How to edit audios so that they have the same length

Question

I have a dataset of audio files, which differ on their lengths by a couple of milliseconds. Some are little shorter than 1s and some are little longer than 1s. They all have the same sampling rate, namely 20kHz, what differs is the number of samples.

Does anyone know a way to automatically transform them into 1s audio files with 20kHz sampling rate?

EDIT: I need to transform the audios to having the same length so that I can input them into a standard fully-connected neural network.

Cutting the audio in the beginning and at the end would be the best solution, I think, if the audio is longer than 1 second. For example, if the audio is longer 1.1s, then I would like to cut the audio by 5ms in the beginning and the end to transform it to a 1s audio? And if the data is shorter, somehow putting some "silence" at the end of the data?

Thank you in advance!

Answer 1

ImageMagick is a great tool to edit multiple images in the command line. Thanks to your question, I just discovered SoX, or Sound eXchange:

SoX is a cross-platform (Windows, Linux, MacOS X, etc.) command line utility that can convert various formats of computer audio files in to other formats. It can also apply various effects to these sound files, and, as an added bonus, SoX can play and record audio files on most platforms.

It can trim, crop, cut multiple files (and resample too), see for instance:

• Editing Multiple Mono Sound Files in SoX,
• Audio format conversion cheat sheet (aka how to)

For shorter files, an option is to create a silence file, happen it, and crop again, see for instance: Linux command to extend the duration of audio files.

Answer 2

In order to make your audio wav file longer you can use sox in linux :

sox input.wav output.wav trim 0 00:01

This will take 1s of the audio input and put it in output.wav. If it's shorter than 1s it will probably work, but I have no idea.

Another solution would be to actually do what you want with a C/whatever program, that won't be very hard to do but you probably don't want to go that way.

I guess Sox has probably a simple and efficient solution to your problem.

Answer 3

I wrote this code in python, it’s not yet optimized but still it worked perfectly for me

from pydub import AudioSegment
from pydub.utils import mediainfo


in_wav = "input.wav"
out_wav = "output file"

with contextlib.closing(wave.open(in_wav,'r')) as f:
            frames = f.getnframes()
            rate = f.getframerate()
            duration = int(((frames / float(rate)))*1000) # input file duration
            
            if(duration < 1000):  #if its less then 1s
                newd=1000-duration
                if(newd%2==0):
                    pair=0
                else:
                    pair=1
                
                durbeg=(newd)/2+pair # the time of silence in the begging
                
                durend=(newd)/2 # the time of silence in the end
                
                
                # create silence audio segment of begging
                beg_sec_segment = AudioSegment.silent(duration=durbeg)  #duration in milliseconds
                # create of silence audio segment of end
                end_sec_segment = AudioSegment.silent(duration=durend)  #duration in milliseconds
                
                song = AudioSegment.from_wav(in_wav)
                
                
                #Add above two audio segments    
                final_song = beg_sec_segment + song + end_sec_segment
                
                final_song.export(out_wav, format="wav")