How to trim, the extra audio portion after audio processing in two channel ?

Hi ,

i have a problem in audio processing . I have trasmitted audio in one hardware did some processing and reconstructed the audio in receiver . But during recostruiction some exctra rows got added in the both the channels , the adiitions are different in different channels . I need to remove the same for comapring . Please suggest a method .

I did it like this .

desiredFs=48000; 

audio_frame_CH1_FF=0;
audio_frame_CH2_FF=0;

[x,OriginalFs] = audioread("file.mp3"); % y samples from audio with Fs sampling frequency in [Hz].
%
N = length(x); % sample lenth
duration_audio=N/OriginalFs;
tx = linspace(0, N/OriginalFs, N);
plot(tx, x);
hold on 

afr = dsp.AudioFileReader('file.mp3','ReadRange',[1 8.007*desiredFs]);% value of duration_audio 
framecnt=0;

while ~isDone(afr)
     audio_frame = afr();
     audio_frame_CH1=audio_frame(:,1);
      audio_frame_CH2=audio_frame(:,2);
     
     audio_frame_CH1_FF=[audio_frame_CH1_FF ; audio_frame_CH1];
     audio_frame_CH2_FF=[audio_frame_CH1_FF ; audio_frame_CH2];
        
     
     framecnt=framecnt+1;
end 
release(afr);

fixedAudioFrame1=Audiotrim(x,audio_frame_CH1_FF);
fixedAudioFrame2=Audiotrim(x,audio_frame_CH2_FF);

fixedAudioFrame=[audio_frame_CH1_FF ; audio_frame_CH2_FF] 

plot(fixedAudioFrame,tx);
hold off

audiowrite('processed.wav',fixedAudioFrame,desiredFs);

The code of Audiotrim is like this

function trimmedAudio=Audiotrim(x,audio_frame_CH1_FF)

    slength =length(x(:,1))
    elength =length(audio_frame_CH1_FF(:,1))
    
    extraRow =elength-slength
    i=1;
    %for i=1:floor(extraRow/2)
    while i< floor(extraRow/2)
        
        if  i== floor(extraRow/2)
            i=1;
        end 
        
        if (length(audio_frame_CH1_FF(:,1))==length(x(:,1)))
           break;
        end 
            audio_frame_CH1_FF(slength+i,:) = [];
       
    end
    trimmedAudio=audio_frame_CH1_FF;
end 

Issue is. Due to the different rows the thsi function does not work in both the channel . I know thios method is not the right approach . Please suggest a better approach towards this problem.

Answer 1

[Abhimenyu]

Hi Mainak,

From the information shared, I could infer that you're trying to align the lengths of the audio frames from two channels to compare them properly. The signal processing technique of "cross-correlation" would help to align the audio frames using the time lag between them rather than manually adjusting the lengths. The initial code that loads and reads the files will remain the same as given below in the MATLAB code:

% Load your audio file (replace "file.mp3" with your actual file path)
[x, OriginalFs] = audioread("file.mp3");

% Desired sampling frequency
desiredFs = 48000;

% Read audio frames
afr = dsp.AudioFileReader('file.mp3', 'ReadRange', [1 8.007 * desiredFs]);
framecnt = 0;
audio_frame_CH1_FF = [];
audio_frame_CH2_FF = [];

while ~isDone(afr)
    audio_frame = afr();
    audio_frame_CH1 = audio_frame(:, 1);
    audio_frame_CH2 = audio_frame(:, 2);
    
    audio_frame_CH1_FF = [audio_frame_CH1_FF; audio_frame_CH1];
    audio_frame_CH2_FF = [audio_frame_CH2_FF; audio_frame_CH2];
    
    framecnt = framecnt + 1;
end

release(afr);

Then we use the MATLAB function "xcorr" to find the time delay between the two channels. The peak of the cross-correlation gives the point of highest similarity, indicating how much one signal is delayed relative to the other.

% Cross-correlation to find time delay (lag) between channels
[acor, lag] = xcorr(audio_frame_CH1_FF, audio_frame_CH2_FF);
[~, I] = max(abs(acor));
lagDiff = lag(I);

With the lag determined, the starting point of each channel is adjusted to align them in time. This step is crucial for ensuring that any further processing or analysis on the channels considers their correct temporal relationship.

% Align the channels
if lagDiff > 0
    audio_frame_CH2_aligned = audio_frame_CH2_FF(lagDiff + 1:end);
    audio_frame_CH1_aligned = audio_frame_CH1_FF(1:end - lagDiff);
else
    audio_frame_CH1_aligned = audio_frame_CH1_FF(-lagDiff + 1:end);
    audio_frame_CH2_aligned = audio_frame_CH2_FF(1:end + lagDiff);
end

After alignment, the signals are trimmed to the length of the shorter one to ensure they have the same duration. This is important for stereo playback and any comparative analysis.

% Trim to match lengths
minLength = min(length(audio_frame_CH1_aligned), length(audio_frame_CH2_aligned));
audio_frame_CH1_aligned = audio_frame_CH1_aligned(1:minLength);
audio_frame_CH2_aligned = audio_frame_CH2_aligned(1:minLength);

Finally, the aligned and trimmed audio is saved back to a file, ready for playback or further processing.

% Now you can compare or process the aligned channels
% For example, save the processed audio
audiowrite('processed.wav', [audio_frame_CH1_aligned, audio_frame_CH2_aligned], desiredFs);

For a better understanding of the "xcorr" function, please follow this MATLAB R2024A documentation link: https://www.mathworks.com/help/matlab/ref/xcorr.html

I hope this helps!