Standalone spectral processor gets an adrenaline shot and new features
Fig. 1. SpectraLayers Pro 2’s new Extract Shape tool removes clicks (shown here as orange lines) embedded in a voiceover track. The clicks were copied to a new layer (depicted as a red horizontal bar on the right side of the GUI), and the layer’s phase was reversed to null the click’s amplitude when combined with the layer containing the original, unprocessed program (green horizontal bar on the right). FROM ITS initial release, Sony SpectraLayers Pro could extract various elements embedded in a stereo track and process each independently before recombining them to create an entirely new blend. You could also paint frequencies with your mouse on, for example, selective drum hits in a dense mix to reinforce their bottom end. Powerful noise-reduction tools jettisoned hiss and AC hum and buzz.
That said, the original release was hamstrung by an extremely cumbersome GUI and incomplete and sometimes baffling documentation. The new SpectraLayers Pro 2 (SLP2) promises a refined GUI, souped-up engine, and new tools
See Me, Hear Me Audio is displayed in SLP2 in an auto-scrolling spectrogram: The timeline occupies the horizontal axis, and progressively higher frequencies are illustrated farther up the vertical axis (see Figure 1). Louder elements appear brighter. Stereo material is split into top (left-channel) and bottom (right- channel) display panes.
Imported audio appears in SLP2 as a layer. You can copy bits of material from this layer— limiting your selection to a specific frequency range (or one fundamental frequency and its harmonic series) over a select period of time—and paste them to a new, empty layer. For example, you can extract from an audio file only the midrange frequencies in a dog’s bark, keeping or eliminating the extracted material in the original file as you wish. You can also add a frequency or noise to an audio file by painting it with your mouse.
Fig. 2. SpectraLayers Pro 2 allows you to create and name markers. These are for visual reference only and can't be used as navigation aids. Offline processing—using either SLP2’s built-in spectral processors or your third- party 32- and 64-bit VST plug-ins—can be auditioned before rendering. You can drag and drop edited and processed files into your DAW or mix them in SLP2 to create a new file. Files and sections of files can be directly imported from and exported to Sony Sound Forge Pro 11.
Exploring the New Features Spectral displays require a ton of horsepower, especially when zooming way in to see microscopic details. Unlike with the original software, SLP2 never got sluggish even when zooming in to view a slice of audio spanning only a tenth of a second and 300Hz bandwidth with maximum frequency resolution (an analysis-window size of 32,768 samples); I used an 8-core Mac Pro with SLP2. I could zoom the frequency axis while keeping the bottom of the display anchored to 0Hz and switch among five different frequency scales (including linear and logarithmic) to optimize my view of various spectra. After making a selection over time and frequency ranges in the spectral display, I could loop playback of my selection. Although I could set and name markers, these were only useful as visual cues and couldn’t be used as navigation shortcuts (see Figure 2).
The new Extract Noise command makes the process of removing broadband noise and AC hum throughout the total length of program material exponentially faster and more intuitive than in the original software. I also used the new Extract Shape tool—which selects objects in the spectrogram based on your specified power range—to virtually eliminate clicks embedded in a voiceover track (shown in Figure 1). This entailed copying the clicks to a new layer and reversing the layer’s phase with respect to the original, unprocessed program to null the click’s amplitude.
I used SLP2’s new Spectral Casting function to subtract the spectra of a music bed from a voiceover (VO) track (working with bed and VO contained in discrete layers). This essentially eliminated frequency masking and made the VO clearer when the two layers were played together. When soloed, the processed music bed exhibited mildly phasey artifacts and occasional dropouts (inaudible when the two layers were played at once, unless I made subsequent EQ or level changes to either layer). I liked the new Spectral Molding function for sound design; I used it to make a synth pad voice using only the spectra of a discrete VO, creating a vocoder effect for the pad.
As with the original release, SLP2 wouldn’t let me make time-range selections starting at 0:00:00:00; make sure you import audio with a handle (blank pre-roll). Because SLP2 cannot extrapolate content that borders extracted spectra to fill in a resulting gap, there is the potential to create “holes” in your program. SLP2 is also very buggy, and the learning curve remains steep (in part due to documentation that’s vague or incomplete in places). (The developer notes that the SpectraLayers Pro Seminar Series has been posted in the Training section at sonycreativesoftware.com.)
Compared to What? People working in sound design will want to compare SLP2 to iZotope Iris, a competing spectral processor that’s far easier to use and more feature- packed than SLP2 for that application. But if you also need powerful noise reduction tools for restoration, post-production audio sweetening or audio-forensics work, SpectraLayers Pro 2 is the only spectral processor I know of that gives you tools for executing all these tasks under one roof.
STRENGTHS: Dramatically improved engine. Streamlined noise extraction. New spectral processors. New loop-playback function. Can set and name markers. Allows offline processing with third-party VST plug-ins.
LIMITATIONS: Buggy. Vague and incomplete documentation. No interpolation function. Can’t snap playback to markers. Can’t make time-range selections starting at 0:00:00:00.
SpectraLayers Pro 2: $399 MSRP
(upgrade from SLP1: $199.95)