QUOTE
Originally posted by Cygnus X1
That is possible, or the person who encoded the file may have purposely downsampled from 44.1 or 48Khz when encoding to kill off some of the high-end (anything above 16Khz) as to obtain less artifacting, being that the sample was only encoded at 128kbps. Back in the old days I used to downsample .wav files to 32Khz before encoding, since the mp3 encoders of the day performed quite poorly at 128kbps, especially with certain HF material.
Down sampling to 32 kHz should only be used at 96 kbps and below.
1. Most encoders, so also lame, are nearly only tuned with 44.1 kHz
2. Short blocks enlarge from 8.7 ms to 12 ms with 32 kHz. Optimal would be values around 5...6 ms for typical Top 40 music, for pathological music around 2 ms.