I know what you mean, wingsman, but it's unavoidable: music is getting louder, and that includes with remasters.
The following waveform is that of "Something" from Abbey Road, throughout it's various re-releases.
The first is taken from the 1983 Japanese Release of Abbey Road. The second is taken from the 1987 worldwide CD remasters. The third is taken from the 1993 reissue of 1967-1970. The fourth is taken from the 2000 compilation album, 1.
But I'm sure you all see the ongoing pattern. The next logical progression is 'louder'! It's gonna be a nightmare.
Fantastic posts thefadedline!
For those that are just learning about this - there are really two issues in play here that change/degrade the music between the time it leaves the musician's fingers/vocal chords and the time it gets to our ears: a)
compression, and b)
clipping. Remember, "loud" and "soft" are just subjective terms, and there is no right or wrong with either. If a listener has a problem - reach for the volume knob on your radio/CD player! The problem comes where, in the recording/mastering process, the music is intentionally changed from what the musicians played/sang and wanted us to hear, or even partially destroyed - all so that it can sound louder than the competitor's CD (evil icon here).
Compression is not an inherently evil process, it's been used for decades, it was even one of the Beatles' favorite production techniques (they liked the Fairchild 670). One of its original purposes was to make sure that really soft passages didn't get overpowered by louder ones that came before or after it. One example would be a vocalist who can't help but sing certain notes loud, and others soft (as opposed to doing that intentionally for musical/artistic effect). So, in that example, compression would be used to make the overall vocal track more even in terms of volume. Another common example is the bass guitar where some notes are played unintentionally louder than others. What compression does here is "compress" or squash the loud passages down so they are closer in volume to the soft passages, and the overall volume is the same throughout the track.
But wait ... wouldn't this make the song softer, I thought we were talking about problems with the songs being too loud?! Right, there is a second part of compression that takes place after the "squashing" of the loud passages -that is to then make the whole newly compressed song louder (this part is called "
make up gain ", and no, it doesn't refer to the benefits of saying you're sorry to your boyfriend/girlfriend!). So now, you might have a song that at it's loudest is the same as before being compressed, but - the soft bits, which have also been made louder, are now closer in volume to the loud bits. Voila, a vocal track that is relatively even in terms of volume.
As it turns out, the whole vocal track sounds louder than before, so this process is heavily used in music production/mastering nowadays. It's important to realize that this comes at the expense of there not being as much variation in volume throughout the vocal track. That can be useful as in this example, but it would fight against you if the volume variations were intentionally being done for musically expressive purposes.
So where's the problem? If it's too loud, can't we just turn the volume down a bit?
Well, when compression is used/overused, there are two things that happen to the music that can't be fixed by twiddling the volume knob. The first one is that all the music on the song is at the same volume. Imagine if in a song like "I'm So Tired" the first bit ("I'm so tired ...") was just as loud as the chorus ("You'd say, ...") . That whole contrast between soft and loud that helps make the song work so well is lost, and now it's just not as interesting/good. Lots of other songs are great in part because of variations in volume throughout the song (another example is "Happiness is a Warm Gun", contrast the beginning ("She's not a girl ...") with the middle and later parts of the song. The whole thing builds up, almost certainly intentionally, to AWESOME effect). Then, imagine if this constant volume process were applied to a whole album/CD ... the brain just interprets that as less interesting than where there are variations in dynamics/loudness (imagine if "Yesterday" sounded just as loud as "I'm Down"!). Incidentally, this is also one of the things that separate a garage band from a better one - thrashing all the songs at the same volume generally doesn't sound as good as soft/loud variations.
The second way compression can ruin a song (in other words, have us hear something different than the musician wanted to play/sing) is that, in an effort to make the records louder, the process
intentionally destroys some of the digital musical information! Just so the record producer can win the "loudness war" against his/her competitor! Remember that make up gain we talked about, the 2nd part of the compression process? Well, if the make up gain is too high, then you can imagine that the peaks of the music get pushed up to the top of those little screens in the incredibly cool posts above by thefadedline , and then when you turn the make up gain
higher still , that's where
clipping occurs. What that simply means is that the peaks get
clipped off - and that bit of music is lost forever! If you do this too much, it sounds like distortion. Why would someone intentionally do that? It's so that the record sounds louder. The producers/mastering engineers of music today have figured out to a fine art how much they can push this before distortion becomes apparent, and they go right up to that point. Sure, there's no distortion, but we are not hearing everything that the musicians are playing/singing!
So, for real life examples,
let's look at some of those examples that thefadedline posted (fantastic examples, dude!). For "Something", clearly each successive mastering is getting louder. If that were all that was happening, not a problem - if it's too loud, turn the volume down. Or, if it's still not loud enough, turn it up! Either way, there is no information lost, no degradation of the signal (until you turn it so loud that your loudspeaker shreds!). The real question is,
is that all that is happening, or are we losing some of the musicianship due to compression? Well, it's hard to tell for sure without without hearing it, but by eye it looks like the differences between the softest and loudest portions are reasonably well preserved between 1983 and 1993. So, some compression may have been used (rather than just turning the gain up at the mastering process), but maybe it wasn't overdone. But look at 2003 ... can you tell which part was supposed to be softer and which part louder? Most people might say, "heck no!". That is a good example of what some people would say is overcompressed - no loud, no soft, just everything at the same volume. Is that how George wanted the song to sound?
Now, that is not as bad as the waveform for "See Your Sunshine". Two things to look for here. First, what is the variation in volume throughout the song? Take a look at the blue part (where the song is) - is there any variation in size of the blue (which translates into variation in volume while listening) from one portion of the song to another? To be kind - not much! It looks like there is about as little change in volume throughout the song as in the 2003 version of "Something". This makes the song less interesting musically. Yes, it allows it to sound louder on playback ($$ to the producer, they believe), but it is less interesting, and probably not the way the composer wanted it to sound. We can't be certain about that last bit, but most songs are written and played with some noticeable variation in volume, even if it's just playing a tad bit louder in the chorus than the verse.
But the other thing here is more significant. Look at the top and bottom edges of the track. You see where there is no white, only blue? At those portions (looks like the vast majority of the song!), the peaks are almost certainly clipped off,
and musical information is lost here. Why, Paul?
I did the same thing on my Cubase when Chaos and Creation came out, I loved the song "Friends to Go" so I wanted to check it out a bit. It started out soft, then got loud, awesome, that is one of the things that drew me to the song in the first place (nice acoustic stuff at the beginning, then killer bass/kick drum at the middle to the end). That is great, a classic use of "louder" - as a contrast to another part that is "softer", to keep the song interesting. BUT, towards the end, it got clipped, i.e., the make up gain was so high that the peaks were cut off ... information lost forever
). We'll never know (until/unless it is remastered) what Paul wanted to have us hear there. I couldn't believe that he would allow that. Maybe he's stopped paying attention to the "technical" stuff? Sad ...
All this is fantastically well described by one of the mastering gurus, Bob Katz on his webpage digido.com. Here are the direct links to two relevant articles:
http://www.digido.com/bob-katz/compression.htmlhttp://www.digido.com/bob-katz/level-practices-part-1.htmlHope this helps someone who might be wondering about this stuff, and thanks again
thefadedline for such awesome posts!
