My musical tastes are relatively populistic, but I’ve got the merest hint of music snob in my bones. It’s the part of me that recoils when Antares Auto-Tune is used as an overtly applied vocal effect, or when hearing pop performances so perfectly pitched and sterile that they’re robbed of life. Nonetheless, I appreciate the ability to make surgical strikes for correcting sporadic killer flaws, especially when they rescue a take that’s otherwise emotionally great.
Most of my work is done within Logic Pro, which includes an embedded tool called Time and Pitch Machine (I’ll call it TPM going forward). Used in conjunction with Logic’s sample edit window, it’s a snap to select an errant note and use TPM to shift it a few cents in either direction. The tool’s main purpose is to drive the time stretching and pitch adjustment of whole blocks of audio (e.g., Apple Loops). TPM offers nine choices of algorithms for handling these jobs. When I read that Serato had designed a proprietary algorithm claiming to outperform all of them, my curiosity was piqued.
Pitch ‘n Time LE is most at home in Pro Tools, but it’s equally effective in Logic. The only difference is that it’s prettier under Pro Tools as a standalone AudioSuite tool, where sliders and windows for tempo, pitch, BPM calculation, key shift, and output length are intuitively laid out. Logic’s TPM is fairly straightforward, though, and allows all of the same control. TPM’s “Classic” mode mimics the AudioSuite link that produces a varispeed effect, wherein time and pitch changes are joined. Pitch ‘n Time’s Pro version (which I didn’t evaluate) requires Pro Tools explicitly. It enables finer control with a Variable Pitch panel, Variable Tempo panel, Time Morph panel and a dedicated Varispeed mode.
Pitch ‘n Time LE can compress time down to 12% of its original length, pushing ballad tempos well beyond wind-up monkey speed. It can also stretch to 800% of original length, dragging even Yngwie Malmsteen to leviathan tempos that sound more like a glacially slow version of My Bloody Valentine’s Loveless LP. Logic’s legacy Version 5 algorithm, by comparison, can achieve only 60% compression or 200% expansion. The Universal algorithm can attain 25% compression and 400% stretch. I chose this set of algorithms to run my first tests, reducing a ten bar phrase to six bars. The Universal algorithm did the job quickest, producing exactly six bars of faster audio. Although Pitch ‘n Time LE took a second or so longer, it retained much more of the harmonic information from the original audio sample. The results were not brittle and choppy, as was the case with Universal. Version 5 actually kept more of the low frequency information being reproduced by my Focal subwoofer, but at a price. This algorithm wouldn’t produce an even six bars, settling nearly half a bar longer than requested. It also sounded the most distorted overall.
As an additional check, I tried stretching the sped-up clips back to their original tempo, using the same algorithm which had reduced each original clip. Again, Version 5 underperformed in its primary goal of producing a ten bar clip, falling a beat short. Its result also clearly underperformed the Universal algorithm, full of artifacts that sounded like the worst ring modulation effect money could buy. The squashed-and-then-restretched Pitch ‘n Time LE audio was predictably trashier than its original counterpart, but it retained more clarity at both ends of the audio spectrum than the Universal algorithm. The hi-hats in my clip held much of their crispness, the bottom end didn’t rumble, and the overall level wasn’t attenuated as it was in the Universal sample.
One easy task I’m frequently asked to do is to capture YouTube audio for a client and alter it into new performance keys or tempos for that week’s set by his church group (an action hopefully justified by their CCLI payments). Since the YouTube audio is hashy and badly compressed to begin with, any deterioration exaggerates an existing problem. I chose the same three algorithms to compare changes to a four minute rock tune by Lincoln Brewster, but also incorporated TPM’s Harmonic Correction option on the native Logic algorithms. Although Pitch ‘n Time LE doesn’t support this feature, I didn’t miss it. When slowing the song’s tempo by 40%, Pitch ‘n Time LE’s results were much smoother than either the Version 5 or Universal algorithms — with or without Harmonic Correction engaged. Version 5 ran the fastest at about 48 seconds, but again, didn’t match the destination length of the other algorithms. Pitch ‘n Time LE ran about 20 seconds faster than Universal for this longer section of audio, in addition to clearly sounding better.
The church band task frequently has me dropping keys from D major to A major, or five semitones. Pitch ‘n Time LE can handle 36 semitones either sharp or flat. I can tell you that my dog didn’t appreciate hearing Hillsong’s earnest female vocals after I pushed them three octaves higher. In any case, 500 cents proved to be no problem. I took the same Brewster song and pushed it in both directions. Similarly to the time shifting test, Pitch ‘n Time LE’s pitch shifting capability produced superior results to the benchmark Logic algorithms, though I was surprised to find that the otherwise lousy performance of that native Version 5 algorithm was not so far behind when it came to pitching material lower, without Harmonic Correction enabled.
One note — unlike Logic’s native Factory algorithms, the “Prelisten” function won’t process audio with Pitch ‘n Time LE selected. With no preview capability, audio must be rendered to hear Pitch ‘n Time LE in action. If further tweaking is required, you’ll need to choose Undo from the sample edit window and try again. For the improvement in sound quality, I found this to be an acceptable compromise.
Based on Pitch ‘n Time LE’s performance, I’d certainly be confident relying on its use for minor adjustments to otherwise solid vocal or instrumental takes within a mix. Experiments pushing a vocal syllable or violin note 5 to 10 cents proved to be musical and transparent. Pitch ‘n Time LE also proved adept as a compositional tool, as a way to try different keys when building a song arrangement in sections. I’ve used pitch shifting this way before, either altering a mixed section or changing individual elements, so that’s not new. The difference is that Pitch ‘n Time LE is a bit faster than most of the algorithms I’ve previously used, and more importantly, produces better sounding results. This isn’t a cheap plug-in, so I might not buy it if that was to be my only application. But given its performance (and depending upon your need), Pitch ‘n Time LE could be well worth the money.
(Pro $799 MSRP, LE: $399 MSRP; www.serato.com)
–Jeff Elbel, www.marathonrecords.com/happyclub
by Larry DeVivo
One of the more interesting signal processors to come along has to be the HEDD 192 from Crane Song. HEDD stands for Harmonically Enhanced Digital Device and is quite unlike any other processor on the...