In addition to the size (100k), the audio has to comply with the following characteristics:
Sampling rate of 8000, number of bits 16, and mono channel.
To give the appropriate values to the file, you can use the Audacity software, to which you can load a file and export it with the values you want:
Once the file has been loaded into the camera, all that remains is to select the audio that has been loaded into the camera, since you could load several and save: