Sample-based synthesis is a type of audio synthesis that uses either hardware or software samplers. The process involves the use of instruments or sampled sounds as the basis for its sounds.
Sample-Bases Synthesis: the History
Before digital sampling became popular, machines used analog tape decks to playback sampled sounds. Machines evolved and more powerful samplers appeared. The sample-based synthesis was first used to mimic the sounds of real instruments by recalling actual samples of these instruments while hitting the keys of a keyboard.
The first samplers were quite expensive and had limited memory. Because of this, they used the shortest samples possible. In turn, this increased their length by looping and achieved pitch alternations when stretching one sample across several notes. The first sample-based synthesis provided an unusual sample rate and bit depth; people perceived it as a grainy sound. Modern sample-based synthesis has more memory and storage, so the sounds appear to be more realistic and with more editing possibilities.
Sample-Based Synth: the Mechanics
Sample-based synthesis involves three main issues: looping, pitch shifting, and data reduction.
Looping
When looping is present, the duration of the sampled sounds played is stretched. If the user holds down a key, the sampler immediately scans through the notes until the key releases. This process involves a specific beginning and end loop points within the sampled sound. When the attack note ends, the sampler reads through the loop point until the key liberates.
Modern samplers offer automatic techniques for finding potential loop points. One of them is to perform pitch detection on a sampled sound by using a pitch detection algorithm. This algorithm creates smooth loops that are constant in pitch. It works by looking for patterns in the wavetable that specify a fundamental pitch period. Once this process determines the pitch, the sampler suggests a specific pair of loop points that match the pitch periods in the waveform.
Pitch Shifting
Pitch shifting is the second issue regarding sample-based synthesis. It’s impossible for a simple sampler to store every note played by an acoustic instrument. When a sound is recorded into a sampler memory and played back with different keys, the sampler uses the pitch shifting technique. When this happens, the sound period can either increase or decrease depending on which the key it presses. Additionally, pitch shifting involves two different methods. The first one, varying the clock frequency of the output alters the sampling rate. In this, the duration of the pitch changes. Second, the sample rate conversion shifts the pitch contained inside the sampler. This allows the playback to be set at a constant rate for all pitches.
In conclusion, you can reduce data in three different forms. These are: in a sampler by limiting the sample resolution, by decreasing the sampling rate, or by storing sounds in a reduced form.
If sample-based synthesis interested you…
If you want to learn more about other types of synthesis, please consult our entries on subtractive, granular, wavetable, additive, and physical modeling, linear arithmetic, phase distortion synthesis.
