Lab evaluation of the UP4 Ver.22 was done using the USB-C port, i.e. connected as a DAC/Headphone Amplifier. The amplifier gain was set to “High” and the digital filter to Linear Phase Slow. The 3.5mm TRS jack was used as the output.
The device delivered 96.7mWrms per channel at the typical limit of clipping (0.3% THD+N) into a 32 Ohm load ((1.76Vrms)) which, as expected, dropped significantly to 6.7mWrms (2Vrms) when the load was set to 600 Ohm. It is clear that headphones with such nominal impedance are not designed for use with devices such as the UP4, where the emphasis is on low power consumption and low voltages are involved.
Distortion figures ran low, as we measured values in the region of 0.007% at 1/3 of maximum power with a 32 Ohm load and in the region of 0.002% with a 600 Ohm load. The intermodulation distortion (per SMPTE) was correspondingly low, with measured values in the region of 0.005% and 0.003% respectively.
The noise of the device was also low, around -85dB(A), referenced to 1kHz, at 1/3 of maximum power into 32 Ohm. Performance on channel crosstalk was typically good for the type of output connection, which was via a 3.5mm TRS jack.
The frequency response was good, with negligible deviations in the low-frequency range and a cut-off in the high-frequency range that starts relatively early, which is probably due to the selected, "Slow", digital filter. Channel balance was exemplary and it is also positive that there are no obvious differences in the response as a function of the load, except for a slight increase in range, in the case of 600 Ohm.
The output spectrum, for a 1kHz signal at the input and a reference level that corresponds to 1/3 of the maximum power at 32 Ohm, is very clean, the stronger harmonic being the second (at -80dBr) and the rest of the findings being below - 100dBr, a very good performance, indeed.

With the same conditions, but driving a 600 Ohm load, the amplifier showed an even better performance. The second harmonic is still the strongest but rests just above -100dBr.
The noise spectrum has no notable findings, being in the region of -120dBr, referring to the level corresponding to 1/3 of the maximum power at 32 Ohm. The device has no noise from the power supply (while connected via USB during the measurements) and does not seem to pick up serious level of high-frequency noises, with all cables in place.

Total harmonic distortion and noise (THD+N) change with frequency and level, with a 32 Ohm load, is at expected levels. Readings are in the 0.01-0.02% range, with values at lower levels being slightly increased.
When the load becomes 600 Ohm, the values decrease a bit, especially those related to measurement with higher levels. The behavior of UP4 Ver.22 does not change significantly, however.

The variation of the distortion as a function of the delivered power and the output load is also as expected. With a load of 600 Ohm the level of THD+N is lower but the overload point appears quite early, in contrast to the corresponding performance when the load is 32 Ohm. A point of interest is the fact that the values corresponding to the power measurement for 0.3% deformation are noticeably lower than the findings in the static measurements. This should probably be attributed to the measurement method (which is different between the static measurements and the sweep from which these curves are derived) and not the device itself. Based on what we've seen, the static measurements reflect better both the maximum power and the distortion behavior.

Previous | Next | More Reviews