clipped signal identification and reconstruction

[SoundJudgement]

Consider incoming sound that is clipped such as a face worn mic whose preamp is too sensitive resulting in clipping when the presenter talks too loud. By means of slope analysis, the clipped section of a lobe could be isolated. Between the clip boundary and the zero crossing, original signal remains. This original signal on some lobes is fairly close to symmetric and could be replaced with a rescaled sine lobe that best overlays the two remnants of original signal. When the clipped lobe is asymmetric, a parabola could be calculated to smoothly bridge the signal remnants. Although such a process would introduce potentially dozens of milliseconds of delay, the reconstructed signal would be more listenable than the clipped signal.

[volounteer]

@soundjudgement

just cause something could be done does not mean that it should be done.

the proper solution is to avoid the clipping not try to fix it

get a mike with agc built in (shure used to make them) and turn the sensitivity down so it does not clip.

[Mike C]

At best processing like that would be done in post production as it would take too long to process for live audio.

Make sure the mic is properly positioned.
If the mic element itself is clipping, use a different mic, Countryman
E6’s for example have three different mic sensitivities available.
Set the transmitter pac gain up for the loudest input.
If the mic pre on the mixer is getting hit too hard ride the trim.

Some of the higher end wireless transmitters do have AGC’s as well as being able
to adjust the body remotely from the receiver.
That would still not take care of the mic element itself overloading.

[nottooloud]

iZotope RX is one of many programs that can reconstruct a clipped waveform. I’ve salvaged many recordings with it. One problem is latency. Another problem is you need to reduce the output enough to make room for the reconstructed waveform. Turning down the mic is a better idea.

[Author]

Posts