other two.
I am aware that it is a lot of work to set up a piano this way. There are even pianos that sound odd enough with "second position" that it is likely to be rarely used. The nice thing about a very well-set-up shift is that one can restrict shifting to "position one" with a turn of the stop screw. And then restore its availability for an artist that would like to use it.
For the CAUT setting, I would like to confess that our philosophy at the UW is best summed as triage. We try very hard to have our best performance pianos prepared so that they would be perfectly comfortable in The Basement. Our piano faculty pianos the same, except that the service interval is much longer, resulting in periods where the shift is perhaps less than ideal. But when they are freshly regulated, full shift.
And so on. We have pianos that pretty much get tuned when they are used, even if it is once a decade. The level of "neglect" increases as the importance of the instrument goes down. Our very best work when we work, and if "they" want more of that, they can find $$$.
There's lots on HOW to prepare the shift, but knowledge is only a start. It is, for me, definitely a skill that has to be practiced regularly.
And the result for us is that, even though I do not have notes or recordings to prove this, I would assert that ALL of the artists that come to play in our President's Piano Series (5-6 top level players each year) use the shift. Some of them quite a lot.
It seems to me that they use it "by ear", not by the position that's "supposed" to be used.
University of Washington School of Music
How are four – or even two – shift positions available with only one pedal. I must admit my ignorance since I thought that the left pedal initiated one shift only. Does the depth of pedal actuation give different shift effects? Enquiring minds want to know.
Rhea Piano Service
I guess the first thing to bear in mind is that neither the shift pedal nor the damper pedal is an on/off switch. There is an "infinite" gradation of in between. (Sostenuto IS an on/off switch, and simply malfunctions in between). This is not to say that all pianists take advantage of the range of pedal positions, but at a high level, in concert work, they do.
The basic premise, when you set up the shift pedal to miss the left string, is that this will cause less energy to be imparted to the soundboard, therefore lower volume. In fact, there is also a significant timbral difference, because the sound of two strings is different from that of three, and because you have a sympathetic string at the same pitch. Sometimes this timbral difference is pretty big, even disturbing and not easy to use in a musical way.
Going a little further, often pianos are set up so that when you depress the left pedal all the way, the movement of the keyframe stops when the left two grooves line up with the right two strings (shift is precisely the distance between strings, or nearly so). This means that full shift will produce a voicing result (in terms of the hammer striking the string, its hardness) that is the same as in at rest position. The felt between the grooves is likely not to have been played on as much or with as much force, so it will be softer. Thus, you will get a softer attack sound with the pedal partway depressed, then a harder sound with it depressed all the way.
We want the pedal to "work as advertised" - to make the sound different and softer. So we need to take action to make sure this happens, and happens in a predictable way. One of the first steps, if setting up so the left string will be missed, is to align the hammers so that the left sides of all of them will begin to miss the left string at the same time (otherwise the effect will be different for different notes). A good way to do that is to insert a cardboard shim between the stop felt at the left and the keyframe, blocking the keyframe at about "half pedal" or so, then align the hammers by listening to them: mute the right two strings of each unison, and adjust the hammer so that its edge just "ghosts" the left string, maybe it doesn't make it sound on a soft blow but does on a hard blow, or something of that nature. You really need to be aligning hammers in situ for this, so this is where a flange spacer comes into its own.
Well, I've got to get some work done, so will return to this later.
Picking up from aligning the hammers so they all will miss the left string at the same time, this will typically shift the hammers toward the treble, so they aren't centered on the three strings. This is because if you center the hammers, depending on their precise width, usually by the time they have shifted so that they are missing the left string entirely, the left two grooves will be lined up to the right two strings, which isn't a good idea. This is one of the conundrums we face.
For further specifics, I'll refer both current and possible future readers to the same thread on the CAUT list, as well as some material in a previous thread durability of hammers. But essentially the idea is to give a voicing gradient depending how far the hammers are shifted relative to the strings. This gradient should be in keeping with the overall sound, and it should be possible to "play expressively" in the shift positions, by which I mean that there must continue to be a voicing gradient at every shift position - the upper partials need to become more prominent the harder the hammer strikes the strings. In practical terms, this means relatively shallow (2 - 3 mm or so) needle insertion, because that will mostly affect the attack and especially at p and pp, but when you play loudly the harder felt underneath will come into play. As a result, the pianist can play a very light, ethereal background with a prominent melody ringing out above it, for example.
I hope this helps in connecting the various thoughts that have been expressed about shift voicing and prep. There isn't complete consensus (about shifting off the left string or not, for instance), and it isn't taught very much outside Steinway and I guess Bosendorfer tech sessions.