unable to plot semilogy properly

hello everyone, I have a code that simulates QAM modulation below:

% QPSK
M = 8;
k = log2(M); % Bits per symbol
EbNoVec = (5:15); % Eb/No values (dB)
numSymPerFrame = 1000; % Number of PSK symbols per frame

berEst = zeros(size(EbNoVec));
for n = 1:length(EbNoVec)
    % Convert Eb/No to SNR
    snrdB = EbNoVec(n) + 10*log10(k);
    % Reset the error and bit counters
    numErrs = 0;
    numBits = 0;
    
    while numErrs < 200 && numBits < 1e7
        % Generate binary data and convert to symbols
        data_in = randi([0 1],numSymPerFrame*k,1);
        
        %----------------------------------------------------------------    
        % Your modulator here:
        modSig = pskmod(data_in,M,InputType='bit'); % PSK Modulation 
        % Pass through AWGN channel:
        rxSig = awgn(modSig, snrdB); % Additive White Gaussian Noise with 
                                     % increasing snrdB
        % Your demodulator here:
        data_out = pskdemod(rxSig,M,OutputType='bit'); % PSK Demodulation 
        
        %----------------------------------------------------------------
 
        % Calculate the number of bit errors
        nErrors = biterr(data_in,data_out);
        numErrs = numErrs + nErrors;
        numBits = numBits + numSymPerFrame*k;

    end
    % Estimate the BER
    berEst(n) = numErrs/numBits;
    
    fprintf("snrdB: %.6f berEst: %.6f " + ...
        " \n \t numErrs: %d numBits: %d\n\n", snrdB, berEst(n), numErrs, numBits);
    
end

fitEbN0 = EbNoVec(1):0.25:EbNoVec(end); % Interpolation values
berfit(EbNoVec,berEst,fitEbN0);
hold on;
for n=1:length(EbNoVec)
    semilogy([EbNoVec(n) EbNoVec(n)],'g-+');
end
hold off;

the code runs well, but unfortunately the semilogy looks like this :

the 'g-+' does not show up on the plot...do you know how to resolve this issue?? any response regarding this problem is really appreciated..thx

Answer 1

[Star Strider]

The easiest way to solve that is to plot each curve on its own y-axis using the yyaxis function —

% QPSK
M = 8;
k = log2(M); % Bits per symbol
EbNoVec = (5:15); % Eb/No values (dB)
numSymPerFrame = 1000; % Number of PSK symbols per frame

berEst = zeros(size(EbNoVec));
for n = 1:length(EbNoVec)
    % Convert Eb/No to SNR
    snrdB = EbNoVec(n) + 10*log10(k);
    % Reset the error and bit counters
    numErrs = 0;
    numBits = 0;
    
    while numErrs < 200 && numBits < 1e7
        % Generate binary data and convert to symbols
        data_in = randi([0 1],numSymPerFrame*k,1);
        
        %----------------------------------------------------------------    
        % Your modulator here:
        modSig = pskmod(data_in,M,InputType='bit'); % PSK Modulation 
        % Pass through AWGN channel:
        rxSig = awgn(modSig, snrdB); % Additive White Gaussian Noise with 
                                     % increasing snrdB
        % Your demodulator here:
        data_out = pskdemod(rxSig,M,OutputType='bit'); % PSK Demodulation 
        
        %----------------------------------------------------------------
 
        % Calculate the number of bit errors
        nErrors = biterr(data_in,data_out);
        numErrs = numErrs + nErrors;
        numBits = numBits + numSymPerFrame*k;

    end
    % Estimate the BER
    berEst(n) = numErrs/numBits;
    
    fprintf("snrdB: %.6f berEst: %.6f " + ...
        " \n \t numErrs: %d numBits: %d\n\n", snrdB, berEst(n), numErrs, numBits);
    
end

snrdB: 9.771213 berEst: 0.030222  
 	 numErrs: 272 numBits: 9000

snrdB: 10.771213 berEst: 0.019667  
 	 numErrs: 236 numBits: 12000

snrdB: 11.771213 berEst: 0.012611  
 	 numErrs: 227 numBits: 18000

snrdB: 12.771213 berEst: 0.006867  
 	 numErrs: 206 numBits: 30000

snrdB: 13.771213 berEst: 0.002806  
 	 numErrs: 202 numBits: 72000

snrdB: 14.771213 berEst: 0.000893  
 	 numErrs: 201 numBits: 225000

snrdB: 15.771213 berEst: 0.000333  
 	 numErrs: 200 numBits: 600000

snrdB: 16.771213 berEst: 0.000063  
 	 numErrs: 200 numBits: 3198000

snrdB: 17.771213 berEst: 0.000008  
 	 numErrs: 84 numBits: 10002000

snrdB: 18.771213 berEst: 0.000001  
 	 numErrs: 11 numBits: 10002000

snrdB: 19.771213 berEst: 0.000000  
 	 numErrs: 2 numBits: 10002000

fitEbN0 = EbNoVec(1):0.25:EbNoVec(end); % Interpolation values

figure
yyaxis left
berfit(EbNoVec,berEst,fitEbN0);
yyaxis right
semilogy(EbNoVec, EbNoVec,'g-+')

figure
berfit(EbNoVec,berEst,fitEbN0);
hlgd = findobj(gcf, 'Type','legend');
hlgd.AutoUpdate = 'on';
hold on;
semilogy(EbNoVec, EbNoVec,'g-+', 'DisplayName','EbNoVec')
hold off;
ylim([min([berEst EbNoVec])  max([berEst EbNoVec])] .* [0.1 10])

get(hlgd)                                                           % Show Changed Legend Properties

            AutoUpdate: on
          BeingDeleted: off
                   Box: on
            BusyAction: 'queue'
         ButtonDownFcn: @bdowncb
              Children: [0×0 GraphicsPlaceholder]
                 Color: [1 1 1]
           ContextMenu: [1×1 ContextMenu]
             CreateFcn: ''
             DeleteFcn: ''
             Direction: 'normal'
             EdgeColor: [0.1500 0.1500 0.1500]
             FontAngle: 'normal'
              FontName: 'Helvetica'
              FontSize: 9
            FontWeight: 'normal'
      HandleVisibility: 'on'
               HitTest: on
           Interpreter: 'tex'
         Interruptible: off
            ItemHitFcn: @defaultItemHitCallback
                Layout: [0×0 matlab.ui.layout.LayoutOptions]
             LineWidth: 0.5000
              Location: 'southwest'
            NumColumns: 1
        NumColumnsMode: 'auto'
           Orientation: 'vertical'
                Parent: [1×1 Figure]
         PickableParts: 'visible'
              Position: [0.1490 0.1363 0.2413 0.1155]
              Selected: off
    SelectionHighlight: on
                String: {'Empirical BER'  'Exp Fit'  'EbNoVec'}
                   Tag: 'legend'
             TextColor: [0 0 0]
                 Title: [1×1 Text]
                  Type: 'legend'
                 Units: 'normalized'
              UserData: []
               Visible: on

To plot them on the same axes without using yyaxis, set the ylim limits to accommodate both sets of data. (There must be something about the berfit function that does not allow this to be the default.)

EDIT — (2 Mar 2024 at 14:47)

Added second plot and explanation.

EDIT — (2 Mar 2024 at 15:47)

Changed legend.

.

[Abdul Hamzah]

thank you for the response...however, which plot is right on this case?

[Star Strider]

As always, my pleasure!

I am not absolutely certain because I am not familiar with your analysis. However I would choose the last one (plotting both curves and associated data on the same axes without using yyaxis).

[Abdul Hamzah]

and how can I change the legend of the plot that is made from the this line here ? berfit(EbNoVec,berEst,fitEbN0);

Btw the plot made by berfit is the right one (if that is what you meant by the last plot)..but do you know how to change its legend?

[Star Strider]

Yes.

First, I got its handle with:

hlgd = findobj(gcf, 'Type','legend');

then I changed it to turn 'AutoUpdate' to 'on' and that added ‘EbNoVec’ to it, which is what I wanted to do. (You can change it to not do that by turning 'AutoUpdate' to 'off'.)

I am not certain what else you want to change, however that should also serve as an example of how to change its properties.

The:

get(hlgd)

function call displays the legend properties. Comment that line (or delete it) to prevent the properties from displaying.

.