Section 7 — Sampling/Signal Source Concepts
Looping
153
Looping
One of the fundamental challenges of sampling is to make efficient use of the sampler’s memory.
You simply cannot sample the entire duration of every event. Plus, you usually want a sound to
continue playing as long as you’re holding the key down, regardless of how long the original
event was. This is why we employ Looping.
When we loop a segment of sound, what we are doing is playing through to the end of the
looped segment and then instantly starting back at the beginning of the loop, as if playing one
continuous sound. It’s a lot like a tape loop — a segment of recording tape with the two ends
spliced together so that it cycles repeatedly.
Consider that a low note on a grand piano can last 30 seconds or more from the time the key is
struck to the time it decays into silence. Sampling this entire event does not make sense because:
• You could use up all your sampling memory on one note, leaving no memory for
multisampling; and
• After the initial attack transients die down, the sound of the piano string settles into what is
essentially a repeating waveform. This means that you can select one cycle of this repeating
waveform and loop it (play it over and over). All the data after the loop can then be removed
(truncated) because you don’t need it. This conserves memory.
One cycle of the wave
can be "looped"
The data after the loop can be
Truncated to conserve memory
The sound will reach the loop and then
play that single cycle over and over.
The decay of the original sound can be
recreated using the Amplitude Envelope
A piano, solo violin, guitar, or woodwind are all sounds that you could loop in this way.
However, there are other sounds that do not settle into a repeating waveform. A full pipe organ,
string section or a choir would be good examples of such sounds. These sounds remain
extremely complex throughout their duration. Using a single-cycle loop on a choir sample would
cause the sound to suddenly become very static and one dimensional when it reaches the looped
section. These complex, ever-changing sounds require that we employ longer loops — repeating
a large segment of the sound rather than just a single wave cycle.
Generally, successful loops fall into one of two categories — single-cycle loops, which we refer to
as short loops, and loops of about a second or more, which we call long loops. Which type to use
depends on the sound. If the sound settles into something relatively static, use a short loop. If
the sound is composed of many elements that constantly chorus and interact with each other over
time, you will need to use a long loop.
Auto-Loop Finding
When you loop a sample, it is important that the beginning and end of the loop “match up” to
form a continuous wave shape — otherwise the loop will pop or click each time it reaches the
splice point. For this reason, the best loop points are those which occur where the wave crosses
the zero line (called a “zero crossing”) and where the direction of the wave at the loop start
matches that at the loop end.
The ASR-10 has a system parameter called AUTO-LOOP FINDING that, when ON, allows you to
choose only loop start and loop end points which meet these criteria.
Press Edit, then System•MIDI, and scroll until the display shows AUTO-LOOP FINDING =
<ON,OFF>, and set it to ON. Now, whenever you move the loop start of any WaveSample, the
ASR-10 will automatically jump to the next zero crossing which matches the wave direction at the
