WHAT IS NORMALISE?

Normalise is a function used to increase the volume of music.   The point of maximum sound output of a recording, will be increased to the specified volume level, without distorting.

For example, we have just recorded Mozart's Symphony 40.   As soon as we have recorded the file to computer, the loudest point of the piece (peak level), registers a volume of -25 db.  We want to make this piece of music ready for CD, so we'll want to increase the volume to it's maximum loudness without distorting it.  When applying normalisation, Sound Forge will scan the entire piece of music.  It will then find that the loudest point (peak level) registers -25db.  It will then increase this peak volume to 0db and adjust the rest of the music accordingly. 

On some setups, the waveform can appear very small.  This is because either the amplitude through the amp is very low, or the same may be the case from the record you're recording from. 

Throughout the 1970's and 1980's, record companies like K-Tel and EMI used a technology called, micro-groove pressing. This was the process of pressing records with a much tighter groove spiral, to fit on more  audio material.  Unfortunately this technically leads to much less sound output, consequently drowning the music further into the surface sound threshold.

To rectify this, you can either turn the input level higher on the amplifier, or normalise. 

However before normalizing, they are a few things to note.  If you are going to record an entire album, it is strongly recommended to record the whole album as one session.  If you normalise individual tracks, sound forge will scan to the loudest noise, which in many cases with vinyl maybe a pop, or a click.  As these vary in loudness,  normalising individual tracks may result in inconsistent volume levels through an album.

It is recommended to normalise to -3db, during the clean up and processing stages of restoring vinyl.  If normalisation is taken to 0db (maximum loudness),  some declicking and denoising processes may distort the signal.

Will continue through at the moment, assuming we are making a compilation.

To start the normalisation process, first click on the "process" on the top of the screen.  Then on the drop down menu, select normalise.  You should have a dialog box that resembles figure 8.  You'll find a slider to the left (1), that can be dragged either higher or lower, depending on how loud you want the file to sound.  On top of the slider is the normalisation value.  This is where the loudness level will be specified (2).  This setting will most probably be set at 0.00 db (100.00%).  You can use your mouse the change the values, but this can be fiddly, so I just set it in the rough proximity of -3.00db.  Then by touching the up and down arrows on my keyboards keypad, I would make the finer adjustments.    The setting should now be at  -3.00 db (70.79%). 

The last thing to always check is the peak level setting (3).  Sound Forge can either normalise to a peak level of a sound file (the loudest level output), or the Average RMS Power (average loudness of the sound file).  To keep things simple, I always normalise using the "peak level" setting.  The other setting leads in to complex territory, which will just confuse things further, so we'll just skip this function, as it's not relevant for this project.