iFi Audio Go Bar DAC-Headphone Amplifier Measurements

Link: reviewed by James Hale on SoundStage! Xperience on July 1, 2022

General Information

All measurements taken using an Audio Precision APx555 B Series analyzer.

The iFi Audio Go Bar was conditioned for 30 minutes at 0dBFS (2Vrms out) into 300 ohms before any measurements were taken.

The Go Bar allows for a USB input only. There are two headphone outputs: one unbalanced over 3.5mm TRS: and one balanced over 4.4mm TRRS.  Comparisons were made between unbalanced and balanced outputs for a 24/96 0dBFS input, in terms of noise, THD, and dynamic range. Other than the extra 6dB of gain and output voltage, the two outputs were virtually identical.

The Go Bar offers an in-ear-monitor (IEM) matching selector, which lowers output voltage for sensitive IEMs and increases the output impedance; this switch was left in the Off position for the measurements. There is also a Turbo Mode associated with the use of volume. Engaging Turbo Mode allows for the Go Bar full output-voltage potential. Turbo Mode was left engaged for the measurements. There are also four user selectable digital filters. All measurements below, unless otherwise stated, are for the balanced output, using the standard STD filter. The four filters are described as follows in the Go Bar manual:

• Bit-Perfect (BP): no digital filtering, no pre- or post-ringing
• Standard (STD): modest filtering, modest pre- and post-ringing
• Minimum Phase (MIN): slow roll-off, minimum pre- and post-ringing
• Giggs-Transient-Optimized (GTO): upsampled to 352kHz or 384kHz, minimum filtering, no pre-ringing, minimum post-ringing

Because the Go Bar exhibits considerable noise above 20kHz (see FFTs below), our typical input bandwidth filter setting of 10Hz-90kHz was necessarily changed to 10Hz-22.4kHz for all measurements, except for frequency response and for FFT measurements. In addition, THD versus frequency sweeps were limited to 6kHz to adequately capture the second and third signal harmonics with the restricted bandwidth setting.

Based on the accuracy and repeatable results at various volume levels of the left/right channel matching (see table below), and the lack of analog inputs, the Go Bar volume control is likely applied in the digital domain. The volume control offers 53 steps in 2dB to 0.5dB increments.

Volume-control accuracy (measured at headphone output): left-right channel tracking

Published specifications vs. our primary measurements

The table below summarizes the measurements published by iFi Audio for the Go Bar compared directly against our own. The published specifications are sourced from iFi’s website, either directly or from the manual available for download, or a combination thereof. With the exception of frequency response, where the Audio Precision bandwidth was set at its maximum (DC to 1MHz), assume, unless otherwise stated, the coaxial digital input (1kHz sine wave sampled at 96kHz at 0dBFS/2Vrms out), the balanced line-level output into 300 ohms, using a measurement input bandwidth of 10Hz to 22.4kHz, and the worst-case measured result between the left and right channels.

* into 300 ohms, THD for the left channel was the same as the right, at 0.0026%

Our primary measurements revealed the following using the coaxial input and the balanced output (unless specified, assume a 1kHz sine wave sampled at 96kHz at 0dBFS/2Vrms out, 300 ohms loading, 10Hz to 22.4kHz bandwidth):

Frequency response (16/44.1, 24/96, 24/192)

The plot above shows the Go Bar frequency response as a function of sample rate. The blue/red traces are for a 44.1kHz sampled dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 96kHz sampled dithered digital input signal from 5Hz to 48kHz, and, finally, orange/pink represents 192kHz sampled data from 5Hz to 96kHz. The behavior at low frequencies is the same for the digital input—perfectly flat down to 5Hz. The behavior at high frequencies for all three digital sample rates is as expected, offering sharp filtering around 22k, 48k, and 96kHz (half the respective sample rate). The -3dB point for each sample rate is roughly 21.1, 45.9 and 89.7kHz, respectively. This corroborates iFi’s claim of a frequency response of 20Hz-45kHz (-3dB). It is also obvious from the plots above that all three responses offer “brick-wall”-type behavior. In the graph above and most of the graphs below, only a single trace may be visible. This is because the left channel (blue, purple or orange trace) is performing identically to the right channel (red, green or pink trace), and so they perfectly overlap, indicating that the two channels are ideally matched.

Frequency response (96kHz, xBase on/off)

The plot above shows the Go Bar frequency response with and without xBass engaged, for 96kHz dithered sampled data. The blue/red without xBass and the purple/green traces are with xBass engaged. We can see that xBass applies about 8dB of boost at 20Hz, and is flat by about 300Hz.

Frequency response (44.1kHz, all four filters)

The plots above show frequency-response for 44.1kHz sampled input data for filters 1 through 4 into a 300 ohm-load for the left channel only. The Standard (STD) filter is in red, Giggs-Transient-Optimized (GTO) in purple, Bit-Perfect (BP) in green, and Minimum Phase (MIN) in pink. The graph is zoomed in from 1kHz to 22kHz, to highlight the various responses of the filters around the “knee” of the response. At 20kHz, the STD filter is at +0.1dB, the GTO filter is at -0.3dB, the BP filter is at -3dB, and the MIN filter is at 0dB. The STD and MIN filters have nearly identical frequency response behaviors.

Note: the filter characteristics are described under General Information section above. Our measured frequency responses generally match the descriptions provided by iFi Audio.

Digital linearity (96kHz)

The graph above shows the results of a linearity test for 96kHz sampled input data. The digital input is swept with a dithered 1kHz input signal from -120dBFS to 0dBFS, and the output is analyzed by the APx555. The ideal response is a straight flat line at 0dB (i.e., the amplitude of the digital input perfectly matches the amplitude of the measured analog output). The data is perfectly linear down to -100dBFS and only -1dB or better at -120dBFS.

Impulse response (44.1kHz, STD and BP filters)

The graph above shows the impulse responses for the first two filter types (STD and BP), for 44.1kHz sampled dithered data, using the Audio Precision’s Transfer Function/Impulse Response function. STD is in blue and BP in pink.

Note: the filter characteristics are described under General Information above. Our measured impulses responses generally match the descriptions provided by iFi Audio.

Impulse response (44.1kHz, GTO and MIN filters)

The graph above shows the impulse responses for the two other two filter types (GTO and MIN), for 44.1kHz sampled dithered data, using the Audio Precision’s Transfer Function/Impulse Response function. GTO is in purple and MIN in green.

Note: the filter characteristics are described under General information above. Our measured impulses responses generally match the descriptions provided by iFi Audio.

J-Test

The plot above shows the results of the J-Test test for the USB input measured at the balanced line level output of the Go Bar. J-Test was developed by Julian Dunn the 1990s. It is a test signal—specifically a -3dBFS undithered 12kHz square wave sampled (in this case) at 48kHz (24 bits). Since even the first odd harmonic (i.e., 36kHz) of the 12kHz square wave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sine wave (the main peak). In addition, an undithered 250Hz square wave at -144dBFS is mixed with the signal. This test file causes the 22 least significant bits to constantly toggle, which produces strong jitter spectral components at the 250Hz rate and its odd harmonics. The test file shows how susceptible the DAC and delivery interface are to jitter, which would manifest as peaks above the noise floor at 500Hz intervals (e.g., 250Hz, 750Hz, 1250Hz, etc.). Note that the alternating peaks are in the test file itself, but at levels of -144dBrA (fundamental at 250Hz) down to -170dBrA for the odd harmonics.  The test file can also be used in conjunction with artificially injected sine-wave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The FFT shows some of the alternating 500Hz peaks in the audio band but at low levels—below -130dBrA. This is an indication that the Go Bar should not be sensitive to jitter.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (STD filter)

The plot above shows a fast Fourier transform (FFT) of the Go Bar balanced output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sine wave at -1dBFS sampled at 44.1kHz (purple/green), using the STD filter. The sharp rolloff above 20kHz in the white-noise spectrum shows the implementation of the brick-wall-type reconstruction filter. There are two imaged aliasing artifacts in the audio band above the -135dBrA noise floor, at around 6kHz (-110dBrA) and 13kHz (-115dBrA). The primary aliasing signal at 25kHz is at -115dBrA, with subsequent harmonics of the 25kHz peak near -80 and -70dBrA.

Note: the filter characteristics are described under General information above. Our measured FFTs generally match the descriptions provided by iFi Audio.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (BP filter)

The plot above shows a fast Fourier transform (FFT) of the Go Bar balanced output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sine wave at -1dBFS sampled at 44.1kHz (purple/green), using the BP filter. As advertised, the BP filter uses no filter at all, which is evident in the very slow rolloff in the noise spectrum. As a consequence, there are obvious imaged aliasing artifacts in the audio band above the -135dBrA noise floor, as high as -90dBrA at 6kHz and 13kHz. The primary aliasing signal at 25kHz is hardly suppressed at all, as is expected, and sits at -5dBrA.  

Note: the filter characteristics are described under General information above. Our measured FFTs generally match the descriptions provided by iFi Audio.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (GTO filter)

The plot above shows a fast Fourier transform (FFT) of the Go Bar balanced output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sine wave at -1dBFS sampled at 44.1kHz (purple/green), using the GTO filter. The GTO filter offers a softer rolloff above 20kHz in the white-noise spectrum compared to the STD and MIN filters. There are two imaged aliasing artifacts in the audio band above the -135dBrA noise floor at around 6kHz (-100dBrA) and 13kHz (-100dBrA). The primary aliasing signal at 25kHz is only mildly suppressed at -20dBrA, with subsequent harmonics of the 25kHz peak below this level.

Note: the filter characteristics are described under General information above. Our measured FFTs generally match the descriptions provided by iFi Audio.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (MIN filter)

The plot above shows a fast Fourier transform (FFT) of the Go Bar balanced output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sine wave at -1dBFS sampled at 44.1kHz (purple/green), using the MIN filter. The sharp rolloff above 20kHz in the white-noise spectrum shows the implementation of the brick-wall-type reconstruction filter. There are two imaged aliasing artifacts in the audio band above the -135dBrA noise floor, at around 6kHz (-110dBrA) and 13kHz (-115dBrA). The primary aliasing signal at 25kHz is at -65dBrA, with subsequent harmonics of the 25kHz peak near -80 and -70dBrA. The behavior of the MIN filter is similar to that of the STD filter.

Note: the filter characteristics are described under General information above. Our measured FFTs generally match the descriptions provided by iFi Audio.

THD ratio (unweighted) vs. frequency vs. load (24/96)

The chart above shows THD ratios at the balanced output into 600 ohms (blue/red), 300 ohms (purple/green) and 32 ohms (pink/orange) for a constant 2Vrms output as a function of frequency, for 96kHz sampled dithered input data. Into 600 ohms, the right channel clearly outperforms the left by almost 10dB. Into 300 ohms, THD ratios are between 0.002 and 0.003%, the same as the right channel into 600 ohms. At 6kHz, THD ratios into 300 ohms are nearing 0.005%. Into 32 ohms, THD ratios are a bit higher, between slightly above, and below 0.005%.

THD ratio (unweighted) vs. output (600, 300, and 32 ohms) at 1kHz

The chart above shows THD ratios at the balanced output into 600 ohms (blue/red), 300 ohms (purple/green), and 32 ohms (pink/orange) at 1kHz as a function of output voltage, for 96kHz sampled dithered input data. Up to about 50mVrms, all three data sets track, with a THD ratio of about 0.005%. From 50mVrms to about 2Vrms (125mW), the 32-ohm data is fairly flat between 0.005% and 0.01%. At 200mVrms, the 600 and 300-ohm data still track closely, at just above 0.001%. From 200mVrms to about 2Vrms (13mW), the 300-ohm data is fairly flat between 0.001% and 0.002%. The 600-ohm data reaches a THD low of about 0.0006% at 400mVrms, then up to 0.002% at 2Vrms (6.7mW). The 1% THD mark is close to the same into 600 ohms (95mW) and 300 ohms (177mW), just over 7Vrms. Into 32 ohms, the 1% THD mark is just over 5.5Vrms, at 1W.

THD+N ratio (unweighted) vs. output (96kHz for 600, 300, and 32 ohms) at 1kHz

The chart above shows THD+N ratios at the balanced output into 600 ohms (blue/red), 300 ohms (purple/green), and 32 ohms (pink/orange) at 1kHz as a function of output voltage, for 96kHz sampled dithered input data. All data sets track closely up to about 200mVrms, where THD+N ratios measure 0.015%. Beyond 200mVrms, the 32-ohm data reaches a low of 0.005% at 2Vrms, while the 300 and 600-ohm data track almost perfectly, and reach a THD+N low of just over 0.002% at 2Vrms.

Intermodulation distortion vs generator level (SMPTE, 60Hz:4kHz, 4:1 for 600, 300, and 32 ohms)

The chart above shows intermodulation distortion (IMD) ratios measured at the balanced output into 600 ohms (blue/red), 300 ohms (purple/green), and 32 ohms (pink/orange) from -60dBFS to 0dBFS, for 96kHz sampled dithered input data. Here, the SMPTE IMD method was used, where the primary frequency (F1 = 60Hz) and the secondary frequency (F2 = 7kHz) are mixed at a ratio of 4:1. The SMPTE IMD analysis results consider the second (F2 ± F1) through the fifth (F2 ± 4xF1) modulation products. The 600-ohm data yielded the highest IMD ratios, from 2% down to 0.06% at -30dBFS, then up to 0.3% (left) at -18dBFS, then back down to roughly 0.02% at 0dBFS. The 300- and 32-ohm data track perfectly down to -30dBFS, where IMD ratios measured 0.02%. Beyond this threshold, the 32-ohm data flattens out and reaches 0.01% at 0dBFS. The 300-ohm data reaches an IMD low of nearly 0.002% just shy of 0dBFS.

FFT spectrum – 1kHz (44.1kHz data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sine-wave stimulus, measured at the balanced output into 300 ohms sampled at 44.1kHz. We see the third signal harmonic (3kHz) dominating at -95dBrA, or 0.002%, while the second harmonic (2kHz) is between -100 and -110dBrA (left/right), or 0.001 and 0.0003%. Subsequent odd harmonics (3, 5, 7, 9kHz) can be seen at and below -110dBrA, or 0.0003%. There are no power-supply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (96kHz data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sine-wave stimulus, measured at the balanced output into 300 ohm sampled at 96kHz. Within the audio band, we see essentially the same FFT as with the 44.1kHz sampled data above.

FFT spectrum – 1kHz (44.1kHz data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sine-wave stimulus, measured at the balanced output into 300 ohms, sampled at 44.1kHz at -90dBFS. We see the main peak at the correct amplitude, surround by an elevated noise floor at around -100dBrA.

FFT spectrum – 1kHz (96kHz data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sine-wave stimulus, measured at the balanced output into 300 ohm, sampled at 44.1kHz at -90dBFS. We see the main peak at the correct amplitude, surround by an elevated noise floor at around -100dBrA. The noise floor is identical to the 44.1kHz FFT above because as a USB DAC, bit depth could not be altered and is held at 32 bits for all sample rates. 

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 44.1kHz)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sine-wave stimulus tone measured at the balanced output into 300 ohms sampled at 44.1kHz. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 2Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz) is at -95/100dBRa (left/right), or 0.002/0.001%, while the third-order modulation products, at 17kHz and 20kHz, are also at -95dBrA.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 96kHz)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sine-wave stimulus tone measured at the balanced output into 300 ohms sampled at 44.1kHz. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 2Vrms (0dBrA) at the output. Within the audioband, we see essentially the same IMD FFT as with the 44.1kHz sampled data above.

Diego Estan
Electronics Measurement Specialist