Tilted Forum Project Discussion Community - View Single Post

mpedrummer2 · 08-08-2003, 02:30 PM

OK...I'm a recording engineer, so perhaps I can shed some light on the subject...

All normalization does is adjust the volume so that the current highest point in the song is equal to the level you set it at. For example, if the max in the song is 75%, and you set the normalization level to 95%, it'll turn everything up until the max level is 95%.

Now here's the catch...that maximum level in the song might only happen for a very short period of time, even as short as a single sample. To a normalizer, duration of the peak means nothing; it was the peak, and that's good enough for it.

This means that if your song has one relatively loud hit, at that 75%, and the rest of the song is, say, 35% of the maximum value, when you normalize to 95%, only the 75% is at 95%, the 35% is now only at 55%.

If you had a second song that also had a peak of 75%, but the rest of the song was 70%, when you normalize, the song will sound much lounder, because the average volume is considerably higher.

If the normalizer in your program has an option for RMS normalization, use that. It will take into consideration the average volume, and normalize accordingly. It will use something called limiting to basically chop off the momentary peaks, so the average volume can be increased.

If this was more confusing than helpful, sorry. Ask any questions you may have, and I'll try to answer next time I'm around.

MPEDrummer

08-08-2003, 02:30 PM	#3 (permalink)
mpedrummer2 Insane	OK...I'm a recording engineer, so perhaps I can shed some light on the subject... All normalization does is adjust the volume so that the current highest point in the song is equal to the level you set it at. For example, if the max in the song is 75%, and you set the normalization level to 95%, it'll turn everything up until the max level is 95%. Now here's the catch...that maximum level in the song might only happen for a very short period of time, even as short as a single sample. To a normalizer, duration of the peak means nothing; it was the peak, and that's good enough for it. This means that if your song has one relatively loud hit, at that 75%, and the rest of the song is, say, 35% of the maximum value, when you normalize to 95%, only the 75% is at 95%, the 35% is now only at 55%. If you had a second song that also had a peak of 75%, but the rest of the song was 70%, when you normalize, the song will sound much lounder, because the average volume is considerably higher. If the normalizer in your program has an option for RMS normalization, use that. It will take into consideration the average volume, and normalize accordingly. It will use something called limiting to basically chop off the momentary peaks, so the average volume can be increased. If this was more confusing than helpful, sorry. Ask any questions you may have, and I'll try to answer next time I'm around. MPEDrummer __________________ My sig can beat up your honor student.