Having loudness differences within an album is an important part of the artistic choices the artist/producer have to make. It is quite rare that all songs in an album has the same structure, arrangement and energy level, so you don't want all your songs to have the same average loudness.Įven if 2 songs are similar in energy in the main sections (verses and choruses), if song A has a long, soft intro and song B has a full-on arrangement throughout, the main sections in song A would be significantly louder if both were normalized to a target LUFSi. Generally, they want a slow song quieter than an upbeat song. When I master an album, I make sure that the loudness differences between each song are exactly as intended by the artist/producer.
but the normalization itself is not changing anything but the level. It bears mentioning that some plugins respond very differently depending on the level of the input signal, especially plugins that model analog units (where -18 dBFS / 0 VU is the expected input level), so normalizing a file could have significant sonic effect due to the plugins. RMS normalization operates a little differently, it's looking at average and/or perceived level rather than peak, but the goal is the same - raise or lower the level of the entire file uniformly, to hit a target.ġ. 1įor peak normalization, a target is determined (typically, but not always, 0dBFS), and then the level is raised or lowered so that the loudest sample (the peak) hits the target. And the only thing that is changed is the level - dynamics, transients, etc. Normalization is nothing more than a programatic level adjustment. Given the rampant misuse and misunderstanding of the term in this thread alone, I think we need to be more specific about what normalization is actually doing to achieve that effect. That is one use of normalization (RMS normalization to be specific), but it's not actually what it is. Normalization is when you get multiple songs to sound the same volume so when one song leads to another the listener doesn't need to adjust their volume.
These normalized versions of songs for DJing are kept in a separate library away from the originals. I specifically use it to lower the volume of most tracks and rebuild the peaks that were lost to limiters during loudness wars. I use a somewhat advanced normalization algorithm and application called Platinum Notes which is specifically designed for DJs. That being said: I do normalize tracks before DJing them when I DJ songs.
The mastering process usually has a compressor and a limiter - both of them are used kind of like a normalizing type function. Or sometimes I'll use the samplers built in functions to re-sample a normalized version alongside the regular version. I might make a normalized version of a sample before loading it in a sample player for example. The only time I find myself using the actual Normalizing function is usually dealing with some sample or repair. Normalizing as a process is a "destructive edit" meaning you can't easily "go back" to how it was before so most people choose to use a non-destructive method like a compressor instead. They can and do go everywhere from individual tracks, to busses, to the master.Ī compressor is a form of normalization more commonly used these days. Most people do some form of normalizing (lower case as in the function itself) but use a different technique to produce a similar type of effect. Normalizing is rarely if ever done - that's with a capital "N" in normalizing as in the actual noun name and not the generic "normalizing function".
0 kommentar(er)
