Why the Silence at Beginning of Some Samples?

[McLilith]

When I prepare a snippet of digital audio for use as a sample, I always trim off any silence or unwanted bits at the beginning of the audio file. I usually zoom right down to the individual sample level, and start the file at a zero-crossing.

When I examine some of the 3rd-party samples that other folks produce, I frequently find significant bits of silence at the start of their samples. It could be anywhere from 0.1 seconds to maybe a full second or more of silence, before the actual audio begins. It seems to me, having random bits of silence at the beginnings of your samples would just mess up the timing of your performances.

Are there any common reasons why someone would actually want this extra silence at the beginning of a sample?

To me, it seems like it's just probably sloppy sample editing and preparation, but if there's a good purposeful reason for it, I'd really like to know. I'm always interested in learning new things!

[bigcat1969]

Much editing is automated or semi-automated so that might be part of it.

I do remember a case with one piano I was working on where I released it with what probably too much space at the front of the notes and later tightened up the samples after some complaints about input lag. Some people said the reason they liked the original was because of the pre-note sound (which I couldn't hear) and that the new tight version lost the specialness of the piano. Also my version of VSCO2 is more heavily edited and normalized (I love consistency) than the SFZ version done by Sam and there is enough difference that I recommend people try both and many prefer the more organic SFZ version. I did actual loosen that up as well, and generally am a bit looser than I used to be, if I remember right especially on sustains as the swell into notes can be important.

[McLilith]

For those folks that like a bit of delay before the sample starts playing, you could always create really tight samples, then add a "random delay" parameter that people can turn on or off to suit their own tastes. You could even let them dial up the particular amount of randomness they prefer with a "front panel knob" on the GUI of the sample library, or maybe use some special key switches to select the amount of randomness. It's fairly easy to do with SFZ. So, I'm sure it could be done in Kontakt and other top-end sampler formats as well. That would probably keep everyone happy.

[DSmolken]

In some cases it's things like pick attack noise, or the hi-hat pedal being stepped on before a foot chik - quiet, but important to realism. I know Ample guitars have that, but have a knob to reduce or get rid of the delay by skipping that sound.

Most of the time, though, it's probably just a lack of appreciation for just how much the extra effort of trimming samples consistently feels better when playing them from a keyboard, e-drum controller, MPC or whatever. Consistent, predictable, quick response (along with well-scaled dynamic layers) really boosts confidence and lets the player focus on expression instead of trying to compensate for the samples' delay.

[sjm]

I am not a professional sample producer or anything, but when I created a free pack recently, I had some samples with silence at the beginning. These were longer tempo-synced samples that don't begin making sound on the 1. In that case, removing the silence from the beginning would have brought the sample out of whack and made it hard to align in your project.

So I think there's a use case for longer tempo-based things. I trimmed the silence off all the one shots though, and would find it strange not to do so.

[enroe]

To me, it seems like it's just probably sloppy sample editing and preparation, ...

Most probably that is the reason.

Because whenever you use sloppy samples with sloppy sample-starts in
a real song you'll get mad, because the timing is weird up to the pole.

So there's no other way than to adapt the sfz-files with the "offset"-
code.