# Noise before P-onset in iterative deconvolution
**URL:** https://discourse.obspy.org/t/noise-before-p-onset-in-iterative-deconvolution/1385
**Category:** Related: rf
**Created:** [March 23, 2022, 9:30am UTC](https://discourse.obspy.org/t/noise-before-p-onset-in-iterative-deconvolution/1385 "2022-03-23T09:30:40Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@Thea92](https://discourse.obspy.org/u/Thea92)
#### Post date: [March 23, 2022, 9:30am UTC](https://discourse.obspy.org/t/noise-before-p-onset-in-iterative-deconvolution/1385/1 "2022-03-23T09:30:40Z")
Hi all,
for some receiver function techniques (e.g waterlevel), the noise ahead the P-onset can be used to estimate the reliability of the peaks at 0+ seconds.
Do you know, how the iterative deconvolution is working and if it is reliable to calculate the RFs like from 20s ahead the P-onset? I wonder if for the iterative deconvolution is really showing the noise, or if the peaks ahead 0s are artefacts and actually ‘later arrivals’, which are placed ahead 0s.
Many thanks in advance,
Thea
---
### Author:  [@trichter](https://discourse.obspy.org/u/trichter)
#### Post date: [March 24, 2022, 7:59am UTC](https://discourse.obspy.org/t/noise-before-p-onset-in-iterative-deconvolution/1385/2 "2022-03-24T07:59:52Z")
Hi Thea,
I did not really use iterative deconvolution extensively. Check out the references given in the doc string for a description of the method.
I’d say it is OK to use the signal before the P-onset as noise estimate. If you are afraid of wrap-around effects you can append some zeros to the signal before applying the deconvolution.
Cheers,
Tom
---
### Author:  [@Thea92](https://discourse.obspy.org/u/Thea92)
#### Post date: [March 29, 2022, 8:19am UTC](https://discourse.obspy.org/t/noise-before-p-onset-in-iterative-deconvolution/1385/3 "2022-03-29T08:19:16Z")
Hi Tom,
thanks for your advices.
Cheers,
Thea