Alright, it’s been a long time coming, but I’ve finally found some time to get down to finishing touches.
First, I have a confession to make. I’ve been doing PC-based audio work for so many years that my workflow preferences are somewhat coloured by history. When I first started doing this stuff, computer horsepower was a serious impediment. You could only work with one bit of processing at a time, and never actually preview anything you were doing. Over time, this changed through successive stages, from realtime preview of an effect all the way to where we are now with the ability to stack tons of effects all over the place and hear fully realized renditions of them, even while tweaking parameters in realtime. I still haven’t fully embraced all that power, but there are exceptions…
Which brings me to my point, which is that remastering tool suites are awesome sometimes. There are a ton of plug-ins for different things, and at least as many opinions. Almost every discussion I’ve seen online includes mention of Izotope Ozone. If you’re serious about mastering, or remastering, it might be worth dropping some cash on it, or something like it. It can be used in any set of tools that support VST plugins, and I use it ALL the time.
Having said that, you can achieve similar (and maybe even better) results with individual effects, but I used Ozone for this part, so I’ll be showing it off a bit.
A further aside about things like Ozone. They come with stock presets for various things, like this:
The thing is, these presets are designed, at least partially, to be dramatic and show off how awesome the tool is, so when you see it demo’ed at your local music shop (upstairs at Tom Lee around these parts), you’ll be impressed and walk out with a boxed copy of the tool right on the spot. But if you actually use them as-is, you’ll end up with something that totally goes against my remastering philosophy of “subtle changes that improve things in a way that most people can’t quite put their finger on”. That’s not to say that they can’t make a great starting point for something good, which is in fact what I did with this project. You just have to dial things back (a lot).
Waveform pictures won’t make a difference at this point, since we’re messing mostly with EQ now, and not volume, so I’ll spare you a view of where we were at before now.
I actually started with a preset here called “Sparkle”. If you look at the picture above, you can see that there are six main sections to Ozone, and Sparkle is really just a combination of Paragraphic EQ (a hybrid between parametric and graphic EQs, but either would work for the simple tweak we’re doing here), and Multiband Harmonic Exciter. The other effects are all inactive.
So what exactly is a harmonic exciter? In a nutshell, it is a bit of black magic fakery. You take frequencies in your original source, artificially create integer multiples of those frequencies, and add them into the output. If, for example, you’ve got a pure note at 3500 Hz in your original source (roughly three octaves above concert A, for the musically inclined). You create some basic sine waves at 7000 Hz, 10500 Hz, 14000 Hz, 17,500 Hz. and add them to the signal. Sure it’s fake, but it gives you some higher frequency content that wasn’t there before. If you were only using EQ, sure you’d be boosting high frequencies, but if the high frequency content just isn’t there, it won’t accomplish much. Because the frequencies you’re adding with the exciter are harmonic multiples of something that does actually exist as a base frequency, you can be somewhat assured that the results will sound “musical” and won’t clash with the original signal.
Aside: Harmonic exciters are truly amazing. Even if you don’t cough up the cash for something like Ozone, you should get one if your audio editor of choice doesn’t have one. Freeware VST plugins are a google search away. They give you the ability to add that “sparkle” to just about anything. It’s especially important when remastering stuff from older sources where simply using an EQ to boost existing high frequencies results in the nasty side effect of also boosting tape hiss and other artifacts from analog sources. IMHO, no remastering toolkit is complete without some kind of harmonic exciter.
Alright, so using this stuff. I basically just pull the wave file I’ve got into Adobe Audition, go into multi-track mixing view (hit F12), drag the wave into the first track, click the fX button at the top of the track control column, and then select Ozone from the effects. Alright, that’s confusing…here’s what it looks like:
Now it’s just a matter of picking “Sparkle” from the list of presets, and dialing in the settings you want. As always, this comes down to listening carefully, and going for something that sounds good, and not over the top. Here are some screenshots that show you where I ended up:
You can also click on “Graph” on either of these views to show the order in which effects are chained. In this case it’s EQ first, and then the Exciter. As you can see, I did what I could with EQ settings to gently boost the high frequencies that were there first, and then applied the Exciter. Only band 2 does anything for the exciter, which uses frequencies in the range of 5.38 kHz to 20 kHz for the multiplier I was talking about before. I can’t remember what the amount was in the original preset, but I dialed it back to 1.6, and that seemed reasonable to my ears.
After you’ve got a setting that works, you need to mix down the results to a file. In Adobe Audition, this is under Edit->Mixdown to New File->Master output in session (stereo). The resulting file will look a little like the one below (on which I’ve highlighted a couple of areas for the next section):
This sounds great, but there are two problems, and they’re both in the red boxes above. Unsurprisingly, they’re both a result of the last step adding too much high frequency content to the the sibilant “S” sounds at the start of both instances of “South Dakota” in the vocals. In addition to being too loud compared to their surroundings, these two minor peaks keep us from normalizing the entire track to a more suitable volume. Rather than attempt some kind of compression, we’re going to spot-fix these.
This is where it’s again important that you really learn the navigation controls for your wave editor of choice. I’ll show one of the two fixes here, but it’s pretty much the same for all minor spot fixes to errant peaks. In this case, I’ll use the second, because it’s much more obvious. First you zoom right in on the part you want to fix, and select it. It will be something like this:
As you can see, this is barely two tenths of a second that is really at issue. As long as you only drop this by as little as absolutely necessary, it won’t even register much to the listener, and make a big difference overall.
For this kind of thing, you’ll want to use envelope processing. In Adobe Audition, this is in the amplitude section under effects. I played around with an inverted bell curve to get what I wanted. Make sure you have spline curves checked so you get a smooth transition. Here’s what mine looks like:
Here’s what that same peak looks like after processing:
As you can see, the peaks have been brought into line with the peaks in the adjoining audio. While auditioning this including a few seconds on either side, it’s barely noticeable. You might need a fair bit of trial and error to get the ramp amount exactly right, but it’s pretty quick to undo and redo this a few times to get it perfect.
So that’s really about it. After you’ve adjusted the peaks, you can normalize the whole thing as we’ve done in previous steps, and get a final waveform. In my case, here’s the final result after all the final fixes, and normalization:
Here again is the final result:
I hope this has served as a somewhat valuable tutorial. A lot of the ideas in here are mix-and-match, and should hopefully be useful for attacking specific challenges, even if what you’re doing is completely different.
That’s it for this project, but feel free to leave comments for things I’ve missed, been wrong about, or that need more clarification.
I’ll be back soon with more. It’s rare that I go for more than a few weeks at a time without some kind of audio project in the works, and I’ll be sure to write about whatever I’m working on next. At the moment, I’m messing around with Unity, making 3D environments to play around with using my Oculus Rift development kit…but I’ll get back to audio soon, like I always do!