Synth Shop

"Synthesizer preservation via restoration and sampling."


The Web SynthArk

Sampling Process

Our Samples

EMS_Synthi_100

With good patch libraries as our starting point, our sampling process is significantly more involved than just recording and selling a bunch of WAV files. It begins with characterization, continues through the sampling phase, and ends with patch programming in the targeted sampler.

Preparation

Initially, the synthesizers to be sampled are calibrated. Then, both the subject synthesizer and a control (i.e. another synthesizer in the same family) are sampled, if possible (we have controls for most, but not all synths sampled). Any abnormalities can then be identified and addressed, when the samples are compared.

Characterization

Characterization includes research on the synthesizer architecture (hence the museum) and analyzing aspects of that architecture (e.g. LFO and EG features and rates). For subtractive synthesizers, the filter is the primary defining aspect of their sonic character. So, care is taken to capture this aspect in the samples (when feasible), without wasting memory on easily reproducible aspects, such as envelope generators after the filter.

Sampling

The patches to be sampled are stripped of any programming not significantly affecting their character (e.g. LFO, panning, amplifier EG). Each synthesizer patch is then sampled from C0-C7 (as applicable), looped and verified against the original patch for accuracy. WAV file development ends here, but sampler (e.g. EmulatorX, ProteusX) development continues to Patch Programming.

Patch Programming

Once the samples are seamlessly looped, we use the characterization data to re-implement performance features (LFO, panning, amplifier EG, effects) in the destination sampler (EmulatorX, ProteusX) as originally programmed. This provides performance flexibility when using the sample patches, without sacrificing significant accuracy of the original synth patch. Additionally, the volume of each patch is adjusted to avoid clipping with normal playing (with EmulatorX/ProteusX headroom set to -3dB). Once the patch is complete, we compare it once again to the original synth.

Editions

Sample banks can be provided in "Studio", "Live", or "BareBones" editions. Both Studio and Live versions are sampled after the synthesizer's filter (with a gated amplifier) in subtractive synths to retain the character of the synth, without sacrificing sample time, due the to easily reproducible amp envelope and LFOs. The BareBones version contains oscillator samples only, allowing the user to add filtering and performance controls via the host sampler (e.g. EmulatorX, EXS24, Mach5, Kontakt). The Studio sample banks are optimized for accuracy, being sampled at 24bit/96kHz on each minor 3rd, while the Live sample banks are optimized for memory usage, being sampled at 16bit/48kHz on each tritone. Both the Studio and Live versions use the host sampler (e.g. EmulatorX, EXS24, Mach5, Kontakt) performance controls (e.g. amplifier envelope generator, LFOs, etc) to allow user customization.