Forensic audio file

Goodnight everyone, I need to prove that two video files have audio with levels (dB) different and show how they differ. I am new in forensic analysis. I removed the audio from the video and saved in two of the .wma file type. I tried to analyze some graphics, but as I do not know much about audio, I do not know what graphics use. Could anyone give me an idea of how to do this analysis, or where I can start? Thanks for any help.

Audacity doesn’t do forensics particularly well because it’s a sound editor, not a WAV editor. If it’s a choice between making an audio file mathematically perfect or making it sound good. We’ll usually go with sound good. This drives the scientists nuts.

“The bit structure four seconds in doesn’t match the original!”
“Maybe not, but it sounds better.”

You may have hosed yourself right at the top with WMA. Windows Media Audio is a compressed sound format. That means it makes judgement calls whether or not an audio file is going to sound good, and changes the data around until it does. That’s how MP3 works.

So they’re different now even before the content you started out with.

Never do critical audio management in a compressed format.


Try comparing the spectrogram views of the sounds. [ You will probably have to increase the “window size” of the spectrogram from the default setting of 256 to 1024 to obtain a high-enough resolution display].

I’d export sample values to Excel and use Excel’s graphing abilities…

Or is that a bad idea?

Thank Koz. I did not express myself very well. I have two files in the same format and I just wanted to show the difference in sound level between them I do not need to make any more detailed check, for example, voice recognition, etc. Actually I just need to prove that the audio1.wav file has a sound level much higher than the audio2.wav file. As both files were obtained from the same source and with the same characteristics, I thought to graphically show the two levels and prove that difference. Hello Trebor really thought about using the spectrogram, currently my biggest difficulty lies in the concepts involving audio, I’m trying to understand the workings of this type of chart. I also thought of using frequency analysis chart after understanding it well, I think I have a lot of great information in it. Cyrano is a good idea too, but do not know how to export audio to an array of numbers and use in Excel, and even I can do it think it would have a lot of numbers and do not know if excel bear a lot of numbers to analyze. Guys, thanks for the opinions and so I have a conclusion, I share with you. Good afternoon everyone.

Could this be as simple as viewing the two files one above the other on the timeline and one set of blue waves is bigger than the other? There is also an AudioBook loudness test. AudioBooks are required to maintain a certain loudness, so it’s necessary to measure it.

Select the whole clip by clicking just above MUTE.
Analyze > Contrast > Foreground > Measure Selection (Read the number) > Close

RMS values roughly correspond to loudness. Do that to each track or file.


Thanks for the tip koz, I will perform the suggested procedures and get the results as soon as I give feedback for everyone.