The TransFormSynth is Not a Sampler...

Wikipedia defines a sampler as 'an electronic music instrument closely related to a synthesizer. Instead of generating sounds from scratch, however, a sampler starts with multiple recordings (or "samples") of different sounds, and then plays each back based on how the instrument is configured.' The relevant articles are here:

Wikipedia on sample based synthesis

Wikipedia on samplers

Thus a sampler begins with recorded snippets of sound (typically from a well known musical instrument) and attempts to provide a keyboard (or MIDI) controlled emulation of that instrument. The goal of the TransFormSynth is quite different: to generate lively and complex sounds with spectra (timbres) that can be precisely controlled. This is accomplished by separating the tonal portion of the sound (the "peaks") from the noise portion of the sound and processing them separately. The use of a sample is more a catalyst and guide for the (re)synthesis than a template for an emulation.

Here is how the separation is accomplished in the TransFormSynth.

Here is how the separation is used in the TransFormSynth.

Try choosing random sounds and patches and you'll be amazed at the richness and variety of the output (just press the button called 'random patch and sound'). Or take a single sound and play with several different patches: you'll hear some of the variety of which the synth is capable. Or take a single patch and play it with several different samples: you will notice certain similarities in the treatment.

If what you want to do is to regurgitate the samples that are put in, then you really should be using a sampler like an Akai S1000 or NI's Kontakt (or any of the others from the Wikipedia page). Nonetheless, it is possible for the analysis-resynthesis of the TransFormSynth to be nearly transparent and to reproduce the input sound quite accurately -- but it requires that a number of parameters be set correctly.