Something strange (and worrying) in the Winamp 2 plugin.
Brief description of problem:
Very sharp stabs of metallic-sounding noise. It sounds very much like a warped sibliant. At normal to high listening levels using headphones it is loud and unpleasant enough to make me fear for my hearing. Winamp's built-in visualization reacts to the noise, causing a couple of the bars to the right of the spectrum analyzer to jump at least two thirds of the way up.
When it occurs:
The problem only occurs when using the Winamp progression slider immediately after the track begins to play. At first I thought it was random, but I have since learned to provoke it on at least two different files. It seems to be directly related to 1) how "far" you seek" and 2) the program material. It always occurs on the same exact spot, after seeking to approximately the same spot.
Example:
On David Bowie's track "Sunday", first off the album "Heathen": if I grab the slider within the first five seconds of playback, and seek to at least 0:45, the problem will invariably occur at 0:58. Same exact spot; always the "L" in the line that goes "Look for the...". Seeking further into the song, and then seeking back to about 0:45-0:55 will produce the same results.
The other track is completely different material, Schubert's "Death and the Maiden", a Naxos recording from 1992. The problem does not occur at the same time, but can be provoked in a similar fashion.
Details:
I use Winamp 2.81, with in_mpc versions 0.96 and 0.97f. They behave identically. I use Clip Protect and ReplayGain in title mode, with Dithering enabled. No EQ, and no DSP plugins (although I've verified that the problem persists with DSP plugins running). The tracks in question were encoded using 1.14 with --xtreme and --xlevel. I've had the problem occur on files produced with earlier encoders as well, but haven't been able to verify them yet.
/ Uosdwis