Different length of sound files with different sampling frequencies

Dear all,

Im currently struggling to understand what is happening. So, I created a sound using the audiowrite function (the sound is created using two different sounds but I dont think it matters) first with a sampling frequency of 44100 Hz, and another one, the sound file is the same but the sampling frequency is 48000 Hz. Now Im observing that the sound produced at 44100Hz is approx. 30sec longer than the other one (48000Hz sampling). It looks like phase shifting of some sort, but Im not sure. Any help/explanation is appreciated. I also made a amplitude/time plot for better understanding below

[Stephen23]

It looks as if you did not plot against the time (i.e. X-data). If you only supply the Y-data then MATLAB will simply plot the Y-data against their indices.

A similar effect would also occur if you use the same X-data for both of the Y-data, even though they were sampled at different times.

Basically there are multiple reasons why this might happen, and we don't know which one applies to your situation because you have not uploaded any code for us to review.

[Denis Benka]

I intentionally set the x axis to 350sec to see where the 44100Hz audio signal ends, but my question is still why is there a difference between those two frequencies, what is causing the different lengths

[Denis Benka]

Here is the audio processing code Im currently using (both of the sound files (cave,forest) have a sampling frequency of 48000Hz):

[y1,F1] = audioread(cave_file);        %cave and forest files are mp3 files loaded earlier
[y2,F2] = audioread(forest_file);

samp_freq=44100;
%samp_freq=48000;                    
a = max(size(y1),size(y2));
z = [[y1;zeros(abs([a(1),0]-size(y1)))],[y2;zeros(abs([a(1),0]-size(y2)))]]

audiowrite('test_sound.wav', z,samp_freq);

[Jan]

@Denis: All we see is this diagram, but we do not now, how you import the data and create the diagram. Obviously the corresponding code contains a bug. We can find it only, if we see this code, of course.

[Stephen23]

"why is there a difference between those two frequencies, what is causing the different lengths"

If you sample something for the same length of time at two different sample rates, then of course the number of samples in the resulting data arrays will be different. That is by the very definition of "sampling rate". But note that the data "length" is not directly relevant to the "phase shift", which is what you complain about in your question. The "phase shift" I explained in my last comment (i.e. incorrect plot syntax or X-values). My explanation matches your plotted data quite well:

>> (44100*330)/48000
ans =  303.19

The X-data seem to be incorrect for one/both of those Y-data arrays, possibly because you plotted against their indices rather than aginst their times, or because of their times.

[Geoff Hayes]

Denis - if the same number of samples (from your z) are used to create each sound with the only difference being the sampling frequency, then the duration (in seconds) of each sound will be length(z)/44100 and length(z)/48000. It is clear that the former will be longer than the latter.

[Denis Benka]

@Stephen, yes dont mind the figure, Im just wondering how (or if) the sampling frequency is affecting the length of the sound file (in seconds). But the size (in kB) of both files seems equal even though the lengths differ, which is also kinda odd to me

[Denis Benka]

@Stephen, @Geoff, I think I finally understand what is happening. Those calculations are exactly what I wanted to know, shame on me I forgot about these. Thanks a lot!