RESAMPLING: BUT NOT AS WE KNOW IT
What do we expect from our samplers, or our hard disk recorders, when it comes to capturing digital audio? Broadly, we expect quality. Most of the time, we want our recording devices to capture sound in a way that is a pristine representation of the original performance. Again, this wasn’t always possible.
THE DILEMMA
Scratchy, low-fidelity audio; for every semblance of signal, there was a disproportionately high amount of unwanted but unavoidable noise too. A montage sequence fastforwarding through time would lead us to the high sample frequency, high sample resolution formats of today, where hard disk space is not only plentiful, it’s available via solid state drives which respond at speeds unimaginable in the not-too-distant past.
Those interested in sampling know that this high quality is particularly important if you intend to start stretching your audio files or taking them to pieces in other innovative ways. It stands to reason that the more information there is in an audio file, the better it will respond to processing.
If you’re not convinced, try copying a 24-bit, 48kHz file and converting the copy to a standard definition mp3. Process both files with a time-stretch to double the original file’s length and you find that the ‘density’ of information in the high res file stands up to processing more robustly than the low-density mp3 equivalent.
A SOLUTION
But what if you want to willfully reimagine your sampling in a way that offers a pleasing balance between ‘interrupting’ the quality of the audio as it plays back as well as bringing a sound all of its own? If those ideas appeal, granular resynthesis may well be interesting.
Imagine a ‘conventional’ digital audio recording as being a little like a high frame rate digital camera. It opens and closes many times per second to capture an individual frame and, when it does, its ‘resolution’ determines how ‘high quality’ the depth of visual field is. Conventional audio sampling is similar – the sample rate (44.1kHz or 44,100 times per second, for example) determines how frequently a section of audio is captured and the resolution (such as 24-bit) determines how ‘deep’ and
‘detailed’ each piece of audio capture is.
But in granular synthesis, what happens at the resampling stage is that each ‘piece’ of audio is captured as a ‘grain’, which you can imagine as a collection of individual pieces of a sample. You have control over the grain size, where the smaller the grain, the more focused and like the original sound the sample will be. Equally, the larger the grain, the more ‘pixelated’ and processed the resulting sound will be.
A number of plugins and processors allow you to experiment with granular synthesis. If you’re a Logic user, you’ll find the Alchemy sampler offers comprehensive granular resampling capabilities. Spectrasonics’ Omnisphere 2 offers granular re-synthesis as one of its many modes of synthesis/sample processing, while there is even the chance to try granular timestretching within NI’s Battery 4.
OUTPUT PORTAL
If you want to experiment with what granular processing can do when inserted across any instrument or audio file, perhaps the best place to start is with Output’s Portal plugin. This allows you to reimagine in new and interesting ways, whether you choose to process an individual sound, or run your entire track mix through its parameters.
Portal is capable of all kinds of different musical effects, adding clouds of pitched overtones to even the most percussive material, or weirdly warping and modulating the pitch of a source signal with break-point envelopes. You can scatter drum loops, toggling their playback speed from a fraction of the original to multiples of many times, whilst also time-stretching and re-ordering vocal parts. And lots more besides.
WHAT IF YOU WANT TO WILLFULLY REIMAGINE YOUR SAMPLING?