“A genius tool or gimmicky toy?” ponders PAR’s technical editor — before finding Roland’s unique R-Mix to be a surprisingly powerful tool.
If you haven’t heard about Roland’s R-Mix software, that is understandable. Found on Roland’s site under Music Education>Practice and Recording Tools, it’s hardly something you’d stumble across. But you should know about it. I’ll make two initial observations after spending some time with it. 1) Having this much power at such a low entry cost is shocking. 2) If you’re an audio purist, read no further. If however, you make a living working with audio and are expected to do things that should not reasonably be done, then you may need a tool like this.
When I first heard about R-Mix, I was told it could isolate individual tracks within a stereo mix and allow the operator to solo instruments, add reverb to individual instruments, change panning within the stereo field, change pitch and tempo or even delete instruments from a recording. Is all that true? Well, based on my experience, the answer is an absolute “That depends.” Is it a magic bullet that will allow you to remix from within a stereo mix with no sonic compromise? Definitely not. Is it a surprisingly powerful tool that works wonders at an acceptable sonic cost? The answer is a conditional “Yes.” So how does it work and what exactly does it do?
Let me start by explaining why R-Mix was developed. It was primarily designed for students who desire to learn component parts within a mix, like learning a bass part. It can also be used to generate a practice track, Karaoke-style, without that part. To accommodate learning, it can also modify the pitch (up or down an octave) and tempo (-50 to +130 percent) and, utilizing Roland’s VariPhrase Technology, it does a very admirable glitch-free job of both of those, even at extreme settings.
To start, open a new Project, which begins by selecting a single audio file, revealing limitation number one: Supported input file formats are mono or stereo .wav or .aif, 16-bit only, 44.1 or 48 kHz only. Export format is 16-bit, 44.1 kHz .wav only. So any working music professional recording at 24-bit or sample rates higher than 48 kHz is out of luck.
Once the Project is opened, I was presented with an extremely easy to understand interface, clearly designed for simplicity. Roland’s V-Remastering technology analyzes the audio according to three parameters: frequency, panning and volume. Frequency is displayed from bottom to top, as expected, but no frequency legend is marked beside the screen. Panning is displayed from left to right while Volume is displayed by colored blobs that change spectrally from blue to red as the amplitude increases. Inside the large Harmonic Placement window, a continuously-variable Frame window (of either rectangular or circular shape) can be used to highlight a sound and then the area inside or outside the frame can be individually turned up or down by +10 dB or –infinity. This allows highlighting of a frequency range that can then be eliminated or soloed. Seriously. The coolest thing is that, given adequate separation of the elements within a mix, it is possible to remove a bass part or vocal part completely from a mix while leaving the other parts intact. But — and this is a big but— this software operates in the frequency domain, so eliminating a lead vocal that is in the center of a mix will also take with it the other instruments in that frequency range, such as snare and any guitars or keys that are panned in the center. Soloing a range will yield the same results, with the vocal having all those similar frequency components audible in it as well. And there is an audible artifact when extracting (adding or subtracting) audio with the resulting audio having an aliasing, swishing sound, depending on how much R-Mix is asked to do. When doing less severe adjustments, like turning up a vocal level or turning down a bass part, it was quite acceptable to my ear. Also, once a part is soloed, it can be panned anywhere within the stereo field. When happy with the results, the file can be exported to save those changes to a new file.
There are also very rudimentary effects with percentage controls (compression, delay, reverb) but these will likely just frustrate a serious engineer, since there are no parameter adjustments. There is also a noise canceling section, but it seems little more than Hi or LoPass filtering to my ear.
When pushed hard like soloing a vocal within a track, I would consider R-Mix completely unusable from a purist sonic standpoint, as the aliasing artifacts are very pronounced. It will do the job it is asked to do but the results aren’t pretty. On the other hand, it does many things incredibly well. On one song with very few instruments, I took a lead vocal that was completely swallowed in reverb and was able to solo the voice and raise it by 10 dB, effectively turning down the reverb to an acceptable level, with results that I would easily put on a record. That’s a cool trick that many in TV post would appreciate I’m sure. And taking out a bass part worked on all the tracks I auditioned, so a user could eliminate an electric bass and replace it with a Moog bass, for instance. In a dense, compressed mix, however, being able to separate out parts will be next to impossible without seriously compromising the audio. Interestingly, I was able to use the classic Led Zeppelin song “Stairway To Heaven” and generate a completely acceptable track mix with no acoustic guitar. It sounded fine and the guitar part was almost completely eliminated. I did the same thing with the harmonica on the intro to “When the Levee Breaks.” Both of these mixes featured instruments panned off center that were relatively dry, making them easy to isolate.
I was shocked, quite frankly, at how good the pitch and time adjustments sounded operating in real time. They were glitch free and I would consider them useful for auditioning, even if not for record production. I was able to slow some songs down by half and listen to them cleanly and they weren’t garbled at all. I expected that taking the tempo up by 30 percent might result in chipmunky garbage, but the audio was very listenable. I was impressed.
Hear audioclips of R-Mix as utilized above at:
This is a very powerful piece of software for $200. For a studio needing to generate Karaoke tracks from stereo masters where limited budget or missing masters preclude doing remixes, this would yield acceptable results, even though definitely not high fidelity. On the plus side, beside the cost, it is incredibly easy to use and is very powerful and could be a lifesaver when someone is asked to do the impossible. On the down side, it doesn’t run as a plug-in and it tops out at 48 kHz and only exports to 44.1 kHz. If I was making a wish list, it would include 1) a Master Bypass button for all processes, plus individual Bypass buttons for all six processes, 2) Ability to handle higher bit-depths and sample rates, 3) Ability to use it as a plug-in within a session and 4) ability to save selected portions of a file instead of processing the entire file. For the money, though, it’s a very simple yet powerful piece of software. And there’s a version promised for iPad. I’m happy to know about it.
Lynn Fuston is the technical editor for PAR, an accomplished audio engineer and owner of 3D Audio, Inc. http://www.3daudioinc.com
Contact: Roland | http://www.rolandus.com