Not to jump in here, but the technical term for this is velocity layers, when you see this listed elsewhere, or even in the documentation for Vienna, that is what this means. Since MIDI supports 127 volume levels, and it would be infeasible to get a real player to play at 127 different volumes, the playback software will determine which velocity layer to play based on the velocity value the DAW sends to the playback software. If you have only two velocity layers that are, for the sake of making this easy split at 1-63 & 64-127 (these values will be different for every soundfont manufacturer, and can even be different from instrument to instrument from the same manufacturer), if the value received is 63 or under, it will play the audio file associated with a softer timbre, and 64 or higher, an audio file associated with a louder timbre. The playback software further increases or decreases the playback volume based on this number.
You can manipulate this to create some very odd effects, or, get an instrument to cut through a mix by manually setting a velocity level of a note to 127, but setting its main volume very low. This will have the effect of tonally sounding like the player is playing as loud as they possibly can, but the actual volume of the note will be very quiet. I use this technique quite frequently when I need an instrument to stand out a bit more, but I dont want its volume to overpower the mix.
I also use this strategy in reverse quite frequently. I really love the Vienna horns, but, I cant stand how blatty they sound at higher velocity levels (there is a sort of 'buzz' they have at the highest velocity layer). I set the velocity for each of the notes in the horn section to something like 80, which is often too quiet, and then I use the main volume or velocity xfade, depending on the degree of laziness I am feeling at the moment to make them louder.
There is probably a better way of doing this, but understanding velocity layers helped me to make things sound how I hear them in my head, and the beauty with computers is, we can do things with instruments virtually that are not possible in real life. It may not be realistic, the way I do it, but it ends up sounding the way I want it to.
