Following on from last month’s vocal transformation techniques, Liam O’Mullane takes you through sampling the voice and MIDI control with VariAudio in Cubase.

Certain sampling techniques were born from the need to squeeze every last drop of memory from your huge metal box of sampling joy.  Like many era-based limitations, these workarounds created what is now considered a classic sound and in turn they can be used in the present day to borrow from an older musical vibe.

In modern times, though, the tools and subsequent techniques we have at our disposal are worlds apart and this month we will look at embracing these two worlds of sampling. From playing audio in its natural form to taking even the smallest snippet of a vocal and creating drums and instruments, we hope to encourage you to reach for your microphone in the quest for audio exploration. Audio examples and samples are included on the coverdisc, so you can hear exactly what is achievable and save you the embarrassment of making a few silly noises if required.

Sampling naturally
Sampling the voice and keeping it in its natural state can provide an organic sound while adding elements of fakery and electronica. A true human voice can be detected in a recording due to the completely un-repeatable nature of one note and the next. A singer can never quite perform a note to exactly the same length, same volume and same tonal change over time. Although a sampled voice is based on recordings of this natural instrument, a more static and repeatable playback is inevitable as you notice the tones remain the same as they play. This, however, is a familiar sound, from the early use of the Mellotron tape-based samplers to the now pitch-perfect vocals created with pitch correction. Vocal sampling can comprise anything from ooohs to arhhhs and requires recordings of a vocalist either for every pitch you intend to use or the use of sound manipulation to create other notes.

Before we get to the unnatural, we’ll go through the art of sampling multiple notes then maximising those samples in the simplest way. It is relatively easy to create a vocal samples instrument when using basic words like arhh as it is based on a single word. There are singing instruments out there that can turn note information and user-defined words into sung melody lines, but the level of piece-by-piece sampling that’s involved is beyond the scope of this tutorial. If it interests you, Vocaloid (by Yamaha) can do this for a solo vocalist and Symphonic Choirs (by Quantum Leap) covers various layers of singing choral parts.

To create your own arhh patch, first set up a virtual instrument to create a pure saw or triangle wave; this will create a pitch thats easy to hear while singing. Then chromatically create MIDI notes from the lowest to highest pitch you can sing. The note length is important too, so if the MIDI note is one bar in length, its best to leave a rest of one bar to allow for breathing between takes. If you keep the MIDI parts for each note seperate (two bars at a time), you can click each one and select your DAW’s loop-to-part function to quickly focus on each note as you go. Use the metronome as well to encourage clean start and end points for your samples, making them more uniform in duration.

Before recording, select your DAWs multiple-take mode so each take will be kept for you to later choose from or use as an alternate take. The more similar each note is in its vocal tone, the more continuous they will sound when played back from a sampler, so the more takes you do, the better.

If you simply want an instrument that will play samples as they are, export them from your DAW one at a time with the MIDI note information in the title (eg ‘Male Vocal Arh C2’). Then drag them into a sampler so that each file is assigned to relevant MIDI note for logical playback. A DAW such as Logic or Live will give you the option to drag the audio directly from the audio track to the EXS24 or Sampler respectively. Although this makes life a lot simpler, it is still worth rendering each sample as a new audio file before loading it into the sampler, so the files are only as long as they need to be, instead of a small portion of a much larger file. In Logic this is called Convert to New Audio File; in Live its called Consolidate.

Example 1 on the coverdisc demonstrates a few notes from a multi-sample instrument with each vocal sample lasting one bar at 120BPM. Example 2 contains the same files, but the samples have been looped at their best looping points. This is an old technique used to create sampled instruments when memory was tight, and also to enable the user to play each note for as long as they wanted. After the attack portion of the sound has passed, the sample will loop around the sustained section of the audio – this should ideally be as unnoticeable as possible, which takes some trial and error to perfect for each vocal sample.

As well as a playback start control, a sampler has various loop options and selecting Forward-Only loop mode will create the most natural-sounding loop as it cycles from the loop end position back to the loop start. The loop’s length needs to be short enough that it doesn’t have too much tonal change within it, but long enough that the sample doesn’t suddenly sound synthetic in tone. This is why a stable pitch after the attack period makes this task a lot easier to achieve. A part with vibrato or slight pitch changes will be very noticeable when looped. As a special effect or rhythmic trick this is fine, but for a less aggressive loop, the more stable the sustain portion in pitch and tone, the better. Most samplers will show you the loop position on the waveform as a highlighted block. This enables you to look for portions of the audio with similar volume and waveform density, which will produce a more natural-sounding loop.

It’s also important to make sure that the loop creates a continuous waveform without any jumps in position from the end to the beginning. If the waveform position is even slightly out, an audible clicking noise will be heard. It can be avoided, however, by making very fine adjustments to the loop start or end points until it stops, or a fade option can cross-fade the last part of the loop over the beginning to smooth the sound. Some samplers give you a double view of the start and end points side by side to make this task easier.

As in the world of synthesis, ADSR can be a very effective sound-shaping tool and example 3 demonstrates the effective use of basic ADSR settings. It contains a sampled vocal melody as it moves from a tight sound with short release and low sustain to long release and high sustain, making them overlap.

Words to rhythm
Using longer phrases of sung or spoken parts, you can use a sampler as a form of scrubbing tool by moving the start position within the audio file. As you pass over certain sounds you will start to hear some drum-like tones. Low sounds lend themselves to bass drums and hi-hats can be found in areas of sibilence. By fine-tuning the sample start times you can get more or less attack as desired and then either use the sample end control or Amp ADSR to create a deliberate ending. After mapping different instances of the same audio across MIDI keys you can create a variety of short sounds which with a little further tweaking can be very effective in a rhythm section of a track.

Example 5 is a demonstration of this process to create a small vocal drum kit. A strong pitch envelope is added with zero attack, short decay and zero sustain to give them a more drum-like transient attack. To further tighten your drum sounds you can apply some basic EQ principles by removing the unwanted high or low end with a low- or high-pass filter as appropriate. A lot of snare-type sounds can be transformed to hi-hats if you remove all but the higher frequencies.

Another approach is to go back to source and try to record vocal drum hits yourself. Example sounds are provided on the DVD and comprise a range of completely clean and unprocessed vocal hits including kicks, snares, cymbals and miscellaneous noises. These can work as standalone samples with a soft organic tone or they can be used as building blocks for layering with synthetic drum parts or instruments When using a pitch ADSR for adding attack, increasing the decay time can turn kick samples into toms by slowing the pitch descent.

Applying timestretch to your samples is a very effective tool for drawing out the length of vocal drum hits and has a smoothing effect on tone as well. You will have to pre-process the audio inside your DAW first, then export the new samples to then re-import into your sampler. Example 6 is a demonstration of various hits followed by a more stretched version each time. You should be able to get an idea of the effects that can be achieved by stretching out short, soft hits until they become pitched, musical tones.

Words to melody
The looping skills from natural sampling can now be used in the arena of audio abuse to create one continuous sound. If there is a constant enough audio signal in any source of vocal recordings, you can create wavetable-sounding effects by moving the loop point around the audio. A more synth-like shape can then be moulded using Amp ADSR. Filter modulation through either ADSR or an LFO can transform a harsh, loop-generated tone into a rich, filter-swept synth sound. Audio example 7 demonstrates just how unique a vocal-based instrument can sound and how animated they become with just a little bit of pitch and filter modulation.

The Vocal Bass samples folder has various examples of one-note sounds. Example 8 takes one of these bass samples and loops the sustain portion in a more deliberate and rhythmic way. This mimics a popular technique used for sampling synth samples and produces a similar, rhythmic effect to using an unsync’ed LFO to modulate a filter cut-off. It also creates another current preference by having an increase in speed as you play higher in pitch and down as you play lower. In a rigidly quantized composition, this single, out-of-sync timing can produce a funk effect against a robotically timed backdrop.

An alternative approach is to implement the time-preservation techniques discussed in the Artificial Multi-sampling box, as with pitch changes, the loss in quality and definition can be a creative tool in itself. If you find this effect creates a new tone you would like to use, you can re-sample the audio and perhaps then use that note with classic looping, or use it as a new source to create time-preserved copies from to create new surrounding notes which hold the new samples characteristics.

When sampling sounds without any discernible pitch, or loops that are so short they create a new perceived pitch, you need to tune the samples to a reliable source. This can be a chromatic plug-in tuner that you would use the same way as a guitarist, or simply a guide note to match to by ear. You don’t have to work to middle C, just make sure that the reference note you use corresponds to its mapped MIDI note.

Pitch perfect
There are, however, some situations in which it can be very hard to tune a sample due to pitch changes within it. It could be that the sample actually sweeps over various notes or its harmonic content is very hard to discern a pitch from. If this is the case and you have a particularly discordant sampled sound, just try to pick the best tuning possible. It’s important to play with fine-tuning by cents as well because this can help a sample fall in line with the clearly defined tuning of semitones. Fine-tuning can also work to detune an instrument from the universal tuning of the song its in, which creates a sense of slight discomfort and reduces the sense of electronic sterility. A good example of this is the tuning of the bass sample in Cold As Ice by M.O.P. It’s so sharp that it is almost a semitone too high, which helps make the bass part stand out from the track. But, in general, a strong difference between synthetic recordings and those including real instruments is a matter of tuning accuracy. So slightly detuning your instruments from each other is always useful for pronounced separation.
 
Synthetic and organic layers
Whether you’re wanting a richer soundscape with a natural and organic feeling or just want to create sounds that are hard to define, layering can be approached in the same manner as traditional synthesis. You can therefore add rich waveforms to fill out the high and mid-frequencies for larger-sounding pads and choirs or pure sine waves to create that solid, dominating low end required in most modern production.

Example 9 is played with and without a double octave sine wave sub, whereas example 10 uses the natural vocal samples used in earlier examples, blended with a detuned, sawtooth-based pad effect. Example 12 uses a plain square wave lead sound, first played in solitude then with a variety of short, looped vocal sounds, each demonstrating the sheer range of tones you can achieve by using the voice as an unlimited sample source.

Vocal approach
You now have a pretty big audio-manipulation toolkit to utilise the tonal possibilities available with a computer, microphone and your voice. Next time you tire of your samples or synth collection, plug in and make some silly noises for your next killer tone.

This feature first appeared in Music Tech Magazine issue 89
Filed under Home, General Features, Features

New users, please register here

Hector Gomez - 13 November 2011, 10:56 PM

thank you !!!

• Latest
• Interviews
• General Features
• Software Workshops
• Q&A
• Ten Minute Masters
• Ableton Live Tutorials
• Cubase Tutorials
• Guitar Tech
• Logic Tutorials
• Mastering
• Native Instruments
• Pro Tools Tutorials
• Reason Tutorials
• Sonar Tutorials
• TV channel
• Back Track

• Effects
• Internet
• Live
• Mastering
• Music-to-picture
• Recording
• Sampling
• Sequencers
• Synthesis