2
$\begingroup$

I generated a one second audio clip of a sinusoid at 440 Hz (16-bit, 44.1 kHz), ensuring that the last sample is zero. However, I hear a faint "blip" sound at the end of the clip, and have noticed this with similar clips that I have found online. Do you know why this happens, and how I can get rid of the "blip"?

Thanks for your help!

Improve this question
$\endgroup$
3
  • $\begingroup$ If that's not a problem, properly window your sine wave, effectively adjusting its attack and release timing... or fade in fade out depending on the duration.. $\endgroup$
    – Fat32
    Commented Jun 11, 2017 at 1:31
  • $\begingroup$ @Fat32 Apologies if I'm missing something obvious, but why do I need to window the sine wave in order to prevent the blip? Since the last sample is zero, I would expect the transition from non-silence to silence to be smooth. $\endgroup$
    – void-pointer
    Commented Jun 11, 2017 at 2:07
  • $\begingroup$ No you don't have to window it. As you also point out, its natural transition to zero is also another way to decrease that blip. But windowing will also give you zero transition at the edges, irrespective of the sinusoidal phase. $\endgroup$
    – Fat32
    Commented Jun 11, 2017 at 2:56

1 Answer 1

Reset to default
5
$\begingroup$

There's still the first derivative discontinuity. That gives a blip with a -12 dB/octave decaying spectrum. It's less audible (because it has less high frequencies) than the -6 dB/octave decaying spectrum of a proper discontinuity, but still audible, especially with a subwoofer...

Do a (long enough a) fade out like Fat32 says and there will be no audible click.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thanks! Now it's clear to me why the fade out is necessary. $\endgroup$
    – void-pointer
    Commented Jun 12, 2017 at 3:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.