Using R-Mix in Sonar opens up a world of cool EQ tricks.
Sonar's R-Mix VST plug-in is a DAW-optimised version of Roland's R-Mix software, which isolates audio based on frequency range and stereo positioning (for example, a hi-hat panned left, or bass panned towards the centre). It recalls the 'spectral view' in Adobe Audition, but is designed more for musical applications than surgical editing.
There's a misconception that the Sonar version is limited compared to the full version, but the only elements that were removed have no relevance in the context of a DAW. For example, there's no need for built-in WAV recording (after all, if you haven't figured out that Sonar can record audio, you probably shouldn't be reading this column!). Nor is there time-stretching and pitch-shifting, as Sonar already knows how to do that. However, note that, unlike many Cakewalk VST plug-ins, other programs cannot load R-Mix; it's licensed for use only with Sonar X2 and above.
You can't necessarily edit sounds with pinpoint precision without hearing artifacts, but much depends on the program material and how experienced you are in using R-Mix. Sometimes R-Mix works perfectly, sometimes not. The bottom line is that if you expect it to be a miracle worker, you'll probably be disappointed — until the first time it works a miracle for you.
R-Mix is intended for processing stereo files. You can still isolate frequency ranges with mono, but can't isolate within a stereo field, which can be important. (This is true even if you convert mono to 'stereo', but the left and right channels are identical.)
The heart of the program is the 'harmonic placement' window, a display that deconstructs the audio into coloured 'clouds' showing frequency, panning and amplitude. The cloud's colour indicates amplitude (black for quieter and white for louder; colours indicate in-between gradations). Higher frequencies are higher up in the window, lower frequencies are lower down, and stereo placement is mapped from left to right.
To give an example, a round, white element in the window's bottom centre that hits briefly and rhythmically is probably the kick drum. A yellow, elongated element located higher up and towards the left is likely to be a hi-hat. A consistently loud, centred, mid-range collection of elements is almost certainly the vocal.
The way you use this information is by delineating a frame, indicated with a red outline, which you can adjust to isolate a specific frequency range and stereo position. The frame can be either rectangular/square or circular/oval, as chosen by a Shape switch. You can define the frame size by grabbing any of its vertical or horizontal lines (or a corner) and dragging.
Of course, this isolation can't be perfect when working with a mixed track. If you're isolating something like a vocal and another instrument covers the same frequency range in the same stereo position, you'll process part of the other instrument too. As a result, the accuracy with which you draw the frame, and how well you can isolate only the section to be processed, greatly influences R-Mix's overall effectiveness; it takes some practice to learn how best to draw a frame to obtain the desired results.
In practice, though, I was surprised at the potential of framing for accuracy. In one song, a vocal was panned to centre and another was slightly off centre in the same frequency range. I thought they'd be panned too closely together to be able to process one without the other, but drawing a very narrow frame allowed me to add reverb to only the desired vocal. It was almost like reaching inside the original multitrack to add the new vocal reverb. I was also able to remove the kick drum from a loop, making it easy to add a new kick to the loop when it was used in a DJ mix that wasn't compatible with the original kick pattern. (Note: As you have to hear these sounds to believe them, there are audio examples of both of these R-Mix edits at /sos/sep13/articles/sonar-notes-media.htm.)
There are separate level, pan and bypass controls for what's inside and outside the frame. Panning moves what's inside the frame within the stereo field, but note that if the frame moves too far off centre, you can no longer adjust the right or left frame boundaries — so adjust the frame size prior to panning.
Preset effects for signals inside the frame comprise two types of compression, three delays (short, medium, long), and three reverbs (room, hall, plate). If you want to use a different effect, clone the track with the R-Mix. On one track, leave the volume up for the Inside Level and turn it down for the Outside Level, then insert the effect you want to use after the R-Mix plug-in. On the other track, leave the volume up for the Outside Level and turn it down for the Inside Level, then adjust the two tracks' volume for the optimum balance.
Additional effects include four noise-reduction algorithms, for hiss, hum, wind noise and air conditioning, that are independent of the frame; they affect all the audio. These algorithms aren't on the same level as something like iZotope's RX2, but they're fun for sound design, and the hiss-reduction algorithm can be fairly effective with cassettes.
The R-Mix control set isn't all that complicated; this processor is all about applications, such as...
- Remixing existing music: You could isolate, for example, a percussion part and make it softer or louder, or compress it. By turning down the outside frame level temporarily and altering the frame size and position, you can isolate sections with considerable precision. I've also been able to reduce the level of an overbearing hi-hat, and bring up the level of a rhythm guitar part that was panned to the left. The more isolated the part is with regard to frequency range and panning, the more effective the 'remixing' process.
- General EQ: If a song sounds muddy due to build-ups in the 300-400 Hz range, not only can you see this, you can draw a wide frame covering the region and reduce the level slightly to minimise the mud. This is much like conventional EQ, except that the visual component can help zero in on what needs to be edited.
- Loop alteration: It's often possible to remove individual drum sounds from an electronic drum loop, especially if they're panned so that the drums are well separated. Being able to customise loops in this way mitigates a loop's inherent rigidity.
- Vocal removal: Most vocals cover the mid-range and are panned to centre. Drawing a narrow frame around the vocal can remove or reduce it, while not affecting instruments above and below the vocal range that are panned to centre (kick or bass, for example). While an obvious use for this is karaoke, you could also use it to create a harmony for an existing vocal, yet reduce the level of that vocal so that the two voices singing together don't overpower the song.
- Dub and DJ remixing: If you're a DJ, R-Mix is the ultimate implementation of a kill switch. For dub, one really cool application is to isolate just a portion of the mid-range and add echo to that, while leaving the bass, kick and hi-hat unaffected, so they can continue to drive the song.
- Cleaning up narration: Working on a narration track with a few nasty p-pops, I was pleasantly surprised that I could draw a frame around the p-pop range, turn down the inside frame level, and solve the problem.
- Sound design: Narrowing in on very specific ranges and adding processing can turn just about any sound into something completely different.
R-Mix is arguably more of a creative tool than a problem-solver, although it certainly is capable of solving problems. Just remember that it takes practice to get good at 'R-Mixing': the more you use it, the better you'll be able to exploit it.