Csdrhrt

This question of mine started when I heard the piano score of Mars by Gustav Holst. Here is the piano score:

How do the pianists even reach triple f? I myself have come across extreme dynamics in piano pieces even by Romantic Period composers such as Mendelssohn. Not really in Beethoven or earlier and Chopin's extreme dynamics are just the extremes of quiet. That is relatively easy.

Extremes of loudness though are hard, especially for a pianist. This is partly because at some point increased force doesn't equal increased dynamic, just increased chance of breaking the piano. It also has to do with the attack of the piano compared to other instruments:

Here is the typical attack of a piano:

As you can see, no chance of a crescendo during the note length so whatever dynamic level it is at at the peak is the loudest it will be. The rest of it will get quieter at a relatively linear rate until the key is released.

By contrast, here is a violin's attack:

As you can see, there is a natural crescendo in the attack so reaching extreme dynamics is no problem. It is one of many instruments that can do a 1 note crescendo or diminuendo. In principle all bowed strings, woodwinds, and brass can crescendo or diminuendo a single note.

When I see something like ffff in a score, here is how I treat it both in my playing and in my compositions:

Orchestral: Yeah, go ahead and crescendo further, no screech will be noticeable at the level of the symphony orchestra

Soloist that isn't piano: Lower the absolute dynamics so that the same relative dynamic is acheived without a screech(so piano becomes pianississimo for example)

Piano: Extend the fortissimo until a quieter dynamic is reached, maybe using octaves to get a crescendo of mass rather than of volume. Once a dynamic of forte or quieter is reached, change the dynamic as usual.

So how do pianists reach dynamics past fortissimo in scores such as Mars by Gustav Holst that include these extreme dynamics? Do they start extremely quietly to get a relative dynamic rather than an absolute dynamic? Do they add mass(like octave doublings) to the music to get a crescendo of mass once maximum dynamic from force is reached? Or do they do something completely different?

Professionals are generally old enough and wise enough not to overplay a piano. Any given instrument has a ceiling of available sound, and the player adapts the dynamic range of the music accordingly. For example, if the music includes fff, then ff would generally be scaled back. On the other hand, in music with no higher than ff marked, those passages might be afforded full tone.

Extremes like fff can be distinguished from ff by flaring up the sound with the pedal, a technique available to any experienced player. They can also be achieved by compromising beautiful sound. Neuhaus had an amusing way of referring to the normal dynamic continuum as extending from the bounds of "not yet tone" to "no longer tone". "No longer tone" is relevant here. The use of a harder, more percussive, forced tone can signal to the listener that a limit of dynamic potential is being reached. Rubinstein used this approach when playing certain Mozart like K466: he would play ff quite stridently, though actually within himself, conveying that the instrument was at its limits, without actually introducing the modern piano's extremes of sound that can compromise the clarity of the music. In this way, the dreaded Dresden doll Mozart - underplayed, over-refined - was avoided.

Sometimes, the dynamic ceiling can also be distinguished by working with the balance of chords. Very often, the outer notes in the right hand will be the doubled melody, with harmony notes in the middle. In normal play, the harmony notes may be scaled back somewhat; in playing with maximum power, they can then suddenly blare out like an awakened brass section.

Sometimes, for instance in Tchaikovsky concertos, fff in the midst of general ff is interpreted as prominence: the part so marked need not necessarily be louder than the others, but must contrast, rather than blending into the texture. Sometimes, all that is necessary here is to draw the listener's attention to the part at the beginning, after which it will be readily followed. This can be achieved on the piano by a tiny delay in the taking of a chord, for example, so that the attack is not swallowed up by the attacks of all the other instruments.

Incidentally:

Chopin's extreme dynamics are just the extremes of quiet

Not true at all; he writes fff quite often. Examples: Op 10 No 4, Op 25 No 10, Op 25 No 12, Op 23, Op 28 No 24, Op 39, Op 52.

I've seen a pianist play a piece by Olga Neuwirth that I can only assume was marked "as loud as humanly possible", because he was playing so violently that with each chord his body would jump up from his seat. It actually looked as if he was trying to damage the piano or hurt his hands; I was expecting the key mechanisms to break at any moment. It was not just very loud, the physicality of it was truly shocking to watch.

So I guess you can always add another "f" to the dynamics; someone will be able to play it, and a concert piano can apparently take it.

Dynamics are not absolute. But you can hit a (real) piano pretty hard! It's not like a MIDI keyboard where velocity is limited to 127.

You seem to be confusing the ADSR envelopes with the amplitude of the sound. It's true that the piano has a well-defined envelope which you can't do much to change barring the point of release, but the relative amplitude of the peak of the Attack is all completely within the player's control.

As such, if a pianist want to change dynamics, they just play softer, resulting in the hammers striking the strings with less force. The ADSR envelope is (largely) the same, but the peak reached changes.

For a single note, there's nothing that can be done, unlike on a constantly excited instrument such as the violin, but you'll notice that all of the crescendos and diminuendos in the piece you've posted are over a sequence of notes, meaning that each successive note in the sequence is played with a different attack, which results in the dynamic of the phrase following the marked indication. The same wouldn't be true if the piano was meant to crescendo over a held note, because as you correctly point out this isn't possible on a piano.

Following your statement that "increased force doesn't equal increased dynamic", the starting point for a 'default' (for want of a better term) mf shouldn't be anywhere near that level. In 20 years of playing the piano I don't think I've ever got to the point where I thought "I can't physically play this piano any louder".

The point that dynamics aren't absolute from the other answers is key, for example, ff on a harmonica is very different from ff on a trumpet!