Normalizing with FFmpeg is a two-step process. First, you need to use the volumedetect
filter which will tell you exactly how much dB you're allowed to crank up :
ffmpeg.exe -i "D:\Project\AC3.ac3" -ac 2 -af volumedetect -y NUL
This will show you the maximum volume of the stereo (downmixed) track along with additional information :
[Parsed_volumedetect_0 @ 0x20fb060] n_samples: 155043840 [Parsed_volumedetect_0 @ 0x20fb060] mean_volume: -26.5 dB [Parsed_volumedetect_0 @ 0x20fb060] max_volume: -3.2 dB [Parsed_volumedetect_0 @ 0x20fb060] histogram_3db: 23 [Parsed_volumedetect_0 @ 0x20fb060] histogram_4db: 87 [Parsed_volumedetect_0 @ 0x20fb060] histogram_5db: 672 [Parsed_volumedetect_0 @ 0x20fb060] histogram_6db: 2157 [Parsed_volumedetect_0 @ 0x20fb060] histogram_7db: 5848 [Parsed_volumedetect_0 @ 0x20fb060] histogram_8db: 15951 [Parsed_volumedetect_0 @ 0x20fb060] histogram_9db: 36078 [Parsed_volumedetect_0 @ 0x20fb060] histogram_10db: 73237 [Parsed_volumedetect_0 @ 0x20fb060] histogram_11db: 138626
And then you can normalize your track :
ffmpeg.exe -i "D:\Project\AC3.ac3" -ac 2 -af volume=3.2dB "D:\Project\WAV.wav"