matplotlib-users

[Matplotlib-users] matplotlib power spectral density (PSD) value discrepancy

[xkryptor <xkr...@gm...>]

Hi all,

 I tried calculating the power spectral density using matplotlib mlab
function psd (). I plotted using two methods:

 1. At first I plot it directly using plt.psd (red line in plot)
 2. Then I output the values from psd() to variables and plotting the
    variables (blue line in plot)

The code I used:

power, freqs = plt.psd(P * 100000, len(P), Fs = 1 / dt, scale_by_freq=0)
plt.psd(P * 100000, len(P), 1 / dt, scale_by_freq=0)
plt.plot(freqs, power)

But the plots are different, I expected it to be coincident. From where
does the discrepancy arise?

[image: Inline image 1]

Re: [Matplotlib-users] matplotlib power spectral density (PSD) value discrepancy

[Thomas C. <tca...@gm...>]

What version of mpl are you using? These functions got a lot of work
between 1.3 and 1.4.

I also suspects that there is a normalization issue going on with power vs
power density.

Tom

On Tue, Dec 16, 2014, 22:28 xkryptor <xkr...@gm...> wrote:

> Hi all,
>
>  I tried calculating the power spectral density using matplotlib mlab
> function psd (). I plotted using two methods:
>
>  1. At first I plot it directly using plt.psd (red line in plot)
>  2. Then I output the values from psd() to variables and plotting the
>     variables (blue line in plot)
>
> The code I used:
>
> power, freqs = plt.psd(P * 100000, len(P), Fs = 1 / dt, scale_by_freq=0)
> plt.psd(P * 100000, len(P), 1 / dt, scale_by_freq=0)
> plt.plot(freqs, power)
>
> But the plots are different, I expected it to be coincident. From where
> does the discrepancy arise?
>
> [image: Inline image 1]
> ------------------------------------------------------------
> ------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&
> iu=/4140/ostg.clktrk_______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Re: [Matplotlib-users] matplotlib power spectral density (PSD) value discrepancy

[Eric F. <ef...@ha...>]

On 2014/12/17, 5:37 AM, Thomas Caswell wrote:
> What version of mpl are you using? These functions got a lot of work
> between 1.3 and 1.4.

Tom,

Sorry, I answered only in his parallel github issue. 
https://github.com/matplotlib/matplotlib/issues/3927.  I'm closing it 
now.  The upshot is that psd returns the power but plots it in db.

Eric

>
> I also suspects that there is a normalization issue going on with power
> vs power density.
>
> Tom
>
> On Tue, Dec 16, 2014, 22:28 xkryptor <xkr...@gm...
> <mailto:xkr...@gm...>> wrote:
>
>     Hi all,
>
>       I tried calculating the power spectral density using matplotlib
>     mlab function psd (). I plotted using two methods:
>
>       1. At first I plot it directly using plt.psd (red line in plot)
>       2. Then I output the values from psd() to variables and plotting the
>          variables (blue line in plot)
>
>     The code I used:
>
>     power, freqs = plt.psd(P * 100000, len(P), Fs = 1 / dt, scale_by_freq=0)
>     plt.psd(P * 100000, len(P), 1 / dt, scale_by_freq=0)
>     plt.plot(freqs, power)
>
>     But the plots are different, I expected it to be coincident. From
>     where does the discrepancy arise?
>
>     Inline image 1
>     ------------------------------__------------------------------__------------------
>     Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
>     from Actuate! Instantly Supercharge Your Business Reports and Dashboards
>     with Interactivity, Sharing, Native Excel Exports, App Integration &
>     more
>     Get technology previously reserved for billion-dollar corporations, FREE
>     http://pubads.g.doubleclick.__net/gampad/clk?id=164703151&__iu=/4140/ostg.clktrk
>     <http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk>_________________________________________________
>     Matplotlib-users mailing list
>     Matplotlib-users@lists.__sourceforge.net
>     <mailto:Mat...@li...>
>     https://lists.sourceforge.net/__lists/listinfo/matplotlib-__users
>     <https://lists.sourceforge.net/lists/listinfo/matplotlib-users>
>
>
>
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>