The hidden bonus track at the end of Magic Man’s 2014 release Before the Waves is a reworking of “South Dakota” which originally appeared on Real Life Color, released in 2010.
The new version is beautiful, but difficult to listen to because of it’s extreme dynamic range, and slightly muddy EQing. Because I’ve been listening to this track endlessly for months now, I decided to do something to make this track sound as great as it deserves to sound, so I could listen to it myself without the mixing issues affecting my enjoyment of it.
If you want to hear it, albeit stripped of some of it’s glory via the anti-miracle of lossy mp3 compression, you can check it out at https://soundcloud.com/umdesch4/magic-man-louth-dakota-umdesch4-remaster
Edit: Here’s a FLAC, so you can hear it properly, and decide if I did a good job or not – https://dl.dropboxusercontent.com/u/4006268/South_Dakota/Magic Man – South Dakota (umdesch4 remaster).flac
Since this actually turned out to be a lot of work, I decided it might be worthwhile to write a tutorial about what I did.
Part 1 – Prepping the track
This is admittedly dull stuff, but even here, there are some things to talk about that I feel worth sharing.
First and most obvious, you have to choose an audio editor in which you’re comfortable with the basic operations of scrolling around a waveform, zooming in and out, and standard cut/copy/paste operations.
I generally use two tools to do the bulk of my work: Sony’s Sound Forge Audio Studio, and Adobe Audition. In general, I find Sound Forge easier to use for extremely basic stuff like this, without too much interface clutter. Anyway, whatever you’re comfortable with should be good enough for this kind of thing.
So, here’s what I started with. You can see that there’s the “regular” part of the track (track 12, It All Starts Here), followed by a big chunk of silence, ending with the track I want to work with:
Rough cropping of this is fairly straightforward, but this particular example presents a bit of a challenge (not unlike many other things I’ve dealt with in the past). Where are the actual start and ending points of the bonus track? It’s tricky because there’s a fade in, and the overall volume of the intro is so extremely low that it’s hard to tell exactly. You can argue that it is simply a matter of zooming in to where you think the beginning section of the track is, cranking your monitor volume, and listening for it. That may not be good enough. Especially when dealing with 24 bit samples, it may turn out that the noise floor of the whole output chain to your ears is higher than the actual signal. This wouldn’t matter much now, but if what you intend to do later involves some heavy boosting (normalization, dynamics processing, whatever), it may turn out that the final result magically rises about this noise floor, and now you get to hear exactly how you’ve missed the start of the track. Whoops!
In this case, it turned out that I could hear the actual start (more on that in a minute), but just in case, you can do a visual inspection too. The simple trick is to zoom in on the area where you think it is, and then zoom vertically (ie. amplitude-wise) as far as your editor will allow you to go. In Sound Forge it ends up looking like this:
In Sound Forge, at least, the +/- buttons on the far left side are what you spam to achieve this view. The +/- buttons on the far right are for zooming in and out in time. (Sorry if that’s obvious, but hey…)
So now you can clearly see that there’s definitely signal at around 5:38. Listening to this at a fairly high volume, it seems like the musical fade in starts much closer to 5:40. Indeed it does, but there’s something interesting going on during those first 2 seconds. The band decided to start the recording a little early, and if you seriously crank it, you can hear a 60-cycle hum coming from their guitar amp. It’s subtle, but it is (IMHO) a powerful subliminal cue that sets up the whole character of the kind of recording you’re about to hear.
What you want to do with this is crop your recording, keeping roughly a half-second before this point. The reason for this is that, especially with advanced dynamics processors, there is often an option for “look-ahead” processing, and you want to give it at least a few extra handfulls of sample-space to work with. Also, caution here is a good thing, as you can always crop more after you’ve done everything else.
For the end point, you can use the same techniques. The difference is that you want to leave even more room at the end. It may be that you feel the need to introduce some slight mastering reverb to the source, and you will need room for the reverb tail. Also (although not as important here), one component of aural enhancers/exciters involves time delay of specific frequency components, so if you’re going to get extreme with some of these exotic effects, you’ll need the room. If the original recording doesn’t have room for these tails, you may want to cursor to the end of the selection and use whatever “insert silence” utility your editor has to drop an extra second or two in there.
So yeah, that’s the easy part, but getting it wrong can lead to headaches later on, so take a few extra minutes doing it right. Measure twice, cut once, right? Once that’s done, save out the cropped results to a new file, and you’re good to start the real work.
Here’s a FLAC file with my results. It’s a handy reference to the original, so you can judge for yourself how well I did at the various stages in the next sections:
https://dl.dropboxusercontent.com/u/4006268/South_Dakota/1.Magic Man – South Dakota (unprocessed).flac
Part 2 coming soon! (here: https://umdesch4.wordpress.com/2014/09/17/remastering-tips-magic-man-part-2/ )