ploting a logaritmic-distributed data

I am trying to plot data obtained from a spectrum analyzer. The x-scale of its measurements is in logaritmic. I would like to plot it as I see it in the instrument's display:

However, I obtain this:

when

    freqs = linspace(startFrequency, stopFrequency, num_of_points);
    semilogx(freqs, measurements)

, when

    freqs = linspace(startFrequency, stopFrequency, num_of_points);
    plot(freqs, measurements)

, when

    freqs = logspace(log10(startFrequency), log10(stopFrequency), num_of_points);
    plot(freqs, measurements)

and when

    freqs = logspace(log10(startFrequency), log10(stopFrequency), num_of_points);
    semilogx(freqs, measurements)

On the other hand, I obtain:

in any case of the ones above but with

    set(axes, 'xscale', 'log');

How am I supposed to do it?

Many thanks.

Edit: if the data was needed, I have attached it.

[dpb]

freqs = linspace(startFrequency, stopFrequency, num_of_points);
semilogx(freqs, measurements)
xlim([150E3 30E6])
ylim([34 64])

will give the same limits. Looks like perhaps the analyzer cut off the next bin at the LH side. Try stretching the actual axis length out to match; not sure why your plotted spectrum seems more compressed towards the higher frequencies; looks like same RH upper limit if I read the scale correctly.

Answer 1

[Lucas Santisteban]

Ok, there was an stupid solution I didn't try.

 freqs = logspace(log10(startFrequency), log10(stopFrequency), num_of_points);
 loglog(freqs, spectrum);
 set(gca, 'yscale', 'lin');

With this I get exacly what I wanted.

Thank you all anyway for your help.

Answer 2

[Star Strider]

I’m not certain what you want the plot to look like.

This is my best guess:

D = load('Lucas Santisteban spectrum.mat');
measurements = D.spectrum;
num_of_points = length(measurements);
startFrequency = 150E+3;
stopFrequency = 30E+6;
freqs = linspace(startFrequency, stopFrequency, num_of_points);

antilogfreqs = 10.^(freqs/startFrequency);

figure(1)
semilogx(antilogfreqs, measurements)
xt = get(gca, 'XTick');
xtl = linspace(startFrequency, stopFrequency, length(xt));
xtlf = regexp(sprintf('%.1f\\cdot10^%.0f\n', [10.^rem(log10(xtl),1)'  fix(log10(xtl))']'), '\n', 'split');
set(gca, 'XTick',xt, 'XTickLabel',xtlf(1:end-1), 'XMinorTick','on')
grid

This code begins by taking the antilog of the frequency vector to create the plot. The ‘xt’ assignment gets the x-tick values the plot call creates, ‘xtl’ creates a linear vector of the frequencies that simply re-samples the frequency vector, the xtlf call creates a formatted version, split into individual cells for the plot labels.

The result:

[Lucas Santisteban]

This is not exacly what I would like to achieve. With your approach, I only get a virtual solution. What I mean is that the real frequencies are not the ones in the Tick, so when you put a marker, the x value doesn't correspond with the actual frequency.

Many thanks anyway.

I can't understand how it can be so difficult, I just want to draw the data I obtain versus a log-spaced frequency and that the plot had a log grid.

Answer 3

[Lucas Santisteban]

Is there a function to do the antifunction of what I obtain in the last graph (the one where I set the scale to log) so the plot was finally correct?