Unexpected beaviour when plotting narrow "peaks"

#1

Hello all,

I'm sorry if this has been treated before, but I haven't found
anything when searching the archives or the net.

Basically, the problem surfaces when plotting FFT spectras using many
data points. The amplitudes of the peaks in the spectra then seem to
depend on the size of the plot window.

Example code:

from pylab import *
t=arange(65536)
plot(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))

[<matplotlib.lines.Line2D object at 0x03871990>]

show()

With the original window size (WXAgg backend), I get maximum peaks at
about 390. Zooming in on the peak will increase the amplitude.
Resizing the window can also make the amplitude increase.

The expected amplitude is:

max(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))

12891.96683092583

Is this a known problem, or should I file a bug report? Is there any
workaround (except decreasing the length of t, and thus the level of
detail)?

I'm using ActivePython 2.6.3.7 for windows, matplotlib 0.99.1 binary,
numpy 1.3.0 and wxPython-unicode-2.8.10.1. I've also tried with the
TkAgg backend, which gives similar results.

Thanks,
Henrik

#2

It looks like the path simplification code is failing on this. As a workaround, you can turn it off:

  rcParams['path.simplify'] = False

or reduce the threshold below which vertices are removed:

  rcParams['path.simplify_threshold'] = 0.0001

In the meantime, I'll look further into why this is failing. Perhaps we need a lower default threshold.

Mike

Henrik Wallin wrote:

···

Hello all,

I'm sorry if this has been treated before, but I haven't found
anything when searching the archives or the net.

Basically, the problem surfaces when plotting FFT spectras using many
data points. The amplitudes of the peaks in the spectra then seem to
depend on the size of the plot window.

Example code:

from pylab import *
t=arange(65536)
plot(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))
        

[<matplotlib.lines.Line2D object at 0x03871990>]
  

show()
        
With the original window size (WXAgg backend), I get maximum peaks at
about 390. Zooming in on the peak will increase the amplitude.
Resizing the window can also make the amplitude increase.

The expected amplitude is:

max(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))
        

12891.96683092583

Is this a known problem, or should I file a bug report? Is there any
workaround (except decreasing the length of t, and thus the level of
detail)?

I'm using ActivePython 2.6.3.7 for windows, matplotlib 0.99.1 binary,
numpy 1.3.0 and wxPython-unicode-2.8.10.1. I've also tried with the
TkAgg backend, which gives similar results.

Thanks,
Henrik

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users
  
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

#3

I have fixed this bug that was causing peaks to be truncated in SVN r7895 and r7896. It should make it into the next bugfix release.

Mike

Michael Droettboom wrote:

···

It looks like the path simplification code is failing on this. As a workaround, you can turn it off:

  rcParams['path.simplify'] = False

or reduce the threshold below which vertices are removed:

  rcParams['path.simplify_threshold'] = 0.0001

In the meantime, I'll look further into why this is failing. Perhaps we need a lower default threshold.

Mike

Henrik Wallin wrote:
  

Hello all,

I'm sorry if this has been treated before, but I haven't found
anything when searching the archives or the net.

Basically, the problem surfaces when plotting FFT spectras using many
data points. The amplitudes of the peaks in the spectra then seem to
depend on the size of the plot window.

Example code:

from pylab import *
t=arange(65536)
plot(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))
        

[<matplotlib.lines.Line2D object at 0x03871990>]
  

show()
        

With the original window size (WXAgg backend), I get maximum peaks at
about 390. Zooming in on the peak will increase the amplitude.
Resizing the window can also make the amplitude increase.

The expected amplitude is:

max(abs(fft(sin(2*pi*.01*t)*blackman(len(t)))))
        

12891.96683092583

Is this a known problem, or should I file a bug report? Is there any
workaround (except decreasing the length of t, and thus the level of
detail)?

I'm using ActivePython 2.6.3.7 for windows, matplotlib 0.99.1 binary,
numpy 1.3.0 and wxPython-unicode-2.8.10.1. I've also tried with the
TkAgg backend, which gives similar results.

Thanks,
Henrik

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users
  
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA