Hello,
It probably would be better to do it by defining a region but we never got
around to allowing the user to do that.
Wayne
On Tue, 25 Sep 2012, Beate Bersch wrote:
> Dear all
>
> I am currently interpreting relaxation data and hetnoe using ccpnmr.
> My experiments were recorded in an interleaved manner so I would expect that the spectral noise should be the same in all matrices within the same experiment.
> But, when I have a look into the spectrum window, noise levels seem to be correlated with peak intensity.
> For instance, in my hetnoe experiment, due to the dynamic nature of roughly half of my protein, ca 50% of the peaks in the saturated spectrum are negative. This seems to result in a noise level that is half as big as the noise level in the reference experiment, or, both experiments have been acquired at the same time.
> Same is true for the R1 series in which the noise decreases with the increase of the relaxation delay.
>
> What I found on this subject:
> ccpnmr.analysis.core.ExperimentBasic.getNoiseEstimate(dataSource, nsamples=1000, nsubsets=10, fraction=0.10000000000000001)
>
> Estimate the noise level for a spectrum by choosing a random nsamples points and finding subsets with the lowest standars devation/
>
> Wouldn't it be better to calculate noise as for instance in nmrview by defining a region without peaks before performing this operation ?
>
> Cheers
>
> Beate
>
|