I want to build a simple program with a simple GUI to help my company with an audio/music processing task.
The program should work on Mac (important) or Windows (secondary), and should work properly with any major type of audio file, in stereo or mono.
It will be given a folder of stereo audio files, and should create a subfolder within to work on all of them.
The program should be able to recognize a short set of related sounds, based on a selection of 20-30 samples I can provide. It can use any machine learning methodology to do so, but it should be fairly quick.
It should search for this sound within the first 3 seconds of the piece.
If it finds the sound, it should search for the musical start of the piece after the sound, and go back .75 seconds and delete everything before that, removing the sound in the process.
If it doesn't find the sound it should search for the musical start after the first sample, and go back and delete .75 seconds before that, or add silence so that there's .75 seconds before the first impulse. (With these two steps we are trying to get about .75 to .5 seconds of plain silence before the start of the song more or less, I imagine this is the second hardest part)
It should cycle through the files and find the integrated loudness (LUFS value) of all of them, taking note of the quietest integrated loudness among them.
It should then go back and normalize all of them to the quietest integrated loudness value of the batch, with a true peak limit of .5 dbTP.
The GUI should let us drag a folder of files into it, and output the processed files into a subfolder in that folder. It should tell us what LUFS value was used to normalize, and what was done to each file.
IMPORTANT: With your response, please leave an estimate of the number of hours you would need to complete the project so I can estimate overall budget. Thank you.
This job is already closed and no longer accepting applicants, sorry.
