So your speech2 filename is the same as your speech filename.
[y,Fs]=audioread(speech2); %don't use wavread or audioread
So you read from coch4.wav via speech2 filename
audiowrite(speech, speech_trimmed, 44100);
and then you write to the file name given by speech, which is the same as the filename given by speech2, so you are writing to the same file that you read the data from.
Best Answer