femto clock is mainly useful to sync up the digital stream from a computer
Other than Asynchronous Rate Converter used in some DACs (like Benchmark), most of them use PLL (Phase Locked Loop) and have variable D/A converter clocks with S/Pdif transmission, to adjust conversion rate to average incoming words rate (otherwise samples would be lost). It used to be done with VCO (Voltage, Controlled Oscillator) adjusted by comparator comparing phase of incoming signal to phase of the VCO util they are even (same frequency). Today it is likely all in digital domain, but still, D/A clock has to follow incoming signal rate, hence it is variable. These "Femto Clocks" are very stable, I'm sure, but average frequency stability is not important since we cannot detect such tiny clock differences/changes. Jitter of the clock is important and likely audible when above about 50ps. It is "picoseconds" - not femtoseconds, making me suspicious of snake oil. With async USB samples come in packets (frames). Each frame delivered at some fixed rate (like 1kHz) contains multiple samples. DAC places them in the buffer and upon buffer over/underflow signals back to increase or decrease number of the samples in next frames. That way D/A clock is fixed, but samples are not lost. Unfortunately connecting computer also connects high frequency noise. There are extremely low noise switching power supplies (SMPS) used by companies like Rowland or Benchmark, but those used in computers are cheap and extremely noisy. This noise is finding any alternative return to ground, including one thru DAC grounds. Optical isolation with USB would be my first choice of action.