Maximize vs. Loudness

GoldWave general discussions and community help
Post Reply
crobannon
Posts: 5
Joined: Fri Aug 04, 2017 10:48 am

Maximize vs. Loudness

Post by crobannon »

1. The problem that I am trying to correct is that various tracks seem to be recorded at varying volumes. Since I am usually listening to my library with random/shuffle, I can't be running back and forth to the stereo to change the volume. I have one recently downloaded, purchased mp3 of modern classical/new age music that was so quiet my remote speakers shut off.
My understanding is that maximize/normalize looks for the loudest sound and adjusts the volume and that loudness looks at the average volume and adjusts. It occurs to me that if, say, I'm processing the 1812 Overture with real cannons, normalize would result in a very quiet play until the end, but that loudness would result in blowing my speakers at the end. So, if my goal is to have all my music play at the same volume, which is the better process: loudness or maximize?
2. I have been sharing tracks with an audiophile friend who just yesterday flipped out when he found out that I had been maximizing/normalizing volume on everything I sent him. "Normalizing causes distortion!" Does it? I must say that all my tracks sound fine to me and he did not complain about distortion until he found out I had been maximizing. So, do loudness and normalize cause distortion? Can the human ear hear it?
crobannon
Posts: 5
Joined: Fri Aug 04, 2017 10:48 am

Re: Maximize vs. Loudness

Post by crobannon »

I have been advised that I shouldn't be asking if volume processes causes distortion, but whether it changes the sound. One friend says that volume controls causes "compression" so that everything will sound like AM radio-- decreased dynamics, no highs or lows, and muddy sound, which is not technically "distortion". Please comment on this.
GoldWave Inc.
Site Admin
Posts: 4372
Joined: Wed Mar 10, 2004 6:43 pm
Location: St. John's, NL
Contact:

Re: Maximize vs. Loudness

Post by GoldWave Inc. »

Maximizing a file to 0dB does not cause distortion. In fact it can prevent distortion if the file you are working on has peak amplitudes above 0dB (which can happen when mixing and applying various effects). All samples are scaled uniformly, so there is no compression. For lower volume sounds, Maximize potentially can increase dynamic range when used with the Noise Reduction effect.

When using the Loudness effect, you'll need to be careful applying the same LUFS value to all files. It is possible that a high value may cause the volume to exceed the maximum for a particular file, which could cause clipping distortion. You'd need to select the appropriate "Clipping control" setting to avoid that or choose a lower value.
DougDbug
Posts: 2172
Joined: Wed Feb 16, 2005 3:33 pm
Location: Silicon Valley

Re: Maximize vs. Loudness

Post by DougDbug »

..."Loudness" is a new feature in GoldWave and at the moment I'm on an older 32-bit system running an older version of GoldWave, so I can't look at it right now.

Maximize simply looks for the peak and adjusts the level for 0dB Maximized peaks. The peak levels don't correlate well with perceived loudness. Most commercial music is maximized (AKA normalized), including many "quiet sounding" songs.

If you don't go over 0dB it will not distort! With integer audio data, 0dB represents the maximum value that you can "count to" with a given number of bits. Nothing bad happens when you get close to 0dB, but it will clip if you try to go over.

GoldWave itself works in floating-point so it can go over 0dB and it essentially has no upper limit. However, if you save-as a regular WAV file or make a CD, etc, it will be clipped at 0dB. You analog-to-digital converter (recording) and digital-to-analog converter (playback) are also hard-limited to 0dB. So for example, you can boost the bass in GoldWave and the peaks might go over 0dB. Then you can run Maximize to bring the peaks back down to 0dB before saving.

Perceived loudness is more related to the (short-term) average level and the frequency content. For example, an acoustic guitar that hits 0dB on the peaks will sound quieter than a distorted electric guitar hitting 0dB on the peaks. The electric guitar has a higher average level.

The EBU R128 Loudness standard, ReplayGain, and MP3Gain, all try to find/match the perceived loudness. Since you have MP3s, you might look-into MP3Gain.

So, here's the dilemma... Since many quiet-sounding songs are already maximized/normalized so they can't be boosted to match the louder songs. The only way to match volumes is by reducing the louder songs. (And then a lot of people complain that all of their songs are too quiet.)

Highly dynamic classical music is a "difficult" problem. I don't think two people would agree when two (or more) dynamic recordings are "equally loud".

You might consider dynamic compression for "casual listening". If you're shuffling music you probably don't want the full dynamics of The 1812 Overture! You can use as little or as much compression as you want. If you boost the quiet parts by +3dB (or reduce the peaks by -3db) you probably wouldn't even notice. And of course, you can keep the original for more serious listening.
I have one recently downloaded, purchased mp3 of modern classical/new age music that was so quiet my remote speakers shut off.
Maybe you need some different speakers...
I'm processing the 1812 Overture with real cannons, normalize would result in a very quiet play until the end, but that loudness would result in blowing my speakers at the end.
In an ideal world you'd have a big amplifier and big speakers/woofers/subwoofers. You'd hear the normal loud-music parts "full orchestra volume", and you'd have headroom for the cannon shots without distortion. You might blow your ears out, but you wouldn't blow your speakers. :twisted:
2leftfeet
Posts: 149
Joined: Sun Jan 20, 2013 10:28 pm
Location: Napier New Zealand

Re: Maximize vs. Loudness

Post by 2leftfeet »

Ii have been using the "maximise" function for many years because this function alters the whole file in the same way. This preserves the dynamics of the music especially with classical music. As far as modern music today is concerned it has already been altered in some way by the recording studio and then by the broadcasters. I just maximise all the tracks to 90% of maximum (no reason really but my choice). This is the LAST thing I do after any other processing which might be needed. Now, when I play any or all the tracks, I know that I will not have to mess with the fader on the mixer!
I have found that, in amateur dramatic theatre setting if I use Videomeld to play, I can always use the "fade" function if I need to make adjustments to cater for artists singing or talking during the playing of an audio track. Again, I don't have to adjust the fader :) BTW, I put my cues in a text box on another Videomeld track.
Post Reply