If recording your own samples and pitching them down, then it makes sense to record at a high sample rate to avoid losing high‑frequency information — though you’ll need to make sure you use a mic, such as this Earthworks QTC50, that is capable of capturing frequencies well beyond the audible band.
I pitch down a lot of samples as a production technique, from drums to all sorts (don’t ask — I make weird music!) and I use the re‑pitch algorithm in Ableton Live. But, while this works, the process means a lot of high‑frequencies are usually lost. What’s the best technique, or combination of methods, for refreshing the high end to compensate for this? Ideally, this would be achievable with the Ableton Suite plug‑ins or equivalents, though I also have an iZotope subscription and their Trash 2. I have heard gentle saturation is good but want to learn even more!
SOS Forum post
SOS Technical Editor Hugh Robjohns replies: The top end is lost because there’s nothing there to keep: if your source file is sampled at 44.1kHz, it will have no information over 22kHz (at best), so if you then slow/pitch it down by a factor of four there’ll be nothing over 5.5kHz!
If your source file is sampled at 44.1kHz, it will have no information over 22kHz, at best, so if you then slow/pitch it down by a factor of four there’ll be nothing over 5.5kHz!
If you’re recording/creating the original sounds yourself it would be worth recording at 96kHz or even 192kHz, so that the file has an extended high‑frequency response. That way, when you pitch it down there will still be something going on in the higher frequencies. Having said that, to record acoustic sounds at such high sample rates you will also need to use mics with an extended high‑frequency response, since most standard mics capture nothing useful above about 20‑25 kHz. Some Earthworks mics extend to 40kHz, and Sanken offer one with a 100kHz bandwidth!
I sense that you’re talking about manipulating existing samples, though, and in that scenario (or if you can’t record at high sample rates) then you’re looking at ways of synthesizing high‑frequency content based on what you’ve already got. Lots of non‑linear processes (anything that results in harmonic distortion) can do that to some extent, but the most obvious is an Aural Exciter-type process. Alternatively, you could split the source track (either using an aux send or just duplicating it) and run one path through a saturation, distortion or clipping process — or all three! Then you high‑pass filter, and possibly EQ, the result and blend that track back in with the original to taste.
Since you refer to an iZotope subscription, it’s probably worth mentioning that iZotope RX has some spectral rebuilding tools that might help. Some other companies also make plug‑ins that aim to rebuild high‑frequency content lost through filtering — Zynaptiq’s Unveil, for example.