Computer Audio Reviews

Sidebar 3: Measurements

To measure Logitech's Squeezebox Touch, I used Stereophile's loaner Audio Precision SYS2722 system. (See www.ap.com and "As We See It" in the January 2008 issue.) For some tests I also used my Audio Precision System One Dual Domain and the Miller Jitter Analyzer. I used audio files fed to the Touch both over my WiFi network from my iTunes library and on a solid-state USB drive plugged into the Squeezebox. (For the latter condition, a version of the Squeezebox Server program runs on the Touch itself). The WiFi performance was a lot more robust than my old Squeezebox 3, using the same antique Apple Airport router. Before doing any testing, I updated both the Touch's Squeezebox Server program and the version running on my Mac mini.

The Squeezebox Touch correctly decoded files with sample rates of 32, 44.1, 48, 88.2, and 96kHz, and was compatible with the WAV, Apple Lossless, FLAC, MP3, and AAC file formats with which I tried it. However, when it came to bit depth, its compatibility with data formats depended on whether the data was being read from a USB drive or from the host PC. With the files sourced from the PC, the Touch correctly decoded 24-bit and 32-bit AIFF, Apple Lossless, and WAV files, but with the same files sourced from a USB drive, the 32-bit files played back with distortion and a 48dB positive level error.

The TosLink output correctly followed the sample rate and bit depth of the input file data, with the exception of 32kHz-sampled material, which appeared to be upsampled to 96kHz. Files stored on a USB drive and sampled at 176.4kHz and 192kHz were identified as having an "unsupported sample rate" and wouldn't play. However, when the same files were played back from the host computer, they were downsampled to 88.2kHz and 96kHz, respectively (footnote 1). It looks as if the Squeezebox Server program running on the Touch has less flexibility than the version on the host PC.

Measured at its RCA output jacks, the Squeezebox Touch's maximum output level at 1kHz was 2.06V, which meets the CD standard. The output preserved absolute polarity (ie, was non-inverting), and was sourced from an impedance of 599 ohms at high and midrange frequencies. This impedance increased inconsequentially, to 878 ohms, at 20Hz. The Touch's frequency response was flat within the audioband for both 44.1 and 96kHz data (fig.1), with a slight passband ripple in the top octave for 44.1kHz data (blue and red traces), and a slight droop above the audioband for 96kHz data (cyan, magenta). Channel separation (not shown) was superb, at >112dB in both directions below 3kHz, and still 100dB at the top of the audioband.

Fig.1 Logitech Squeezebox Touch, frequency response at –12dBFS into 100k ohms with 44.1kHz data (left channel blue, right red) and 96kHz data (left cyan, right magenta, 0.25dB/vertical div.).

The headphone output on the rear panel also preserved absolute polarity, and offered a maximum output level of 975mV at 1kHz. The output impedance was a suitably low 2 ohms over most of the audioband, rising to a still low 5.5 ohms in the low bass.

For consistency with the digital tests I have performed for Stereophile since 1989, I first examine a digital product's resolution by playing a dithered 1kHz tone at –90dBFS while sweeping the center frequency of a 1/3-octave bandpass filter from 20kHz to 20Hz. The result of that test with the Squeezebox Touch is shown in fig.2: with both 16- and 24-bit data, the traces peak at exactly –90dBFS, suggesting excellent DAC linearity, and no harmonic spuriae are visible. A spectral bump in both channels at 60Hz is presumably related to the power supply, but at –112dBFS, this is negligible. The increase in bit depth drops the noise floor by 7–8dB, implying a resolution of just over 17 bits, which is not bad for such an inexpensive product. However, this is not enough to resolve a dithered tone at –120dBFS (bottom traces in fig.2). Fig.3 repeats the spectral analysis using an FFT technique; the same increase in resolution with increasing bit depth can be seen.

Fig.2 Logitech Squeezebox Touch, 1/3-octave spectrum with noise and spuriae of dithered 1kHz tone at –90dBFS with (from top to bottom): 16-bit data, 24-bit data, dithered 1kHz tone at –120dBFS (right channel dashed).

Fig.3 Logitech Squeezebox Touch, FFT-derived spectrum with noise and spuriae of dithered 1kHz tone at –90dBFS with 16-bit data (left channel cyan, right magenta) and 24-bit data (left blue, right red).

Linearity error with 16-bit data (fig.4) was negligible to well below –100dBFS, and the Squeezebox Touch easily resolved the three DC voltage levels of an undithered 16-bit tone at exactly –90.31dBFS (fig.5), though with a little more noise than do the best digital players. Increasing the word length to 24 bits gave a good if noisy sinewave (fig.6).

Fig.4 Logitech Squeezebox Touch, left-channel linearity error, 16-bit data (2dB/vertical div.).

Fig.5 Logitech Squeezebox Touch, waveform of undithered 1kHz sinewave at –90.31dBFS, 16-bit data (left channel blue, right red).

Fig.6 Logitech Squeezebox Touch, waveform of undithered 1kHz sinewave at –90.31dBFS, 24-bit data (left channel blue, right red).

The Touch's analog output stage was comfortable driving low impedances. Even with a full-scale signal into 600 ohms, all the distortion products lay at or below 97dB (0.0014%), with the subjectively benign second and third harmonics the highest in level (fig.7). The Touch performed equally well on the demanding high-frequency intermodulation test, with the 1kHz difference tone resulting from an equal mix of 19 and 20kHz tones lying below 100dB, and the higher-order products at –110dB (0.0003%, fig.8). This graph was taken into 100k ohms; dropping the test load to 600 ohms increased the level of the higher-order spuriae by 6dB, but the spectrum remained otherwise clean (not shown).

Fig.7 Logitech Squeezebox Touch, spectrum of 50Hz sinewave at 0dBFS into 600 ohms, 24-bit data (left channel blue, right red; linear frequency scale).

Fig.8 Logitech Squeezebox Touch, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 100k ohms, 24-bit data (left channel blue, right red; linear frequency scale).

Finally, the Squeezebox Touch's jitter performance remained unchanged, whether it was playing the 16-bit diagnostic Miller-Dunn tone via WiFi or stored on a USB-connected drive. Fig.9 shows a narrowband spectrum analysis of the Touch's output while it decoded WiFi data. What harmonics of the low-frequency squarewave can be seen above the noise floor lie at the residual level, though the central tone is widened at the base due to random low-frequency jitter.

Fig.9 Logitech Squeezebox Touch, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz, 16-bit data via WiFi. Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz (left channel blue, right red).

Like the Centrance DACport, which I reviewed in June, Logitech's Squeezebox Touch offers excellent audio engineering with no sign that it has been compromised to reach its low price point. This is especially commendable when you consider that, in addition to its audio circuitry, the Touch includes a Linux-based server complete with touchscreen display. I am amazed that all this can be done for $299.99.—John Atkinson

---

Footnote 1: I believe the down-conversion is performed with the high-quality sox algorithm.