You are here

Creative Use Of Your Sampler

Tips & Techniques By Paul Ward
Published February 1995

Are you forever coming up with the same old 'N‑N‑N‑N‑Nineteen' effects on your sampler? Do you reach for that jaded Peruvian noseflute sample every time you want a 'really original sound' for your latest masterpiece? Fear not, for top tipmeister Paul Ward is back, with more off‑the‑wall sampler uses for you to try out at home...

There can be little doubt that the digital sampler has had a profound effect on the way that modern music is produced. But, like many other forms of musical technology, it wasn't until the sampler became affordable to the masses that we began to learn of its real capabilities. From a slightly shaky start of vocal stutters and massed choruses of blown milk bottles, the sampler has spawned a whole host of methods for the creation of musically usable results. Here is a selection of techniques from my own personal armoury.

1. Feeling solo...

When recording on multitrack, instrumental solos (guitar, sax, and so on) are often left towards the end of the track‑laying process. This has the advantage of giving the performer a reasonably accurate picture of the backing to his/her solo. Besides being generally more inspiring to play to, this arguably helps to produce a performance more in sympathy with the finished product. Unfortunately, by this time in the recording process, there are invariably very few tape tracks left with which to try out numerous versions of the solo. In the case where the solo is being laid on the only empty track left, the player becomes very reluctant to record over a 'nearly perfect' attempt, and the later portions of a performance are often 'played safe' to avoid messing up an otherwise flawless take. The result can often be a stilted, 'not‑quite‑there' solo that no‑one will ever be quite happy with.

A sampler can make all of this a thing of the past. By sampling each pass of the solo, and hiving it away to disk, you can make many attempts, trying for that elusive 'perfect' take before the best one is committed to tape. Indeed, the solo need never see a piece of tape at all.

2. Syncing samples simply

Recording into a sampler is all well and good, but how do you get this to play back in time with the backing track? Rather than using manual or audio‑triggered sampling, I set my sampler to begin recording on receipt of a MIDI note (this is done on the 'REC1' page on the Akai S1000). If the same note from the sequencer is used to initiate both sampling and playback, the result should be replayed in perfect sync. If your sampler doesn't offer MIDI sample triggering, then use the same MIDI note to play a trigger sound from, say, a drum machine (a cowbell is ideal), and feed this into the sampler as part of the recording source to initiate sampling. Since your trigger sound will become part of the recorded sample, you should ensure that there is a reasonable gap between the trigger and the actual solo — a second or so should suffice. If you use a long attack time, MIDI volume commands, or even edit the sample data to zero where the cowbell is recorded (the Kurzweil K2000 is amongst samplers that can perform this function), it should be possible to remove the trigger sound from playback. When playing back the sample, you may find it necessary to introduce a small negative delay to the playback note to counteract the inherent delay produced by using an audio trigger, but in most cases this will be negligible.

3.Old splice: cutting samples the easy way

So, you have 34 versions of a guitar solo conscientiously archived to disk, and the band are confident that they have the killer solo of all time — if you could use the first three bars of the seventh attempt, the next six bars from the thirteenth take and the dive‑bomb ending of the last! 'Easy‑peasy', you say, 'just chop up the samples and play each portion by triggering them from the sequencer'. Well, not necessarily. For one thing you'll spend a lot of time that way finding the ideal splicing points where clicks and glitches can be avoided (itself not an easy task). Then you'll spend an awful lot more time trying to get the length and start points of the samples correct, and moving the trigger notes to get each part in sync with the rest of the track. I have known hours to go by during such tasks. There is a simpler way...

Load up all three solos into your sampler, but set each to respond on a different MIDI channel. Make two extra copies of the sequencer track with the trigger note on it, so that the trigger is transmitted simultaneously on those three channels (in reality there will be a tiny delay between each note, but that's life!). Now, create three more tracks for the same channels, with a blank part/pattern in each (ideally, these patterns should start a bar or two before the solo actually begins). Use whatever editor is available in your particular sequencer to insert a MIDI volume command at the start of each pattern. Set a volume of 127 for the channel that will play the first part of the composite solo, and set the volumes for the other two to zero. Where the second part should take over, insert a volume of 127 for the corresponding channel and a volume of zero for the first part. Similarly, set the volume of the third part to 127 and the second part to zero where necessary.

On playback, you should hear each portion of the solo introduced from the corresponding sample, as the MIDI volume commands turn them up and down at the appropriate moment. There are no problems with synchronisation, since all the samples are playing back from the start each time. In reality, a little tweaking will undoubtedly be necessary to get the volume commands in just the right place, but this is far easier than all that splicing and moving of note events (and it eats up less MIDI bandwidth, too). If the volume changes are too abrupt, try cross‑fading between them, by adding multiple MIDI volume messages, and 'ramping' them up or down as necessary. Cubase's MIDI mixer page is invaluable for this kind of job.

Although this method is a little costly on sample memory and polyphony, it is very much quicker for producing the desired result — the choice is yours. A variation of this technique will allow you to 'mute' out any stray clicks, squeaks or other nasties, by turning down the volume for brief periods as the sample plays. Samples can also be faded in and out. I have achieved excellent results by getting a guitarist to hold a high note at the end of a solo, and then fading it out on playback, while slowly bending the pitch down by an octave or so!

4. Recording effects with your sampler

Once in a sampler, a great deal more can be done to a recording before laying it to tape. A particular effect that would have been risky to commit to tape (such as flanging, echo or compression) can be tried out with impunity. Simply sample the material to be effected completely dry. You can then experiment with different effects on this dry recording to your heart's content. When satisfied, you can commit the effected material to tape. This method also has the advantage of freeing up effects processors at mix‑down. If the effect is subsequently found to be undesirable, then the signal can always be re‑laid to tape without it.

5. Creating your own effects

Many effects can be created by the sampler itself without resorting to an effects processor at all. The simplest of these is to re‑trigger the sample to produce echoes. By using different velocity values to control both amplitude and filter, and by experimenting with the playback timings, a wide variety of delay/pre‑delay and multi‑tap echo effects can be created.

By playing two copies of a sample simultaneously, and slightly modulating the pitch of one of the samples with an LFO, typically 'whooshy' chorus and flanging sounds can be produced. Filter settings can have a profound effect on the result here too. Don't overlook the use of an envelope to modulate the pitch of the second sample — the flanging effect can be controlled in a much more predictable and repeatable manner.

6. Thicken to taste

The time‑honoured way to 'thicken' a sampled sound is to play two simultaneous copies, usually panned to left and right, pitching one slightly sharp and the other slightly flat. This creates a very satisfying stereo spread from a single sample, and is relatively simple to achieve. An alternative method is to reverse playback of one of the samples. This works particularly well on vocal pads or strings, where subtle variations in pitch occur throughout the duration of the sound. Playing one in reverse pits these variations against one another in a pseudo‑random manner, which can result in a more natural chorusing effect. A little extra pitch‑shifting or vibrato on one or both of the samples never goes amiss either.

7. The grand old days of 8‑bit

Time has a way of filtering the good memories from the bad. Nowadays, there are those who get all glassy‑eyed when recalling the low‑tech charms of grainy 8‑bit samples. But even now, it's quite easy to get 8‑bit‑type sounds from a 16‑bit sampler. All you need to do is under‑record your sample. This will force the sampler to use fewer bits to resolve the sample levels. The drawback to this is that you then have a very low‑volume sample that will struggle to match other levels from your sampler. There are a couple of ways around this. Firstly, if your sampler allows you to 'normalise' a sample level, this can be implemented to push the volume up to a usable level, whilst retaining the low resolution. Secondly, you could digitally mix the sample with itself inside the sampler until the volume attains a reasonable level. The third method is simply to record the sample back onto tape and re‑sample it at a more healthy volume.

8. Return to sender: sampler drum effects

I've described the next tip before in these hallowed pages, but I find it so useful that it's worth mentioning again here. When using a sampler for drums, it is often the case that various effects levels are required for specific sounds. A bass drum or hi‑hat might need to be left fairly dry, whilst a snare or tom might require a large amount of reverb. To avoid using up several outputs from the sampler (and the corresponding number of channels on the mixing desk), a single separate output can be used as an effects send. To do this, simply duplicate the sounds in the sampler, and allow them to play at the same time as the originals. Many samplers (such as the Akai S1000) allow more than one program to share the same MIDI channel, and this is ideal for the job. Feed the duplicated sounds to the separate output, and connect this output to an effects processor. The level of the individual duplicated sounds appearing at this output will then determine how much effect each sound takes on. The effects level of each sound can be controlled by editing the duplicate programme. For added sophistication, the duplicate sounds could be played over a different MIDI channel, and their playback controlled by a different track in your sequencer. Then, velocity values will control effects level — even down to a single beat.

9. Less fuzz, man

As a final thought, consider the following scenario. A guitarist plays a blistering solo, but on listening back to it, you think it would have sounded better with a little less overdrive on the amplifier. No problem. Make sure you record the guitar by feeding it through the sampler first, and then onto the amp. If you're not happy with the amp settings, you just make the necessary alterations to them, and play back into the amp from the sampler. Everyone is happy. The guitarist's first child is named after you. Whaddya mean? I think 'Paul' would make a nice name for a girl...