Eventually I realised it was nothing to do with the source WAV, and that Replay Gain was the culprit. Replay Gain is an admirable attempt to tackle the problem of the different perceived loudness in mp3s from different sources. It is basically an algorithm to estimate the perceived loudness of a track, and then store an offset against a reference level in the track meta-data. Its a big improvement on the awful peak normalisation that some mp3 players were attempting a few years ago, but its still not perfect.
When you make your own music, you go to considerable trouble to master the tracks to get the perceived loudness to the level that you want. You dont really want some little algorithm suddenly sneaking in and resetting the levels. Admittedly, Replay Gain only adds meta-data and doesn't actually touch the music, but when its all done without your consent and then your mp3 player honours the Replay Gain tags without you realising, its equivalent to someone sneaking into your studio and messing around with the master volume knob while you're looking the other way.
Perhaps I was particularly unlucky in this case because my noisy track was getting given a Replay Gain of -9.5 db, a massive difference.
MediaMonkey is a pretty useful tool, even in its Free version, but I couldn't find any options in it to turn off the Replay Gain interference. Basically, any mp3 tag editing resulted in Replay Gain being slapped on the track.
Using this I was able to clean out the unwanted Replay Gain meta-data that had sneaked in. I also experimented with setting the Replay Gain to 0.0 or -0.l db to see if that stopped MediaMonkey from interfering, but it didn't. However I cleared or set the Replay Gain, MediaMonkey would re-set it to its own calculated value on every tag edit.
No comments:
Post a Comment