Sampled instrument designers have always struggled with the Vibraphone – the ‘snapshot’ nature of sampling doesn’t really suit it at all; if you sample the single notes, then each notes tremolo will be out of time with the next, and of course the speed will be fixed at one rate anyway. Sometimes people go ahead and sample it anyway, and offer it out of sync with a choice of rates ( fast, medium, slow), but that tends to sound pretty wrong. The more common ‘workstation’ approach was always to simply switch the motor to off on the Vibraphone, record it static, then fake the tremolo in the synth/sampler using an LFO with a bit of filter and amplitude modulation. Which, in the context of workstation sounds, doesn’t do too bad a job, in that ‘Mock-Tudor’ style that workstation sounds have.
But about a year or 18 months ago, Dan hit on an idea – whether in the bath, I know not – and like many breakthrough ideas, it was beautiful in its simplicity.
Powell hypothesised that if one recorded the vibraphone twice, in a static state each time, once with the motorised fans fully open, in the vertical position (v) and then again with the fans completely closed, in the horizontal position (h), then by simply crossfading between the two with a simple sine LFO in the sampler, you would get a very natural tremolo effect (Tr).