The MIDI protocol has always had – quite unfairly so – a bad reputation among music purists, based on the belief that it generates a robotic and artificial sound (e.g. Karaoke backing tracks).

In fact, MIDI itself does not sound per se: it only carries the information on the performance of the musician and communicates them from one machine to another. A simple 3-byte MIDI message called “Note On”, for instance, instructs the synthesizer or sampler, tuned to the channel you are communicating with, with regard to the note to be played and to the speed of the attack (note volume).

Since its creation in 1983, the MIDI protocol has never been modified, but today, thanks to the rapid growth of RAM and Hard Disk capacity, it finally can make the most of the possibilities offered by sampling.

In the past, samplers were not able to reproduce all the nuances of human execution or the sound details of the instrument that they had to simulate. The piano of an old Yamaha PRS 300 arranger has 3 samples, one for the low, one for the medium and one for the high octaves, spread over 61 keys through pitch shifting (pitch change by means of digital processing). The result of using such a primitive technology consisted in the mechanical and unnatural sound we already mentioned.

Today, to reproduce the sound of a single snare drum, it is customary to record a hundred or more samples in order to reproduce the dynamics of the instrument (from ghost notes to the strongest hits), to simulate the use of the right and the left hand (round robin), to turn off the snares through key switching and so on. Some libraries also offer the possibility of activating and mixing various mikings (e.g. in the case of a snare drum, close miking, overhead and room mics).

Such a high level of realism makes it easy to trick the ear of listeners into believing that you have recorded a drummer in flesh and blood. After all, is this not precisely what happens before every single beat is stored and mapped in order to be played by a sampler?