Vintage synths are perfect for sampling, and modern multisampling software makes the process easy and creatively rewarding. We show you how it's done in our in‑depth guide.
For many of us, adventures in sampling began with hardware samplers from Akai, Emu and Roland that offered a new way to capture and manipulate sound. But, as exciting as the early days of sampling were, hardware samplers were hampered by limited memory and storage facilities, so creating multisampled instruments was labour intensive and time consuming.
As music technology evolved, a new generation of samplers emerged in software form, free of the memory and storage limitations of the past. Better still, not only did these software samplers offer features previously unavailable to their hardware predecessors, they improved upon the sampling workflow immeasurably, especially with regard to more ambitious multisampling projects. Sampling could now be accomplished far more efficiently than ever before, and the sonic improvements were significant. Evidence of this are the excellent multi‑gigabyte sample libraries from companies such as East West, Vienna Symphonic Library and Project SAM, as well as the development of outstanding sample‑based software instruments from the likes of Spectrasonics.
Despite this, much of the focus of 'user sampling' over the past few years has been on creating drum kits and rhythmic loops, which make little use of these improvements in multisampling. Many avoid the sampling process entirely, instead relying on software samplers merely as sample playback devices, but with the tools available today there's no better time to dig deeper into sampling and create your own sample libraries. A great source for material may be sitting there right in your studio, in the form of vintage analogue synthesizers that you can sample. In this tutorial, I'm going to show you how to accomplish that.
The steps required are relatively simple. One: record samples to your DAW of choice. Two: name the samples appropriately and export for use in your software sampler. Three: import and map the samples in your sampler for playback.
In this tutorial, I'll use some classic analogue synths of yesteryear as examples for the sampling process: the Oberheim Matrix 12, the Sequential Circuits Prophet 600, and the Moog Taurus Bass Pedals. Before we begin the sampling tutorial, we'll first need to address a few technical issues and then determine what really needs to be sampled in the first place.
Sample rates: Modern samplers convert sample rates on the fly with little CPU penalty, so use either 44.1kHz or 48kHz. Most of the work I do these days is in film and television, so my studio is configured for 48kHz and I use that sample rate. I wouldn't advocate higher-definition sample rates for analogue synths, but I'll always opt for 24‑bit over 16‑bit depth for increased resolution. This may incur some processing penalty on larger projects.
Sample note length used to be a significant factor. Traditionally, the workaround was to sample the shortest note length possible and loop a portion of the sound. This is no longer an issue, so I sample on the long side. If I don't have time to loop the samples, I'll be sure to sample six‑ to eight‑second note lengths.
How many notes? The number of notes you should sample is often directly proportional to how much time you have to map and loop them afterwards (we'll examine that part of the process in detail shortly). In this tutorial, we'll cover a few scenarios to give you an idea of what's involved. On the short end, we'll sample the Prophet 600 with just two notes per octave. At the other extreme, the Matrix 12 will be sampled every two notes. I found it made a noticeable difference on many patches to sample with such frequency. My general rule is 'the more the better', so decide for yourself after making a test. Finally, we'll sample the Taurus pedal chromatically, capturing all 25 available notes. If you're new to sampling, it may be best to sample just two or three notes per octave. Advanced users will want more exacting results and to sample more notes. The best way to learn is by doing, so why not power up one of your hardware synthesizers and follow along. If you don't own one, just download the tutorial files from /sos/jul11/articles/sampling-vintage-synths-audio.htm and import them into your sampler of choice.
What to sample? Deciding on what to sample is the next step. I suggest you choose only those sounds that are unique to that particular instrument. It's unnecessary to sample relatively insignificant sounds that turn up elsewhere in abundance. If your synth doesn't offer velocity sensitivity, you may wish to manually change the filter settings; first capturing an initial set of multisamples, then a second set with a different filter setting, which will be layered in your sampler. There is a distinct advantage to capturing the sound of a synth's filter in as many ways as you can, as it is often largely responsible for the distinctive character of the instrument in the first place.
Now we're ready to begin. For the first example, I'll sample the Taurus. I've set my DAW's sample rate to my studio's default of 48kHz, and the bit depth to 24. Assuming that everything is routed properly and that there's a respectable signal level, it's time to begin sampling.
The Taurus has 13 physical pedals, ranging chromatically from C1 to C2 an octave above. Stepping on the Octave footswitch will raise the pitch in order to access the notes from C2 to C3. Once I engage recording, I'll step on each pedal and hold the note for about eight seconds, then allow two or three seconds of silence, and then step on the next pedal for eight seconds. When I've finished, I'll have one sound file with 25 notes sampled chromatically, note lengths of eight seconds, and a few seconds of silence between each note.
The next step is to trim, name and export the samples. In order to play them back within your sampler, you'll have to map them to the appropriate keys so that your sampler will know what pitch the note was sampled at. Most samplers have the ability to do that for you automatically, but the sample files must be named properly.
Most DAWs have the ability to trim, or strip, silence from sound files. In Pro Tools and Logic Pro, this feature is conveniently labelled 'Strip Silence', and can be used to create individual files that can be named appropriately. You accomplish this by setting a threshold for the strip function to engage, being sure to add some 'pad' at the end so as not to cut off the ring-out of the sample too abruptly (Logic calls that the 'Post Release Time').
It's helpful to name your sound files in a specific way: I suggest using the synth name, a group descriptor (more later on why that's helpful) and finally a sequential number — all separated by an underscore or dash. In the case of my Taurus samples, the names could be Taurus_Vari_01, Taurus_Vari_02, Taurus_Vari_03 and so on... The name describes the synth, the patch (the Taurus has only four presets, the first one being named Variable), and finally the sequential number of the sample (01 representing the first note sampled, and in this case the lowest pitch). You could just as easily name the samples by the root pitch (C1, C#1, D1, and so on), or MIDI note number (36, 37, 38, etc), but I'm naming sequentially to demonstrate how that works with a sampler's auto‑mapping functionality.
Once you've named the samples, select them and export them to a folder on the drive where you store your samples. Most samplers will stream files directly from disk, so I recommend that you store your samples on a drive separate from the boot drive.
The next step is to import these samples into your sampler and map them for playback on the keyboard. While most samplers can automatically map sequentially numbered files, some do so a little differently than others, and it's worth giving a few examples.
In Native Instruments' Kontakt:
It's just as easy to do this in Avid's Structure sampler for Pro Tools:
In Logic's EXS24:
It took less than one minute to import and map a series of sequentially numbered samples with each of these samplers!
In my next example, I'm going to use MIDI to trigger the Prophet 600 and sample just two notes per octave, before labelling each sample with the pitch name. I've created a patch that I'd like to sample with three different filter settings. Most samplers allow you to layer your samples, and in this instance I'd like to be able to trigger a different filter setting depending on how hard or soft I play. Samplers accomplish this by creating separate 'parts', or 'groups' that serve as a container for a specific type of sound. An example of this feature would be to label different groups with musical descriptors such as p, mp, mf and f, depending on the dynamics of the sound.
Another example might be to add a descriptor of the filter setting, and in this particular case I'm going to label my Prophet 600 samples as 'bright', 'edgy' and 'dark'. Once the sampler has identified these different sets of samples as groups, it will allow you to trigger them separately in a number of ways using velocity, the mod wheel or with key switches.
Key switching is a method that samplers use to trigger by note different groups of samples for playback. One example of this technique is a patch that lets you switch between a strummed and picked guitar sound. Another example is a patch that will switch between sustain and staccato bowing techniques.
I'm going to automate the sample recording process by creating a MIDI template. Regardless of which DAW you sequence and record in, the steps are primarily the same:
In the Prophet 600 example, the MIDI notes will be eight seconds long with three seconds of silence between each note. You can then change the filter setting on the synth, before recording again to a different track, using the same MIDI notes. I've chosen to sample two notes per octave, beginning with C2 and ending with C5. The sampling procedure is then as follows:
The next step is to export these samples to your samples drive and then import them into your sampler.
In Kontakt, simply drag the samples into the Zone grid. Don't be concerned with the placement of samples; Kontakt will take care of that for you in the next step. Now open the Mapping Editor, and under the Edit tab, select Edit Menu / AutoMap Setup (see screens above). This is where you instruct Kontakt on how to interpret the name. Kontakt can be directed to ignore the synth and patch name: use the filter descriptor ('Dark' in the case of my examples) as the Group Name, and assign the note to 'Set to Single Key'. This maps each note to a single key and creates a Group named 'Dark'. Next go to Edit Menu / Batch Tools and select 'Move Root Keys to Center'. Finally click on the 'Auto Spread Zone Key Ranges' button and Kontakt will fill in all the space between the samples by stretching them above and below the root note. Kontakt will pitch-shift any notes that you didn't sample directly.
It's even easier to map the Prophet samples in Structure: In the 'Import Samples' dialogue box, select 'By Root Key' from the 'Map to Key Zone' panel and 'From Name' and 'Note' from the 'Set Root Key' panel. You have the option to fill notes up or down, or centre them next to the root key to fill in the gaps you haven't sampled.
For my final example, I'll look at multisampling the Matrix 12. In the previous example, I sampled only two notes per octave. The inherent problem with a sampler pitch-shifting notes is that the notes won't have the same length on playback. One way to get around that is to loop the samples (we'll talk about that next) and utilise a release envelope. The other way is to sample more notes, and in the case of the Matrix 12 example I'm going to sample every other note, eight seconds in length, from C1 through C7. Not only that, I'm going to sample five velocity levels. And to top it off, I'm going to loop every sample.
Before you suspect that I've gone mad and have nothing better to do than sit around and sample all day long, I'm going to introduce you to my preferred sampling solution, in the form of a software programme from Redmatica called AutoSampler. First reviewed in SOS in February 2005, AutoSampler now outputs EXS24, Kontakt, NNXT and Structure formatted patches — all in one go. And besides recording and mapping samples automatically, it will also loop samples. The looping features are particularly robust, and in my testing I couldn't find a bad loop. It has actually made the sampling process fun! So for the final tutorial I'm going to go have a cup of coffee while AutoSampler does all the work for me. Life couldn't be easier!
Now on to one of the more time-consuming aspects of this process: looping. For the vast majority of playing situations, six- to eight‑second samples may be all that's needed to avoid having to loop your samples, although you'll probably have to tweak the release envelope of your patch. However, there still may be situations where you need the notes to sustain longer. So if you absolutely must loop, read on...
Like many of their hardware counterparts before them, software samplers typically offer some way to find a good-sounding loop. Depending on the source, and whether it's mono or stereo, this may yield less than acceptable results in the form of a noise at the loop point. In order to help a loop work more effectively, most samplers allow you to create a crossfade at the loop points to blur the transition.
Here are some general guidelines to adhere to when looping samples:
The best sample-loop editor I've come across belongs to Redmatica's Keymap Pro. EXS24 users have an advantage, as EXS24 supports a direct link to Keymap, but luckily Keymap will also import and export Kontakt, Structure and NNXT patch files, so users of those samplers can take advantage of its looping facilities as well. Kontatk and Structure have some fairly robust looping tools built in, but Keymap Pro has the ability to automatically loop sample files in batches and far exceeds the capabilities of any sampler. You would be hard pressed to give it a sample that it couldn't loop seamlessly with the huge number of loop options it offers.
Now that the hard work is done, it's time to get creative and have fun. This is your opportunity to make your sampled instruments do things the original synths never could. It can be pretty wild sampling a 12‑oscillator mono sound from the Matrix 12 and then playing it back polyphonically. Talk about phat! A six‑note Prophet 600 can rival a Prophet 10 now, complete with reverbs, delays, chorus and whatever built‑in goodies are at hand inside the sampler. Try assigning different filter groups to the mod wheel, or layer a Prophet sample with an Oberheim sample and switch between them with velocity...
Nothing beats a vintage synth. Over the years, I've owned some pretty amazing ones, and having sampled them gives me that secure feeling of knowing that even if they give up on me one day, I'll have captured the best that they have to offer.
When it comes to effects, the guiding principle for years has been to record your samples dry. For the most part, I still agree that's the way to go, especially if you're recording fewer samples per octave. In general, samples with effects applied can be much more difficult, or impossible, to loop. Most samplers have an ample supply of built‑in effects to use on your dry samples (Structure has a particularly large number of great-sounding effects). There are a few situations in which I've broken that rule, however. I have a wonderful Eventide delay in my rack, and some great-sounding Lexicon reverbs — both of which far surpass anything built into a sampler. So on occasion, when I'm sampling a synth such as the Prophet 600 that has a mono output, I will run it through an external effect and sample that output in stereo. Doing this selectively on a few types of synth patch can really bring them to life. A vintage analogue synth pad run through some shimmering stereo delays from the Eventide makes an awesome-sounding sampled patch!
Here are some suggestions for additional information about sampling: