Quote:
Originally posted by Stephonovich
You do that to albums you've ripped? Why? I could understand doing an album leveling, so all your albums are roughly the same volume, but the songs themselves the same volume? What about dynamic range in the music?
(-:Stephonovich |
And again, someone isn't READING what MP3Gain actually does.
The dynamics aren't touched at all, this isn't
normalization or limiting like everyone thinks it is. Look at it this way: If you're listening to some kind of high dynamic music, say some classical music with lots of low parts with wind instruments and then WHAM outta nowhere you get hit with a full on crescendo of the entire orchestra blasting away, do you reach for the volume control to turn it down at that point (or if you've heard the song before and you know what to expect so you begin turning it down at *just* the right time)?
If you do, then that's considered
limiting because you're "limiting" the volume of the playback but you're not "compressing" the dynamics, which are still there untouched in the original source.
Compressing/Limiting is what radio stations do to music they're broadcasting, they use devices called compressors and limiters (and no, it's not audio compression like mp3, it is simply keeping the constant audio playback at a specific level). Compressors are used to limit the dynamics of the music. The part about saying "Compressors" and having people understand I don't mean MP3/AAC/Ogg/WMA/etc audio compression can be a major sticking point. This kind of compression/limiting is compressing/reducing/limiting the dynamics and playback levels - audio compression is primarily about smaller file sizes. Tp put it bluntly, compressors control the audio dyamics, limiters control the audio volume. Please note dynamics and volume are two different things yet they are both parts of the final result: audio output.
So using my example above, you didn't compress the music's dynamic range, you just limited it by lowering the volume to account for and negate the effect of the crescendo. The perceived volume of music would be staying roughly the same.
If the actual audio level of a specific section of music was 4 on a scale and yet you boosted it to 6, and the crescendo rated an 8 and you lowered it to 6, you just limited the audio level from the highs and lows to a nice even setting.
Here's a page that gets pretty technical but it might help:
How to best use Compression (not the mp3 kind)
MP3Gain doesn't do compressor/limiter functions. It merely takes an audio source file and performs mathematical operations on it based on psychoacoustic models that compute the perceived level of audio playback and modifies the files to play at a specified level (as defined by db in the settings, 89db is pretty much the defacto standard for audio production).
Hope this helps,
br0adband
ps
As explained in the MP3Gain help files, I could do something to the following procedure and still not affect the original mp3 file:
1) Make an MP3 of my choice (EAC + LAME 3.90.3 would do it) and make two copies of it, one for control and one for modification and testing..
2) Convert that test file to an 89db level with MP3Gain.
3) Convert that test MP3 to a WAV file while retaining the pre-WAV MP3.
4) Take that testMP3 and process it with MP3Gain 100 times, each time choosing a random db level, like 95db the first run, 50db the second run, etc etc.
5) After 100 processing runs, set the level back at the original 89db level.
6) Take that test mp3 file and convert it to another WAV file (not overwriting the first WAV, of course)
7) Compare the two WAV files.
Theoretically, the two WAV files *should* match, bit for bit.
I might do this sometime just for fun.
