Home -> Magazines -> Issues -> Articles in this issue -> View
Sampling Techniques (Part 1) | |
How To Get The Most From Your SamplerArticle from Sound On Sound, December 1992 |
If you only knew it, your sampler is a veritable digital playground, just waiting for you to come and frolic in it — throw caution to the winds and try some off-the-wall creative techniques. In the first part of this new series, Tom's doing a bit of bouncing...
OK, so you've got a sampler, and you use it all the time. But are you really getting the most out of it? The idea of this new column is simple: to explain a full range of sampling techniques and tricks, from the very basic to the advanced, to help you exploit your sampler to the max. For the record, I use an S1000, but the techniques I'll be describing can be applied across the full range of contemporary samplers; however, terminology varies from one manufacturer to another, so just bear in mind that this is the case, and you shouldn't be caught out. Anyway, on with the show...
One useful technique I've been using for several years is what I refer to as sample bouncing. Similar in many respects to bouncing tracks on a multitrack tape recorder, sample bouncing involves layering several sounds (say brass, strings and choir) within one sampler and sampling — bouncing — these as a composite into another sampler. Track bouncing (also called ping-ponging) is used in recording studios around the world every day of the week, so why not sample bouncing? What about the loss of fidelity? Well, what about it?
Whether you realise it or not, modern sampler technology has put within the reach of the masses highly advanced digital recorders capable of capturing any perceivable sound. It might not record more than several seconds at a time, but, yes, your sampler is a digital recorder.
In terms of sound quality, 16-bit samplers come mighty close to state of the art digital tape recorders found in £1,500 a day studios. When sampling at rates of 44.1 kHz and above, for all intents and purposes 16-bit samplers can be considered master-quality recorders. Even 12-bit samplers like the S900 at their highest sampling rate can sonically out-perform most domestic cassette recorders.
So, unless there's something seriously wrong with the input and/or output stage of your sampler, to the average listener there should be no perceptible loss of audio quality when bouncing samples using high sample rates, and certainly much less than when bouncing the same material using analogue tape.
Just because you might not have access to a second sampler doesn't mean that you can't join in the fun. If you're prepared to trade off a little fidelity, there's no reason why you can't layer sounds in your sampler (ie. so they're triggered from the same MIDI note), record this to tape, and then record the composite back into your sampler. If you have a multitrack tape recorder, you can even layer sounds to the maximum your sampler will permit, record this composite onto one track (or two, if you want stereo, and your sampler allows it) and repeat the process with different sounds until the desired mega-result is achieved. Sampling different mixes back into your sampler will enable you to take advantage of velocity switching, if your sampler has this facility. A sync track and sequencer will ensure accurate entries across all tracks.
You could digitally mix your samples, of course — mixing samples entirely in the digital domain adds little (if any) coloration or degradation to the sound — but sample bouncing has several distinct advantages over digital sample mixing:
• You can layer as many samples as your sampler permits. With the exception of Digidesign's Sound Tools, the now defunct PPG Waveterm B, and Dynacord's Add 1, which can mix up to four samples simultaneously, currently available sample mixing software only allows you to mix two sounds together at a time. With sample bouncing, you can layer as many sounds as your sampler permits — more if you are working with multitrack tape.
• Loops are played (and recorded) for the duration of the note(s) held. As far as I know, the Roland S770/750 is the only sampler capable of doing this in the digital domain.
• Calculating sample rate conversions to match the pitches of samples before mixing them is unnecessary, as pitch matching with sample bouncing is carried out on your program page when layering sounds.
• Sample bouncing is faster than digital sample mixing and allows you to audition the composite sound before actually combining samples.
• With the aid of an audio mixer, sample bouncing allows you to EQ each of your sounds to taste, given that each can be sent to a separate output of your sampler. You can also combine external sound sources with your layered samples as you record them. For example, mixing in MIDI and non-MIDI synth sounds, records, CDs, your voice, guitar, and so on.
I won't kid you. The further you move from keeping sounds entirely in the digital domain, the more you leave your sounds open to signal degradation. But don't ever let this stop you when the creative urge is present — the 'essence' of a sound is infinitely more important than its technical perfection. Besides, a little bit of grunge never hurt anyone. You'd be surprised at how many owners of top-end samplers such as Fairlight Series IIIs, Akai S1000s and Roland S770s actually seek out grungy old 8-bit samples to complement their palette of crisp 16-bit sounds. The Emulator II's crunchy 8-bit Mega Brass is still up there on the list of favourite sampled brass for the distinctive quality it adds to a track.
Of course, you'll want to make sure that all leads are in good condition and as short as possible. There's little sense in using a 20-foot guitar lead if one sampler is only perched on top of the other.
When using tape as an intermediate step to realising your mega sounds, make sure tape heads are sparkling clean, de-magnetised, and azimuth adjusted for best results. In a pinch — but only in a pinch — if no tape head cleaner or isopropyl alcohol is available to clean tape heads, try some aftershave or perfume on a cotton bud (it's mainly alcohol anyway), but only as a last resort.
There are three alternative sample bouncing procedures which you can follow. All can give good results, but in ascending order of preference, they are:
"Remember that just because you don't have access to sample editing software doesn't mean that you can't came up with some spiffing sample composites and hybrids."
On the orchestral front, you might want to make large brass, string and/or vocal ensembles from small section or solo samples. Or perhaps layer a variety of different samples, in their proper registers, to create your own personalised orchestral strikes. People who like to bang things might want to combine their favourite snares as one sample — especially if you've found yourself layering these sounds together before — it'll save RAM in your sampler. You may even want to experiment with mixing and matching the crack of one drum with the body of another and the ambience of yet a third drum — fun stuff.
BETTER: Take this option if you have a second but dissimilar sampler, with different resolution (one 12 and one 16-bit sampler, for example) and/or no MIDI sample dump. Bounce your sounds into the second sampler's audio input then sample the composite back into the first sampler's audio input (remember that samplers are digital recorders).
BEST: If you have a second but dissimilar sampler, operating at the same bit rate as the first, and both have MIDI sample dump, bounce sounds into the second sampler's audio input, then MIDI sample dump back to the first sampler.
IDEAL: The best situation is having two identical samplers, in which case you can bounce sounds into the second sampler's audio input, then save to disk.
The first thing to do is plan your attack and gather together the samples you want to mix.
GOOD: If you don't have access to a second sampler, bounce sounds onto recording tape and then sample the composite back into your sampler. DAT, F1, hi-fi video systems, reel-to-reel (highest available speed) and personal multitrack recorders running at double speed will give better audio results than standard audio cassette. Use noise reduction if you have it.
Synth lovers will enjoy mixing samples of their favourite synths together to arrive at new textures and timbres, or to further their search for the ultimate bass sound. With sample bouncing, it is possible to play the equivalent of a room full of synths (or samplers) from one key of your sampler if you desire. Whatever your bag, get out those disks, have a good listen, and choose which sounds you want to work with.
You'll need to assign all the samples you want to bounce together to play back from the same key or range on your sampler. If this is not possible, assign samples to adjacent keys and play them back simultaneously, either manually or with a sequencer. Note ons can be staggered to increase the ensemble effect.
Tune samples as appropriate — in unison, octaves or to the notes of a chord. When working with samples tuned in unison or octaves, a slight detuning (a few cents sharp or flat) between samples has the effect of fattening up ensembles and provides a more natural-sounding chorusing effect than using a dedicated chorus effect unit — especially when there are three or more samples beating off one another. Be sure to leave at least one or two samples tuned to the correct pitch. Gigantic, larger than life drums can be achieved simply by layering the same sample five or more times upon itself, each tuned 10-15 cents (1 cent = 1/100th of a semitone) flat or sharp of the preceding sample.
Don't forget about shaping your samples' loudness and brightness contours with your sampler's amplifier and filter envelopes. These can be used in a corrective fashion, such as using a low-pass filter to clean up tell-tale tape hiss, or an envelope-controlled amplifier to balance the looped sustain portion of one sample against the others. On the creative side of the coin, envelope-controlled amplifiers can be used to fade one sound through another, or to allow only the attack portion of a sample to be heard as part of your composite.
You can balance the samples' volumes against one another, either internally within your sampler or externally via an audio mixer if your sampler has separate outputs. If your sampler is stereo, panning may also be executed internally or externally.
Don't underestimate the power of equalisation to help transform library samples into something more uniquely your own. As most samplers only have low-pass filters, a graphic or parametric EQ can be of immense help in keeping things interesting. By giving each sample you are mixing its own timbral space, they can each hold their own a bit better when combined as a composite. Lightening up the lower frequencies will also help keep your bottom end from getting too murky — a common problem when layering tracks using only samplers and synths with low-pass filters.
Of course, since you're going into a sampler's audio input whichever way you execute your sample bouncing, there's no reason not to experiment, treating the signal with any effects you feel are appropriate. Be careful with reverb and effects, though — once sampled, there's no way to get rid of them. Add reverb sparingly, if you absolutely have to, at the re-sampling stage, otherwise leave it off. It can always be added later on, when you'll probably choose more appropriate reverb levels and settings.
Finally, remember that just because you don't have access to sample editing software doesn't mean that you can't come up with some spiffing sample composites and hybrids. Your sound sources and imagination are your primary assets and the end results always justify the means. Happy sampling!
LIBRARY SAMPLES | ORIGINAL PITCH |
Hi Strings (good bow bite) | G4 |
Synth strings (smooth loop) | C4 |
Solo Violin | A3 |
Solo Violin | D#4 |
Group Violin | G3 |
Assign all samples to G3 through G4 (MIDI note numbers 67-79 on the keyboard), and tune samples to play back at the correct pitch (see sidebar on tuning).
We want this composite to cover the range of an octave, and have chosen samples with original pitches spanning this octave. The individual samples might not travel an octave very well on their own, but en masse stand a better chance of getting away with it.
Playing back the samples at the lowest planned playback pitch (Group Violin's G3) will ensure that most, if not all, of the available top end is within sampling range, but requires more sampling time for all members of the composite to reach a steady, loopable state. Playing back the samples at the highest planned playback pitch (Hi String's G) will require allocating the smallest amount of sampling time, but the trade-off is a considerable loss of top end (with samples recorded and bounced at 44.1kHz, you'll be halving the available top end on the Group Violin's G3 by resampling it at G4). Try it nonetheless: the 'rounding out' of some samples can often be a pleasant alternative to hard, crisp realism. A happy medium is to play all samples at a midway point, say C#4 or C4 when resampling the composite.
SAMPLE | ORIGINAL PITCH | NEW PITCH | STEREO PANNING | |
Library Strike 1 | D3 | C3 | Full Left | root |
Library Strike 2 | G3 | C3 | Full Right | root |
Pizz Violins | G5 | C5 | Left | root |
Analogue Brass | F#4 | G4 | Right | fifth |
Cymbal Crash | C3 | E3 | Mid left | no pitch |
Staccato Celli | A3 | C2 | Mid Right | root |
Bass Drum | D#3 | D#3 | Centre | no pitch |
Each sample should have its own amplifier envelope. Short, staccato shapes are the way to go, with maybe a touch of sustain and release. An envelope setting like A:00 D:48 SS:20 R:45 (on the S1000) could be a starting point.
At the pitches we have chosen here, both strikes, the pizz violins and the celli all play our root C; the brass is tuned to a fifth. The presence of thirds (major or minor) can limit the usefulness of chord samples like orchestral strikes, so I've omitted them here. Of course, non-pitched samples such as bass drum and cymbal are tuned according to taste — in this example, a short, bright cymbal was called for, so I tuned the cymbal a major third higher than its original pitch, but left the bass drum alone as it sounded fine.
Having assigned all the samples to the same key (say, middle C, MIDI note 60), route the composite through your reverb on a large hall setting to complete the effect. With true stereo (or two mono) reverbs, reversing playback sides can often help smooth out any incongruities in the original mix's stereo balance.
Read the next part in this series:
Sampling Techniques (Part 2)
(SOS Jan 93)
All parts in this series:
Photographing Sound - The Art of Sampling (Part 1) |
Soft Options - Cheap And Cheerful Sample Creation On The Atari ST |
An Emulator for £10 |
The Art of Looping (Part 1) |
The Complete Sampler Buyers' Guide |
Sweetening the pill - Sampler test |
A Vocal Chord (Part 1) |
Retro-Sampling - Sampling Classic Electro Sounds |
Hands On: Akai S1000 Sampler |
Hands On: Roland S750 |
Criminal Record? - Sample CDs (Part 1) |
Total recall - Akai the old |
Browse by Topic:
Feature by Tom McLaughlin
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!
New issues that have been donated or scanned for us this month.
All donations and support are gratefully appreciated - thank you.
Do you have any of these magazine issues?
If so, and you can donate, lend or scan them to help complete our archive, please get in touch via the Contribute page - thanks!