Music Remastering and Remixing
It’s not news that the advent of digital technology has drastically changed the way in which music is both consumed and produced. I know from my own personal experience that the shuffle feature on my first iPod changed the way I listened to music, and it actually rekindled my interest in music to a large degree. I was somewhat of a purist from the standpoint that I would listen to an album or disc from start to finish and I was never a fan of ‘Greatest Hits’ or ‘Best Of’ collections. The simple fact was that I eventually started getting tired of carrying around a bunch of discs to fumble through and I was becoming less inspired to choose a complete disc of material to listen to in the first place.
When I finally got my first iPod and discovered I could shuffle individual playlists I had created it was like a breath of fresh air. I began to notice however that there was sometimes a large discrepancy in volume between tracks from older albums and more recent productions. This wasn’t really something I was totally unaware of before the iPod but it wasn’t really an issue when listening to an entire disc. Material on a single disc was produced fairly consistently in terms of volume from track to track.
To tackle the issue of why this was I go back to my own experience with recording music and audio engineering. When it came to the mastering stage one of my first objectives was to make sure that a track was normalized. That is to make sure that the highest peaks were at 100% of the total available digital value. Performing this action however did not often yield on overall volume that was on par with most of the music production in the early 2000’s. It basically all comes down to dynamic range but there are different aspects you have to look at.
When music is performed there are natural variations in the volume of each instrument over the course of a track. When everything is mixed there will be one point where the volume is at its maximum and if we normalize that point to 100% of the digital amplitude everything else will be relative to that, so if that peak is quite loud compared to the rest of the track then the overall track volume will seem quite soft. To counteract this effect the first thing to do is apply peak limiting or compression to the track. With the peaks at a lower level relative to the quieter sections of the track, when we perform the normalization the quiet sections will now seem louder. In the pre-digital era this dynamic compression process would be performed using an analog unit of some type.
Analog dynamic processing units would respond to the input signal in real time. The device would start the compression process according to an attack time the was usually controllable on the front of the unit. You could not set this attack time to 0 however because an analog unit needed to ‘hear’ the signal for at least some amount of time before responding. So there would always be a certain amount of peak however small that you could never get rid of. This not only limited the amount you could bring up the overall volume but the attack period could also be audible.
In the digital age a compressor/limiter algorithm can ‘see’ the whole track at once and perform the processing with no attack time. The perceived overall volume of track can be substantially increased without any audible undesired aftereffects. Having your music simply sound louder than another artists can be a big advantage.
Another factor that sets modern production apart is the increase in low and high frequencies. When albums had to be mastered for vinyl and tape production there was a limit to the amount of high and low end you could have in your tracks without causing distortion in those ranges. When you hear tracks recorded before around the mid 80’s you will probably notice this difference.
Over the last couple of years I have gone through some of my favorite albums and created my own personal remasters and in some cases remixes. Now when they come up in a shuffled iPod playlist they are more similar in overall volume and frequency spectrum. There are a lot of early albums that have been digitally remastered but the changes in dynamic range and frequency spectrum vary and are often quite subtle.
In the sixties stereo recordings were in their infancy. The instruments were often panned hard left, hard right or dead center. When listening to some of these recordings through a set of headphones these hard panned instruments can sometimes sound unnatural and discomforting, especially with bass instruments. Hard panned bass instruments can also sound less full in a set of speakers since bass frequencies require lots of moving air. Cutting off one of your woofers just doesn’t make any sense.
I recently remastered Herb Alpert’s ‘The Lonely Bull’ album originally recorded in 1962. The tracks on the album contained multiple hard panned instruments including hard panned bass and drums. With hard panned drums the kick drum would be included so once again you have hard panned bass frequencies. What I did for this album was isolate the bass frequencies and center pan them. With the main stereo pair I panned them towards the center somewhat and added some reverb to fill out stereo field. I applied some equalization and then the final dynamics processing. In cases where this much alteration is performed it may me considered more of a ‘remix’ than just a remaster.