Details

Parent Category: Products

Category: Phono Preamplifier Measurements

Created: 15 January 2021

Link: reviewed by Jason Thorpe on SoundStage! Hi-Fi on January 15, 2021

General information

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

The V10 was conditioned for 30 minutes at 1Vrms at the output before any measurements were taken.

The V10 offers one pair of unbalanced RCA outputs and one pair of balanced XLR outputs, as well as two pairs of unbalanced RCA inputs, one pair for moving-magnet (MM) and the other for a moving-coil (MC) cartridge. There are a number of DIP switches on the rear panel of the V10, allowing the user to alter MC resistive loading, MM capacitive loading, and gain, as well to turn on or off a subsonic filter. Other than the 6dB difference in gain between the balanced and unbalanced outputs (+6dB for balanced), there were no mentionable differences between both outputs types in terms of THD+N, as long as the gain was kept constant. Settings were left at the manufacturer’s default positions (see specs below). To achieve the reference output voltage of 1Vrms at 1kHz at the balanced output, 10mVrms was required at the MM input and 1mVrms at the MC input.

Published specifications vs. our primary measurements

The tables below summarize our primary measurements performed on the V10. Here we can compare directly against Hegel’s own published specifications for the V10, which are stated as follows:

• Gain (XLR): MM, 40/45/50/52dB; MC, 60/65/70/72dB (default = 40/60dB for MM/MC)
• Gain (RCA): MM, 34/39/44/46dB; MC, 54/59/64/66dB (default = 34/54dB for MM/MC)
• MC load impedance: 100/300/47k ohms (default = 100 ohms)
• MM load capacitance: 100/147/200/220/247/320/420pF at 47k ohms (default = 100pF)
• Subsonic filter: on, -3dB at 20Hz; off, -18dB/octave
• RIAA accuracy: +/-0.1dB 20Hz to 20kHz
• Output noise: -84dB (MM, A-weighted, ref 1Vrms), -81dB (MC, A-weighted, ref 1Vrms)
• Output impedance: XLR and RCA, 200 ohms
• Channel crosstalk: -84dB @ 1kHz (ref 1Vrms)
• Frequency response: 2Hz to 20kHz
• THD: MM, <0.005%; MC, <0.009% at 1kHz ref 1Vrms
• Maximum output voltage: 23Vrms

Our primary measurements revealed the following using the unbalanced MM input (unless specified, assume a 1kHz sinewave, 1Vrms at the balanced output into a 200k ohms load, 10Hz to 90kHz bandwidth):

Our primary measurements revealed the following using the unbalanced MC input (unless specified, assume a 1kHz sinewave, 1Vrms at the balanced output into a 200k ohms load, 10Hz to 90kHz bandwidth):

Gain was measured at 40 and 60dB (Balanced, MM/MC), the same as Hegel’s spec. The MC load input impedance was measured at 124 ohms, closely matching Hegel’s spec and setting of 100 ohms. The input impedance for the MM input was measured at 47 and 46k ohms (L/R), very close to the industry standard 47k ohms.

Output noise (A-weighted) was measured at -85dB (MM) and -79dB (MC), either exceeding or approaching Hegel’s specs of -84/-81dB (MM/MC). Our measured output impedance of 200 ohms confirmed Hegel’s spec of the same value.

Our measured crosstalk values at 10kHz were close to Hegel’s spec of -84dB (at 1kHz), where we measured -83dB for both channels for the MM input and -88/-95dB (left and right channels) for the MC input. For a direct comparison, we also measured crosstalk at 1kHz, and found -98/-100dB (L/R channels, MM input), and -88/-90dB (L/R channels, MC input), bettering the -84dB Hegel spec.

Hegel’s claim of better than 0.005% (MM) and 0.009% (MC) THD was also verified. There are many ways to measure THD, and manufacturers are often shy on showing all of the parameters used for the measurement. Here, we assume that Hegel is referring to THD (not THD+N), unweighted, against a 1Vrms 1kHz output signal. The bandwidth for Hegel’s measurement is unknown, but we use 10Hz to 90kHz for our measurements. Under those conditions, we found the V10’s THD to be below 0.0008% (MM) and 0.001% (MC). Even if Hegel’s THD spec is actually THD+N (A-weighted), they still meet spec for the MM input, as we measured 0.005%, and come very close for the MC input at 0.01%.

Hegel’s maximum output voltage (1% THD+N) claim of 23Vrms (balanced) was also confirmed, where we measured 24Vrms.

Frequency response - MM input

The blue (left channel) and red (right channel) traces represent the frequency response without the subsonic filter turned on. In our measured frequency-response plot above for the MM input, the V10 is within +/-0.2dB or so of flat from 20Hz to 20kHz, just about meeting Hegel’s RIAA accuracy claim of +/-0.1dB. An inverse RIAA EQ is applied to the input sweep, so that if a device were to track the RIAA curve perfectly, a flat line would emerge—the V10 has nearly perfect RIAA accuracy from 5Hz to 80kHz. The purple (left channel) and green (right channel) traces represent the frequency response with the subsonic filter engaged, where Hegel’s claim of -3dB at 20Hz is confirmed. In the graph above and some of the graphs below, we see two visible traces; the left channel (blue or purple trace) and the right channel (red or green trace). On other graphs, only one trace may be visible, which is because the left and right channels are tracking extremely closely, so as not to show a difference with the chosen axis scales.

Frequency response - MC input

In our measured frequency-response plot above for the MC input shown with the blue and red traces (left and right channels), the V10 is within +/-0.2dB or so of flat from 20Hz to 20kHz, just about meeting Hegel’s RIAA accuracy claim of +/-0.1dB. The worst-case deviation was seen at around 8Hz, were the V10 over-responded by 0.5dB. The purple and green trace (left and right channels) represent the frequency response with the sub-sonic filter engaged, where Hegel’s claim of -3dB at 20Hz is confirmed.

Phase response - MM and MC inputs

Above is the phase response of the V10 for both the MM and MC inputs (they measured effectively identically), from 20Hz to 20kHz. The V10 does not invert polarity. Since phono preamplifiers must implement the RIAA equalization curve, which ranges from +19.9dB (20Hz) to -32.6dB (90kHz), phase shift at the output is inevitable. Here we find a worst case -60 degrees around 200Hz and -55 degrees at 5kHz.

THD ratio (unweighted) vs. frequency - MM input

The chart above shows THD ratio as a function of frequency for the MM input, where the input sweep is EQ’d with an inverted RIAA curve. The output voltage is maintained at the refrence 1Vrms. The THD values vary from 0.01% at 20Hz, down to below 0.0001% at 10kHz, then back up just above 0.0003% at 20kHz.

THD ratio (unweighted) vs. frequency - MC input

The chart above shows THD ratio as a function of frequency for the MM input, where the input sweep is EQ’d with an inverted RIAA curve. The THD values vary from 0.04% at 20Hz, down to around 0.0004% at 5kHz, then a climb to 0.001% at 20kHz.

THD ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 30mVrms) - MM input

Above we can see a plot of THD ratios as a function of output voltage for the MM input. We can see very low THD ratio values, ranging from as low as 0.0001% at 5Vrms, up to about 0.0006% at the “knee”, just past 20Vrms, and about 0.005% at the lowest output voltage of 100mVrms. Beyond the “knee” there is sharp rise in THD as the V10 reaches its maximum output voltage. It’s important to mention that anything above 1-2Vrms is not typically required for most line-level preamps or integrated amps.

THD+N ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms) - MM input

Above we can see a plot of THD+N ratios as a function of output voltage for the MM input. We can see THD+N ratio values, ranging from 0.2% at 100Vrms, down to about 0.0015% between 15 and 20Vrms.

THD ratio (unweighted) vs. output voltage at 1kHz (input voltage from 0.1mVrms to 30mVrms) - MC input

Above we can see a plot of THD ratios as a function of output voltage for the MC input. The MC input behaved similarly to the MM input, with 0.007% THD at 100mVrms, dipping to the lowest THD value of 0.0005% at 2Vrms, then up just past 0.003% at the knee at 20Vrms. The 1% THD ratio value is reached at 24Vrms at the output.

THD+N ratio (unweighted) vs. output voltage at 1kHz (input voltage from 0.1mVrms to 30mVrms) - MC input

Above we can see a plot of THD+N ratios as a function of output voltage for the MC input. We can see THD+N ratio values ranging from around 0.5% at 100Vrms, down to about 0.005% between 15 and 20Vrms.

FFT spectrum, 1kHz - MM input

Shown above is a fast Fourier transform (FFT) of a 1kHz input sinewave stimulus for the MM input, which results in the reference output voltage of 1Vrms. Here we see exceptionally clean results. Signal harmonics are non-existent (i.e., cannot be seen above the -120dB noise floor at 2kHz). The frequency components on the left side of the 1kHz peak are mostly due to power-supply noise, where the typical 60Hz and 120Hz peaks can be seen. The worst noise peak (60Hz) is below -80dBrA, with the subsequent harmonic (120Hz) at around -90dBrA. The third harmonic (180Hz) is just below -90dBrA, and the subsequent harmonics are below -100dBrA.

FFT spectrum, 1kHz - MC input

Shown above is the FFT of a 1kHz input sinewave stimulus for the MC input. Signal harmonics are virtually non-existent, only a hint of a peak can be seen below -110dB at 2kHz. The 60Hz and 120Hz peaks are just below -70dBrA, and the third and fourth noise harmonics are just below -80dBrA.

FFT spectrum, 50Hz - MM input

Shown above is the FFT for a 50Hz input sinewave stimulus measured at the output for the MM input. The X axis is zoomed in from 40Hz to 1KHz so that peaks from noise artifacts can be directly compared against peaks from the harmonics of the signal. The harmonics from the 50Hz signal (100, 150, 200Hz, etc) are non-existent. The power supply noise peaks can clearly be seen, which were described in the 1kHz FFT chart above.

FFT spectrum, 50Hz - MC input

Shown above is the FFT for a 50Hz input sinewave stimulus measured at the output for the MC input. Like the MM FFT, here again, the harmonics from the 50Hz signal (100, 150, 200Hz, etc.) are non-existent. The power-supply noise peaks can clearly be seen, which were described in the 1kHz FFT above.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus) - MM input

The chart above is an FFT of the IMD products for an 18kHz and 19kHz summed sinewave stimulus tone for the MM input. The input rms values are set so that if summed (for a mean frequency of 18.5kHz), would yield 1Vrms (reference or 0dBRa) at the output. Here we find the second-order modulation product (i.e., the difference signal of 1kHz) just above -100dBrA. We can also see the third-order modulation products (i.e., 17kHz and 20kHz) are extremely low, below -120dBRa.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus) - MC input

The next chart is an FFT of the IMD products for an 18kHz and 19kHz summed sinewave stimulus tone for the MC input. Here we find the second-order modulation product (i.e., the difference signal of 1kHz) just at -80dBrA, which is very low. We can also see the third-order modulation products (i.e., 17kHz and 20kHz) are extremely low, below -120dBRa. These clean IMD FFTs are reflected in our simplified IMD results (which only account for the sum of the second- and third-order modulation products) in our primary measurement table, where the MM input measured at -92dB for both channels and the MC input at -76/-73dB, for, respectively, the left and right channels. Incidentally, the 3dB difference between left and right channels for the MC input can be seen in the 1kHz peak in the FFT chart above. The fourth-order modulation products are non-existent in this FFT and the one for the MM input.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Phono Preamplifier Measurements

Created: 01 January 2021

Link: reviewed by Doug Schneider on SoundStage! Hi-Fi on January 1, 2021

General information

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

The Bellari VP549 was conditioned for 30 minutes at 1Vrms at the output before any measurements were taken.

The VP549 is designed for moving-magnet (MM) cartridges only and has one set of single-ended RCA inputs and outputs. There are two switches and one knob on the front panel of the VP549. One switch selects between 120pF, 220pF, and 330pF cartridge-loading capacitance. The other switch, Rumble Filter, enables a low-frequency filter that provides attenuation below 20Hz. The knob controls the preamp’s overall gain. The center position is labelled “0,” the minimum position “-10dB,” and the maximum position “+4dB.” Unless otherwise stated, measurements were performed with the cartridge load capacitance set to 220pF, the Rumble Filter switch disengaged, and the gain set to the default “0” position. An 8mVrms 1kHz sinewave was required to achieve the reference output voltage of 1Vrms.

Published specifications vs. our primary measurements

The table below summarizes our primary measurements performed on the VP549. Here we can compare directly against Bellari’s own published specifications for the VP549, which are stated as follows:

• Gain: 41.51dB gain @ 1kHz at the "0" setting, 31.64dB of gain at the -10 setting, 45.66dB of gain at the +4 setting
• Input Impedance: 47k ohms
• Output Impedance: 470 ohms
• THD: 0.005% @ 1kHz
• S/N Ratio: >94dB unweighted
• Equalization: RIAA +/-1dB, 14Hz to 23kHz

Our primary measurements revealed the following (unless specified, assume a 1kHz sinewave, 1Vrms output into 100k ohms, 10Hz to 90kHz bandwidth, gain set to 0):

Our measured input (47.8/46.9k ohms, L/R) and output (460.5/460.8 ohms, L/R) impedances match very closely Bellari’s specs of 47k ohms and 470 ohms. Our measured gain values at all three settings also corroborate the published specifications, with maximum deviations in the order of 2-3%.

For the remainder of the specifications, as is often the case with manufacturer supplied measurements, not enough information is given to directly reproduce their results. With respect to Bellari’s THD of 0.005%, the company has not specified a reference output voltage, gain setting, weighting, or bandwidth filter. Our THD measurement shows an even lower <0.0012% (unweighted) against a 1Vrms output voltage, whereas our THD+N figure (A-weighted) is <0.0063% for both channels.

Bellari’s signal-to-noise ratio (SNR) of >94dB (unweighted) seems generous, however, again, we have very little information provided with respect to measurement parameters. The most charitable SNR measurement we could achieve was with the gain setting at -10dB and the output voltage at 1.35 Vrms (1 kHz), where we measured 95dB (A-weighted). Our reference SNR measurement (1Vrms output, gain set to 0) is 83dB (A-weighted), or 60dB (unweighted, input bandwidth filter set from 20Hz to 20kHz).

Frequency response

The chart above shows our frequency response measurement. An inverse RIAA EQ is applied to the input sweep, so that if a device were to track the RIAA curve perfectly, a flat line would emerge. The blue and red curves are the left and right channel without the Rumble Filter switch engaged, the green and purple curves with the Rumble Filter switch engaged. The measurement was also made with the gain set to its minimum and maximum values, and with every cartridge load capacitance setting, which had no effect on the frequency-response results. We can see here that Bellari’s claim of +/-1dB is valid at the top end of the spectrum (up to 23kHz as stated), but the +/-1dB claim down to 14Hz was not corroborated by our measurement. We find a -5.3dB dip at 30Hz, and at 14Hz, down -5.5dB again. In the graph above and some of the graphs below, we see two visible traces; the left channel (blue or purple trace) and the right channel (red or green trace). On other graphs, only one trace may be visible, this is because the left and right channels are tracking extremely closely, so as not to show a difference with the chosen axis scales.

Phase response

Shown above is the phase response from 20Hz to 20kHz. The VP549 does not invert polarity. Since phono preamplifiers must implement the RIAA equalization curve, which ranges from +19.9dB (20Hz) to -32.6dB (90kHz), phase shift at the output is inevitable. Here we find a worst case +20 degrees at 20Hz, and just below -60 degrees between 5 and 10kHz.

THD ratio (unweighted) vs. frequency

The chart above is the THD ratio as a function of frequency, where the input sweep is EQ’d with an inverted RIAA curve. The output voltage is maintained at the refrence 1Vrms. The THD values vary from 0.05% at 20Hz, down to below 0.001% at 1kHz, then back up to 0.007% at 20kHz.

THD ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms)

Above we can see a plot of THD ratio as a function of output voltage. We can see very low THD ratio values, ranging from 0.005% at 100mVrms down to below 0.001% at around 1Vrms at the output, then a sharp rise in THD at the “knee” at around 3.5Vrms at the output. Beyond this output voltage, the VP549 distortion begins to rise exponentially. The 1% THD ratio value is reached at around 4Vrms at the output. It’s important to mention that anything above 1-2Vrms is not typically required for most line-level preamps or integrated amps.

THD+N ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms)

Above we can see a plot of THD+N ratio as a function of output voltage. At 100mVrms, THD+N values are at around 3%, then down to between 0.1 and 0.2% at 3Vrms. There is a considerable difference in the THD+N vs. THD plots for the VP549, where it’s clear that distortion products are low, but noise, as shown in this chart, is relatively high.

FFT spectrum, 1kHz

Shown above is a fast Fourier transform (FFT) of a 1kHz 8mVrms input sinewave stimulus, which results in the reference output voltage of 1Vrms with the gain set to 0. Here we see that the second harmonic at 2kHz is <-100dB below the reference signal (dBrA). The frequency components on the left side of the 1kHz peak are mostly due to power supply noise. Of note is the peak at 60Hz at a level near -70dBrA.

FFT spectrum, 50Hz

The chart above is an FFT of a 50Hz input sinewave stimulus, which results in the reference output voltage of 1Vrms with the gain set to 0. The X axis is zoomed in from 40Hz to 1kHz, so that peaks from noise artifacts can be directly compared against peaks from the harmonics of the signal. We find no perceptible signal harmonic peaks (e.g., 100Hz, 150Hz) above the relatively high noise floor. The highest noise peak is from the primary source (60Hz) at about -75dBrA, while the next highest peak is from the third harmonic noise peak (180Hz) at about -85dBrA.

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

This final chart shows an FFT of the IMD products for an 18kHz and 19kHz summed sinewave stimulus tone. The input rms values are set so that if summed (for a mean frequency of 18.5kHz) would yield 1Vrms (Reference or 0dBRa) at the output. Although it appears that the second-order modulation product (i.e., the difference signal of 1kHz) is sitting at -105dBrA, the peak is actually at 1020Hz, which is the 17th harmonic from the 60Hz power-supply noise. The 1kHz IMD peak is imperceptible above the noise floor. The third-order modulation products (i.e., 17kHz and 20kHz) are just above -110dBra, and the subsequent modulation products are not noticeable.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Phono Preamplifier Measurements

Created: 01 January 2021

Links: reviewed by Doug Schneider on SoundStage! Hi-Fi on January 1, 2021 and by James Hale on SoundStage! Xperience on January 1, 2021

General information

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

The NAD PP 2e was conditioned for 30 minutes at 1Vrms at the output before any measurements were taken.

On the PP 2e’s back panel are one switch, to switch between moving-magnet (MC) and moving-coil (MC) operation, and two sets of single-ended RCA inputs, with one set labeled MC and the other set MM. The rear panel also has a pair of single-ended RCA outputs. For the MM input, a 15mVrms 1kHz sinewave was required to achieve the reference output voltage of 1Vrms, while a 1.2mVrms 1kHz sinewave was required at the MC input.

Published specifications vs. our primary measurements

The tables below summarize our primary measurements performed on the PP 2e. Here we can compare directly against NAD’s own published specifications for the PP 2e, which are stated as follows:

• Input impedance: MM, 47k ohms; MC, 100 ohms
• Gain at 1kHz: MM, 35dB; MC, 60dB
• Input sensitivity (ref. 200mV output): MM, 2.5mV; MC, 0.3mV
• Signal-to-noise ratio (A-weighted): MM, 80dB; MC, 78dB
• Input overload (20Hz/1kHz/20kHz): MM, 10/102/950mV; MC, 0.8/9/84 mV
• Rated distortion (THD 20Hz to 20kHz): MM, <0.03%; MC, <0.03%
• RIAA response accuracy: MM, +/-0.3dB; MC, +/-0.3dB
• Output impedance: 100 ohms
• Maximum output level: 5.3V

Our primary measurements below revealed the following using the unbalanced MM input (unless specified, assume a 1kHz sinewave, 1Vrms output into 100k ohms load, 10Hz to 90kHz bandwidth):

Our primary measurements below revealed the following using the unbalanced MC input (unless specified, assume a 1kHz sinewave, 1Vrms output into 100k ohms load, 10Hz to 90kHz bandwidth):

Our measured input impedances for both the MM (47.65/47.81k ohms) and MC inputs (117.6/118.6 ohms) match very closely NAD’s specs of 47k ohms and 100 ohms.

Our measured gain values for both inputs (MM: 37dB, MC: 59dB) also corroborate the published specifications of 35dB and 60dB. Input sensitivity is another way to express gain (in volts per volts). Interestingly, NAD’s input sensitivity specs when converted to dB are 38dB for the MM input and 56.5dB for the MC input, which are slightly different than their specs for gain. Nevertheless, all published gain values are very close to our measured values.

NAD’s signal-to-noise ratio (SNR) specs of 80dB and 78dB (MM and MC, A-weighted) are difficult to compare to our measured values, because we do not know what NAD used as a reference output voltage. Nevertheless, our measured values of 93dB and 84dB (MM and MC, A weighted, with a 1Vrms reference) exceed those published by NAD.

NAD’s maximum output voltage value of 5.3mVrms was corroborated by our measurements, where we saw 1% THD+N at 5.32Vrms at the output (1kHz).

Our input overload values are expressed in dB, as overload margin, with a reference 5mVrms and 0.5mVrms input signal for the MM and MC inputs respectively. We measured the input signals required at 20Hz, 1kHz, and 20kHz to achieve 5.32Vrms at the output for both input types. We expressed those values in dB as a ratio over the reference values (5 and 0.5Vrms), but we also show the actual input voltages in parentheses so as to directly compare against the NAD values. We measured lower input overload values; with some rounding, our values compare to NADs as follows: 8/75/711 vs. 10/102/950mVrms for the MM input, and 0.7/6/57 vs. 0.9/8/84Vrms for the MC input.

NAD’s published rated distortion values are <0.03% for both input types; unfortunately, we only know these were measured with a 20Hz to 20kHz filter. We don’t know if NAD measured THD or THD+N, and what they used as a reference output voltage. Our measured THD+N values of about 0.003% and 0.006% (MM and MC) show better performance than the published values, but these are A-weighted. We also measured THD+N ratios with a 20Hz to 20kHz bandwidth (instead of A-weighting), using the fourth-order low-/high-pass Butterworth filters in the APx555, and measured less than 0.008% for the MM input, and less than 0.03% for the MC input, corroborating NAD’s claim.

In terms of our measured output impedance, there is a significant discrepancy, where we measured about 522 ohms vs. NAD’s published 100 ohms. While 500 ohms is on the high side for a line level device output impedance, it should pose no issues for any typical active preamp or integrated amp input.

Frequency response - MM and MC inputs

In the measured frequency-response plot (above), the PP 2e is within 0.3dB of flat from 20Hz to 20kHz, corroborating the NAD claim. We found the same measured response with both the MM and MC input. An inverse RIAA EQ is applied to the input sweep, so that if a device were to track the RIAA curve perfectly, a flat line would emerge. In the graph above and some of the graphs below, we see two visible traces; the left channel (blue or purple trace) and the right channel (red or green trace). On other graphs, only one trace may be visible, this is because the left and right channels are tracking extremely closely, so as not to show a difference with the chosen axis scales.

Phase response - MM and MC inputs

Above is the phase response of the PP 2e for both the MM and MC inputs (they measured effectively identically), from 20Hz to 20kHz. The PP 2e does not invert polarity. Since phono preamplifiers must implement the RIAA equalization curve, which ranges from +19.9dB (20Hz) to -32.6dB (90kHz), phase shift at the output is inevitable. Here we find a worst case +28 degrees at 20Hz, and just above -60 degrees between 200 and 300Hz, and right around -60 degrees between 5 and 8kHz.

THD ratio (unweighted) vs. frequency - MM input

This is the THD ratio as a function of frequency plot for the MM input, where the input sweep is EQ’d with an inverted RIAA curve. The output voltage is maintained at the refrence 1Vrms. The THD values vary from 0.005% at 20Hz, down to below 0.001% from 300Hz to 500Hz, then back up just above 0.01% at 20kHz. One curiosity is the deviation in THD between left and right channels, where we find up to 5dB difference in favor of the right channel, which is most obvious between 1 and 5kHz.

THD ratio (unweighted) vs. frequency - MC input

Above is the THD ratio as a function of frequency for the MC input. The THD values vary from 0.01% at 20Hz, down to just above 0.001% at around 500Hz, then back up just above 0.01% at 20kHz. Here again, we see a deviation in THD between left and right channels.

THD ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms) - MM input

Above is the THD ratio as a function of output voltage plot for the MM input. We can see very low THD ratio values ranging from just above 0.01% down to 0.001% at around 2Vrms at the output, then to a sharp rise in THD at the “knee” at just shy of 5Vrms at the output. The deviation in THD performance (5-10dB) in favor of the right channel can be seen with output voltages between 1 and 5Vrms. The 1% THD ratio value is reached at 5.32Vrms at the output. It’s important to mention that anything above 1-2Vrms is not typically required for most line-level preamps or integrated amps.

THD+N ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms) - MM input

Above is the THD+N ratio as a function of output voltage plot for the MM input. We can see THD+N ratio values ranging from just above 0.1% below 100mVrms, down to 0.002% (R) and 0.005% (L) at around 3-4Vrms at the output.

THD ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms) - MC input

Above is the THD ratio as a function of output voltage plot for the MC input. THD ratio values range from about 0.015% at 100mVrms to 0.001% at around 2Vrms at the output, then to a sharp rise in THD at the “knee” at just shy of 5Vrms at the output. Beyond this output voltage, the DUT distortion begins to rise exponentially. The 1% THD ratio value is reached at 5.32Vrms at the output.

THD+N ratio (unweighted) vs. output voltage at 1kHz (input voltage from 1mVrms to 100mVrms) - MC input

The chart above is the THD+N ratio as a function of output voltage plot for the MC input. We can see THD+N ratio values ranging from just above 0.2% below 100mVrms, down to between 0.005% and 0.01% at around 3-4Vrms at the output.

FFT spectrum, 1kHz - MM input

Shown above is a fast Fourier transform (FFT) of a 1kHz 15mVrms input sinewave stimulus for the MM input, which results in the reference output voltage of 1Vrms. Here we see that the second signal harmonic at 2kHz is at -100dBrA (L) and -105dBRa (R). At 3kHz, or the third signal harmonic, we see at 10dB difference between the left (-100dBrA) and right (-110dBrA) channels. The worst noise peak is at 120Hz (second harmonic of 60Hz) and can be seen at around -90dBrA (left) and -100dBrA (right).

FFT spectrum, 1kHz - MC input

Above is a fast Fourier transform (FFT) of a 1kHz 1.2mVrms input sinewave stimulus for the MC input. The second (2kHz) and third (3kHz) signal harmonics are roughly the same here as with the MM input above. The worst noise peaks are at around -90dBrA at 180Hz (third harmonic of 60Hz) and just below -80dBrA at 60Hz. Of note, is that the 120Hz predominates on the MM input, while on the MC input, it’s the 60Hz peak.

FFT spectrum, 50Hz - MM input

The chart above depicts a fast Fourier transform (FFT) of a 50Hz input sinewave stimulus for the MM input. The X axis is zoomed in from 40Hz to 1kHz, so that peaks from noise artifacts can be directly compared against peaks from the harmonics of the signal. Here we see that the second noise harmonic at 120Hz dominates at -90dBrA (L) and -100dBRa (R). The second signal harmonic (100Hz) is at -100dBrA.

FFT spectrum, 50Hz - MC input

Above is a fast Fourier transform (FFT) of a 50Hz input sinewave stimulus for the MC input. Here we see that the thrid noise harmonic at 180Hz dominates at -80dBrA (L) and -90dBRa (R). The second signal harmonic (100Hz) is barely perceptible above the noise floor at -95dBrA.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus) - MM input

The chart above represents an FFT of the IMD products for an 18kHz and 19kHz summed sinewave stimulus tone for the MM input. The input rms values are set so that if summed (for a mean frequency of 18.5kHz), would yield 1Vrms (reference or 0dBRa) at the output. Here we find the second order modulation product (i.e., the difference signal of 1kHz) at -95dBrA, which puts them a little below the primary power supply noise products. The third-order modulation products (i.e., 17kHz and 20kHz) are considerably different between the right and the left channels. This is also reflected in our simplified IMD results (which only accounts for the sum of the second and third order modulation products) in our primary measurement table, where the right channel outperformed the left by 15dB on the MM input, and by 5dB on the MC input. The worst-case third-order modulation product peaks are at about -85dBrA for the left channel, but closer to -105dBrA for the right channel.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus) - MC input

Above is an FFT of the IMD products for an 18kHz and 19kHz summed sinewave stimulus tone for the MC input. Here we find the second-order modulation product at -85dBrA. The third-order modulation products (i.e., 17kHz and 20kHz) are at about -85dBrA for the left channel, but closer to -105dBrA for the right channel.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Digital-to-Analog Converter Measurements

Created: 15 October 2024

Link: reviewed by Matt Bonaccio on SoundStage! Hi-Fi on October 15, 2024

General Information

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

The Orchard Audio PecanPi+ Streamer Premium was evaluated as a DAC and conditioned for 30 min at 0dBFS (volume set to 2Vrms out) into 200k ohms before any measurements were taken.

The Orchard Audio PecanPi+ Premium is marketed as a network streamer, but does offer one coaxial S/PDIF (RCA) digital input. There are two line-level outputs (balanced XLR and unbalanced RCA) and one headphone output (1/4″ TRS). There is a digital volume control for the headphone and line-level outputs via a potentiometer on the front panel. Comparisons were made between unbalanced and balanced line-level outputs and no appreciable differences were seen in terms of THD and noise, but 1kHz FFTs are provided for both balanced and unbalanced outputs.

The analyzer’s input bandwidth filter was set to 10Hz-22.4kHz for all measurements, except for frequency-response (DC to 1 MHz), FFT (10Hz to 90kHz), and THD vs frequency (10Hz to 90kHz) charts, the latter to capture the second and third harmonics of the 20kHz output signal.

The PecanPi+ digital volume control offers no visual on-screen feedback to the user in terms of level. Volume can be adjusted in 0.5dB steps. Channel-to-channel deviation was good, at around 0.013dB throughout the range.

Volume-control accuracy (measured at line-level outputs): left-right channel tracking

Primary measurements

Our primary measurements revealed the following using the coaxial input and the balanced line-level outputs (unless specified, assume a 1kHz sinewave at 0dBFS, 200k ohms loading, 10Hz to 22.4kHz bandwidth):

*due to very low noise of DUT, analyzer self-noise has been removed from measurement to more accurately report value

Our primary measurements revealed the following using the coaxial input and the headphone output (unless specified, assume a 1kHz sinewave at 0dBFS, 300 ohms loading, 10Hz to 22.4kHz bandwidth):

Frequency response vs. sample rate (16/44.1, 24/96, 24/192)

The plot above shows the PecanPi+’s frequency response as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 24/96 dithered digital input signal from 5Hz to 48kHz, and finally orange/pink represents 24/192 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 soft filtering below 20, 30, and 50kHz (less than half the respective sample rate), The -3dB point for each sample rate is roughly 16, 35, and 70kHz, respectively. With 16/44.1 data, the -1dB point is at 13.1kHz. The PecanPi+ appears to utilize a reconstruction filter that prioritizes a clean impulse response (no pre-/post-ringing behaviour) with virtually no phase shift at the expense of more high-frequency attenuation in the frequency domain. Evidence for this can be seen in other graphs in this report. 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.

Phase response vs. sample rate (16/44.1, 24/96, 24/192)

Above are the phase response plots from 20Hz to 20kHz for a 0dBFS input signal as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital, the purple/green traces are for a 24/96 dithered digital input signal, and finally orange/pink represents 24/192 from 5Hz to 96kHz.  There is essentially no phase shift in the audioband, even for the 16/44.1 data.

Digital linearity (16/44.1 and 24/96 data)

The graph above shows the results of a linearity test for the coaxial digital input for both 16/44.1 (blue/red) and 24/96 (purple/green) input data, measured at the balanced line-level output of the PecanPi+. The digital input was swept with a dithered 1kHz input signal from -120dBFS to 0dBFS, and the output was 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 24/96 input data is essentially perfect down to -120dBFS, while the 16/44.1 input data performed well, only over-responding by 3/2dB (left/right) at -120dBFS. The 24/96 data yielded such superb results that we extended the sweep down to . . .

. . . -140dBFS. Above we see that even at -140dBFS, the PecanPi+ is only over/undershooting by 1 dB between -140 and -130dBFS. This is an excellent linearity-test result.

Impulse response

The graph above shows the impulse responses for a looped 24/44.1 test file that moves from digital silence to full 0dBFS (all “1”s) for one sample period then back to digital silence, fed to the coaxial digital input, measured at the balanced outputs, into a 200k ohm-load for the left channel only. We can see that PecanPi+ yields an impulse response with essentially no pre- or post-ringing behaviour, or one that emulates a non-oversampling DAC.

J-Test (coaxial input)

The plot above shows the results of the “J-test” test for the coaxial digital input measured at the balanced line level output of the PecanPi+. The “J-test” was developed by Julian Dunn the 1990s. It is a test signal, specifically a -3dBFS undithered 12kHz squarewave sampled (in this case) at 48kHz (24bit). Since even the first odd harmonic (i.e. 36kHz) of the 12kHz squarewave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sinewave (the main peak). In addition, an undithered 250Hz squarewave 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 sinewave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The coaxial input shows a strong J-Test result, with peaks visible, but only below the -140dBrA level.

J-Test (coaxial input, 2kHz sinewave jitter at 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the balanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results show no visible sidebands. This is further evidence of the PecanPi+s strong jitter immunity.

Wideband FFT spectrum of white noise and 19.1kHz sinewave tone

The plot above shows a fast Fourier transform (FFT) of the PecanPi+’s balanced-line level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green). There is a soft roll-off above 20kHz in the white-noise spectrum. There are absolutely no imaged aliasing artifacts in the audioband above the -135dBrA noise floor. The primary aliasing signal at 25kHz is barely suppressed at -20dBrA.

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

The chart above shows THD ratios at the balanced line-level output into 200k ohms (blue/red) and 600 ohms (purple/green) as a function of frequency for a 24/96 dithered 1kHz signal at the coaxial input. The 200k and 600 ohms data are identical throughout the audioband, which is an indication that the PecanPi+’s outputs are robust and can handle loads below 1k ohms with no difficulty. There was an evident difference in THD ratios between the left and right channels, with the right channel outperforming the left by about 10-15dB from 20Hz to 1kHz. THD ratios (right channel) ranged from 0.00003-0.00005% from 20Hz to 2kHz, then up to 0.0005% at 20kHz. Despite the left/right THD ratio discrepancy, these values are extremely low and nearing the limits of what the APx555 can measure.

THD ratio (unweighted) vs. frequency vs. sample rate (16/44.1, 24/96)

The chart above shows THD ratios at the balanced line-level output into 200k ohms as a function of frequency for a 16/44.1 (blue/red) and a 24/96 (purple/green) dithered 1kHz signal at the coaxial input. The 24/96 data (right channel) consistently outperformed the 16/44.1 data by 10-15dB from 20Hz to about 1kHz due to the inherently higher noise floor at 16 bits (i.e., the analyzer cannot assign a THD value below the noise floor). THD ratios with 16/44.1 data ranged from 0.0001 to 0.0005% throughput the audioband. 24/96 THD ratios (left channel) were essentially the same as the 16/44.1 data.

THD ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD ratios measured at the balanced output as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green), with the volume set to maximum. Once again, the 24/96 outperformed the 16/44.1 data, with a THD range from 0.2% at 300uVrms to 0.00005% at 2Vrms, while the 16/44.1 ranged from 3% down to 0.0005% at the maximum output voltage of 5.19Vrms.

THD+N ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD+N ratios measured at the balanced output as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green), with the volume set to maximum. The 24/96 outperformed the 16/44.1 data, with a THD+N range from 2% down to 0.0002% at 3-5Vrms, while the 16/44.1 ranged from 30% down to 0.002% at the maximum output voltage of 5.19Vrms.

Intermodulation distortion vs. generator level (SMPTE, 60Hz:4kHz, 4:1; 16/44.1, 24/96)

The chart above shows intermodulation distortion (IMD) ratios measured at balanced output for 16/44.1 (blue/red) input data and 24/96 input data (purple/green), from -60dBFS to 0dBFS. 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 16/44.1 data yields IMD ratios from 2% down to 0.002% at 0dBFS. The 24/96 data yields IMD ratios from 0.1% down to 0.0005% from -10 to 0dBFS.

FFT spectrum – 1kHz (digital input, 16/44.1 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1. We see signal harmonics at 2 and 3kHz. The second (2kHz) harmonic for the left channel (right channel cannot be seen above the noise floor) is at -120dBrA, or 0.0001%, and the third harmonic (3kHz) is at -130dBRa, or 0.00003%. There are also no power-supply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/96. Due to the increased bit-depth, the noise floor is much lower compared to the 16/44.1 FFT at a very low -160dBrA. We see signal harmonics ranging from -120dBrA to -150dBrA, or 0.0001% to 0.000003%, all the way to 20kHz (and beyond). The second (2kHz) signal harmonic shows the significant discrepancy between the left (-120dBrA) and right (-150dBrA) channels. Here also, there are no powersupply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS, unbalanced output)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohms for the coaxial digital input, sampled at 24/96. The main difference with the balanced output is at the second (2kHz) signal harmonic for the right channel, where here we find a level of -125dBrA, or 0.00006%, instead of the -150dBrA for the balanced outputs. The third (3kHz) harmonic is also higher here, at -120dBrA, or 0.0001%, instead of -130dBrA for the balanced outputs.

FFT spectrum – 1kHz (digital input, 16/44.1 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1 at -90dBFS. We see the signal peak at the correct amplitude, and essentially no signal harmonics or noise peaks.

FFT spectrum – 1kHz (digital input, 24/96 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at -90dBFS. We see the signal peak at the correct amplitude, and essentially no signal harmonics or noise peaks.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 16/44.1)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 16/44.1. 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 perhaps barely visible above the noise floor at -130dBrA, or 0.00003%, and the third-order modulation products, at 17kHz and 20kHz, are at -135dBrA, or 0.00002% (left channel).

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 24/96. 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 -130dBrA, or 0.00003%, and the third-order modulation products, at 17kHz and 20kHz, are at the same level.

Intermodulation distortion FFT (coaxial input, APx 32 tone, 24/192)

Shown above is the FFT of the balanced line-level output of the PecanPi+ with the APx 32-tone signal applied to the input. The combined amplitude of the 32 tones is the 0dBFS reference, and it corresponds to 2Vrms into 200k ohms. The intermodulation products—i.e., the “grass” between the test tones—are distortion products from the amplifier and below the -140dBrA, or 0.00001%, level. This is a very clean IMD result.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Digital-to-Analog Converter Measurements

Created: 01 June 2024

Link: reviewed by George de Sa on SoundStage! Hi-Fi on June 1, 2025

General Information

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

The Hegel Music Systems D50 was conditioned for 30 minutes at 0dBFS (volume set to 2.64Vrms out) into 200k ohms before any measurements were taken.

The D50 offers two coaxial S/PDIF digital inputs (RCA and BNC), two optical S/PDIF inputs (TosLink), one AES-EBU balanced digital input (XLR), and one USB input. There are two sets of line-level outputs (balanced XLR and unbalanced RCA). Comparisons were made between unbalanced and balanced line-level outputs, but no appreciable differences were seen in terms of THD and noise; however, 1kHz FFTs are provided for both balanced and unbalanced outputs.

The analyzer’s input bandwidth filter was set to 10Hz to 22.4kHz for all measurements, except for frequency response (DC to 1 MHz), FFTs (10Hz to 90kHz), and THD vs. Frequency (10Hz to 90kHz). The latter to capture the second and third harmonic of the 20kHz output signal.

Published specifications vs. our primary measurements

The table below summarizes the measurements published by Hegel for the D50 compared directly against our own. The published specifications are sourced from Hegel’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), unless otherwise stated, assume a 1kHz sinewave at 0dBFS (24/96) at the input, 2.64Vrms at the balanced output into 200k ohms, and a measurement input bandwidth of 10Hz to 22.4kHz, and the worst-case measured result between the left and right channels:

Our primary measurements revealed the following using the coaxial input and the balanced line-level outputs (unless specified, assume a 1kHz sinewave at 0dBFS, 200k ohms loading, 10Hz to 22.4kHz bandwidth):

* due to very low noise of DUT, analyzer self-noise has been removed from measurement to more accurately report value

Frequency response vs. sample rate (16/44.1, 24/96, 24/192)

The plot above shows the D50’s frequency response as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 24/96 dithered digital input signal from 5Hz to 48kHz, and finally orange/pink represents 24/192 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 brickwall-type filtering right at half the respective sample rates. The -3dB points for each sample rate are: 21, 46, and 92kHz respectively. 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.

Phase response vs. sample rate (16/44.1, 24/96, 24/192)

Above are the phase-response plots from 20Hz to 20kHz for a 0dBFS input signal as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital, the purple/green traces are for a 24/96 dithered digital input signal, and finally orange/pink represents 24/192 from 5Hz to 96kHz.  The 16/44.1 stream shows -180 degrees of phase shift at 20kHz, the 24/96 data roughly -10 degrees at 20kHz, and the 24/192 shows no phase shift within the audioband.

Digital linearity (16/44.1 and 24/96 data)

The graph above shows the results of a linearity test for the coaxial digital input (the optical input performed identically) for both 16/44.1 (blue/red) and 24/96 (purple/green) input data, measured at the balanced line-level output of the D50. For this test, the digital input was swept with a dithered 1kHz input signal from -120dBFS to 0dBFS and the output was 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 24/96 input data is essentially perfect down to -120dBFS, while the 16/44.1 input data performed well, only over-responding by 2.5/1dB (left/right) at -120dBFS. The 24/96 data yielded such superb results that we extended the sweep down to . . .

. . . -140dBFS. Above we see that even at -140dBFS, the D50 is only over/undershooting by less than 1 dB between -140 and -130dBFS. This is an exemplary linearity test result.

Impulse response

The graph above shows the impulse responses for a looped 24/44.1 test file that moves from digital silence to full 0dBFS (all “1”s) for one sample period then back to digital silence, fed to the coaxial digital input, measured at the balanced outputs, into a 200k ohm-load for the left channel only. We can see that d50 yields an impulse response with essentially no pre-ringing but sustained post-ringing.

Wideband FFT spectrum of white noise and 19.1kHz sinewave tone

The plot above shows a fast Fourier transform (FFT) of the D50’s balanced-line level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green). The steep roll-off above 20kHz in the white-noise spectrum shows the implementation of a brickwall-type reconstruction filter. There are absolutely no imaged aliasing artifacts in the audioband above the -135dBrA noise floor. The primary aliasing signal at 25kHz is highly suppressed at -100dBrA.

J-Test (coaxial RCA input)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the balanced line level output of theD50. J-Test was developed by Julian Dunn the 1990s. It is a test signal—specifically, a -3dBFS undithered 12kHz squarewave sampled (in this case) at 48kHz (24 bits). Since even the first odd harmonic (i.e., 36kHz) of the 12kHz squarewave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sinewave (the main peak). In addition, an undithered 250Hz squarewave 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 sinewave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The coaxial input shows a strong J-Test result, with peaks visible, but only below the -140dBrA level.

J-Test (optical input)

The optical input shows essentially the same J-Test result as with the coaxial RCA input.

J-Test (coaxial BNC input)

The coaxial BNC input shows essentially the same J-Test result as with the coaxial RCA input.

J-Test (AES-EBU input)

The AES-EBU input shows essentially the same J-Test result as with the rest of the inputs.

J-Test (coaxial RCA input, 2kHz sinewave jitter at 10ns)

The plot above shows the results of the J-Test test for the coaxial digital input (the optical input behaved the same) measured at the balanced line-level output, but with an additional 10ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results do show visible sidebands but only below the very low -140dBrA level. This is further evidence of the D50’s strong jitter immunity.

J-Test (coaxial RCA input, 2kHz sinewave jitter at 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input (the optical input behaved the same) measured at the balanced line-level output, but with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results do show visible sidebands but only below the low -125dBrA level. This is further evidence of the D50’s strong jitter immunity.

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

The chart above shows THD ratios at the balanced line-level output into 200k ohms (blue/red) and 600 ohms (purple/green) as a function of frequency for a 24/96 dithered 1kHz signal at the coaxial input. The 200k and 600 ohms differ by only about 10dB throughout the audioband, which is an indication that D50’s outputs are robust and can handle loads below 1k ohms with no difficulty. THD ratios into 200k ohms were extraordinarily low, ranging from 0.00003 to 0.00007% from 20Hz to 20kHz. These values are nearing the limits of what the APx555 can measure.

THD ratio (unweighted) vs. frequency vs. sample rate (16/44.1, 24/96)

The chart above shows THD ratios at the balanced line-level output into 200k ohms as a function of frequency for a 16/44.1 (blue/red) and a 24/96 (purple/green) dithered 1kHz signal at the coaxial input. The 24/96 data (right channel) consistently outperformed the 16/44.1 data by 15-20dB from 20Hz to about 1kHz due to the inherently higher noise floor at 16 bits (i.e., the analyzer cannot assign a THD value below the noise floor). THD ratios with 16/44.1 data ranged from 0.0002% from 20Hz to 1kHz, then down to just below 0.0001% past 10kHz.  24/96 THD ratios ranged from 0.00003-0.00007% from 20Hz to 20kHz.

THD ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD ratios measured at the balanced output as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). Once again, the 24/96 data outperformed the 16/44.1 data, with a THD range from 0.3% at 200uVrms to 0.00005% at 1.5-2.6Vrms, while the 16/44.1 ranged from 3% down to 0.0003% at the maximum output voltage of 2.6Vrms.

THD+N ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD+N ratios measured at the balanced output as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). The 24/96 outperformed the 16/44.1 data, with a THD+N range from 3% down to 0.0002% at 2.6Vrms, while the 16/44.1 ranged from 30% down to 0.002% at the maximum output voltage of 2.6Vrms.

Intermodulation distortion vs. generator level (SMPTE, 60Hz:4kHz, 4:1; 16/44.1, 24/96)

The chart above shows intermodulation distortion (IMD) ratios measured at balanced output for 16/44.1 (blue/red) input data and 24/96 input data (purple/green), from -60dBFS to 0dBFS. 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 16/44.1 data yields IMD ratios from 2% down to 0.003% at 0dBFS. The 24/96 data yields IMD ratios from 0.06% down to 0.0005% from -15 to 0dBFS.

FFT spectrum – 1kHz (digital input, 16/44.1 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1. The second (2kHz) and third (3kHz) signal harmonics are barely visible above the noise floor at -130dBrA, or 0.00003%. There are also no power-supply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/96. Due to the increased bit-depth, the noise floor is much lower compared to the 16/44.1 FFT at a very low -160dBrA. We see signal harmonics ranging from -130dBrA to below -150dBrA, or 0.00003% to 0.000003%, all the way to 20kHz (and beyond). The second (2kHz) signal harmonic shows a difference between the left (-130dBrA) and right (-140dBrA) channels. Here also, there are no power-supply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS, unbalanced output)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohm for the coaxial digital input, sampled at 24/96. Other than a few very low level (-150dBrA) spurious noise peaks to the left of the main signal peak, this FFT is identical to the FFT for the balanced output.

FFT spectrum – 1kHz (digital input, 16/44.1 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1 at -90dBFS. We see the signal peak at the correct amplitude, and no signal harmonics or noise peaks.

FFT spectrum – 1kHz (digital input, 24/96 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at -90dBFS. We see the signal peak at the correct amplitude, and one very low-level signal harmonic peak (3kHz) at -160dBrA, or 0.000001%.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 16/44.1)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 16/44.1. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, they would yield 2.6Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is not visible above the noise floor at -140dBrA, or 0.00001%, and the third-order modulation products, at 17kHz and 20kHz, are at -125dBrA, or 0.00006%.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 24/96. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, they would yield 2.6Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is at -150/-135dBrA (left/right), or 0.000003/0.00002%, and the third-order modulation products, at 17kHz and 20kHz, are at -130dBrA, or 0.00003%.

Intermodulation distortion FFT (line-level input, APx 32 tone)

Shown above is the FFT of the balanced line-level output of the D50 with the APx 32-tone signal applied to the input. The combined amplitude of the 32 tones is the 0dBFS reference, and corresponds to 2.6Vrms into 200k ohms. The intermodulation products—i.e., the “grass” between the test tones—are distortion products from the amplifier and below the absurdly low -150dBrA, or 0.000003%, level. This is a very clean IMD result.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Digital-to-Analog Converter Measurements

Created: 01 April 2024

Link: reviewed by Dennis Burger on SoundStage! Simplifi on April 1, 2024

General Information

All measurements were conducted using an Audio Precision APx555 B Series analyzer.

The CD 50n was evaluated as a fixed-output DAC and conditioned for 30 minutes at 0dBFS (2.3Vrms out) into 100k ohms before any measurements were taken.

The CD 50n offers three digital inputs: one coaxial S/PDIF (RCA), one optical S/PDIF (TosLink), and one USB. There are two sets of unbalanced (RCA) line-level outputs (fixed and variable) and one headphone output (1/4″ TRS). There is an analog volume control for the headphone output.

The CD 50n offers a few features and settings. The following are the default settings used for the coaxial input, unbalanced line-level outputs, using a 0dBFS input, unless otherwise specified:

• Line Out Level: Fixed
• Lock Range: Wide (Medium and Narrow also available)
• Filter: Filter 1 (Filter 2 also available)
• H/P Amplifier Gain: High (Mid and Low also available)

The analyzer’s input bandwidth filter was set to 10Hz–22.4kHz for all measurements, except for frequency response (DC to 1 MHz), FFTs (10Hz to 90kHz). and THD vs. Frequency (10Hz to 90kHz), with the latter to capture the second and third harmonics of the 20kHz output signal.

The CD 50n analog volume control for the headphone outputs appears to be a potentiometer. Channel-to-channel deviation proved typical for this type of volume control implementation.

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

Published specifications vs. our primary measurements

The table below summarizes the measurements published by Marantz for the CD 50n compared directly against our own. The published specifications are sourced from Marantz’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 as set at its maximum (DC to 1MHz), assume, unless otherwise stated, assume a fixed 2.34Vrms output (RCA) into 100k ohms and a measurement input bandwidth of 10Hz to 22.4kHz, and the worst-case measured result between the left and right channels.

Our primary measurements revealed the following using the coaxial input and the single-ended line-level outputs (unless specified, assume a 1kHz sinewave at 0dBFS, 100k ohms loading, 10Hz to 22.4kHz bandwidth):

Our primary measurements revealed the following using the coaxial input and the headphone output (unless specified, assume a 1kHz sinewave at 0dBFS, 300 ohms loading, 10Hz to 22.4kHz bandwidth, gain set to High):

Frequency response vs. sample rate (16/44.1, 24/96, 24/192, Filter 1)

The plot above shows the CD 50n’s frequency response as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 24/96 dithered digital input signal from 5Hz to 48kHz, and finally orange/pink represents 24/192 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 that of a shallow reconstruction filter (Filter 2 offers brickwall-type filtering). The -3dB point for each sample rate is roughly 17.5, 36.6, and 64.5kHz, respectively. 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 vs. filter setting (16/44.1)

The plots above show frequency response for a 0dBFS input signal sampled at 44.1kHz for Filter 1 (blue) and Filter 2 (red), into a 100k-ohm load for the left channel only. The graph is zoomed in from 1kHz to 22kHz to highlight the various responses of the filters. We can see that Filter 1 offers soft attenuation around the corner frequency, likely minimizing phase shift, with a -3dB point at 17.5 kHz, while Filter 2 is a brickwall-type filter, with a -3dB point at 20.9kHz. It’s worth pointing out that Filter 1 (the default filter) may be discernible with 16/44.1 content when compared to DACs with ruler-flat frequency responses using brickwall filters, depending on one’s age and high-frequency hearing acuity. The -1dB point is at roughly 12.5kHz.

Phase response (8-ohm loading, line-level input)

Above are the phase-response plots from 20Hz to 20kHz for the coaxial input, measured across at the unbalanced output, using both Filter 1 (blue) and Filter 2 (red) for a 16/44.1 input. We can see that the CD 50n does not invert polarity, with a worst-case phase shift of -80 degrees at 20kHz for Filter 2 (the brickwall filter). What Filter 1 loses in high-frequency response, it gains with zero phase shift in the audioband.

Digital linearity (16/44.1 and 24/96 data)

The graph above shows the results of a linearity test for the coaxial digital input (the optical input performed identically) for both 16/44.1 (blue/red) and 24/96 (purple/green) input data, measured at the unbalanced line-level output of the CD 50n. 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 24/96 input data is essentially perfect down to -120dBFS, while the 16/44.1 input data performed well, only overresponding by 2.5/2dB (left/right) at -120dBFS. The 24/96 data yielded such superb results that we extended the sweep down to . . .

-140dBFS. Above we see that even at -140dBFS, the CD 50n is only overshooting by 1dB with a 24/96 signal (right channel, the left is still at 0dB). This is an exemplary linearity test result.

Impulse response (24/44.1 data, Filter 1 and Filter 2)

The graph above shows the impulse responses for a looped 24/44.1 test file that moves from digital silence to full 0dBFS (all “1”s), for one sample period then back to digital silence, fed to the coaxial digital input, measured at the unbalanced outputs into a 100k-ohm load for the left channel only. Filter 1 is blue and Filter 2 is red. We can see that Filter 1 is a simple filter with virtually no pre/post ringing. Filter 2 shows almost no pre-ringing, but significant post-ringing.

J-Test (coaxial, Lock Range Wide)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output of the CD 50n, using the Wide Lock Range setting. J-Test was developed by Julian Dunn in the 1990s. It is a test signal—specifically, a -3dBFS undithered 12kHz squarewave sampled (in this case) at 48kHz (24 bits). Since even the first odd harmonic (i.e., 36kHz) of the 12kHz squarewave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sinewave (the main peak). In addition, an undithered 250Hz squarewave 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 sinewave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The coaxial input shows a good J-Test result, with two peaks at the -130dBrA level clearly flanking the 12kHz primary peak. This is an indication that the CD 50n may have good jitter immunity with the Wide setting.

J-Test (coaxial, Lock Range Wide, 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results show visible sidebands at 10kHz and 14kHz, but at a very low -130dBrA. This is further evidence of the CD 50n’s strong jitter immunity using the Wide Lock Range setting.

J-Test (coaxial, Lock Range Narrow)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output of the CD 50n using the Narrow Lock Range setting. The result is identical to the Wide Lock Range setting.

J-Test (coaxial, Lock Range Narrow, 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz, without perfect jitter immunity. The result is very poor, with a significant increase in the noise floor. The Narrow Lock Range setting should not be used with sources that may be prone to jitter.

J-Test (coaxial, Lock Range Medium)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output of the CD 50n using the Medium Lock Range setting. The result is identical to the Wide and Narrow Lock Range settings.

J-Test (coaxial, Lock Range Medium, 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz, without perfect jitter immunity. The results show visible sidebands at 10kHz and 14kHz, but at an extremely low -145dBrA. This is evidence of the CD 50n’s very strong jitter immunity using the Medium Lock Range setting.

J-Test (optical, Lock Range Wide)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output of the CD 50n. The optical input is clearly worse than the coaxial input using the Wide Lock Range setting, with peaks as high as -105dBrA flanking the 12kHz fundamental.

J-Test (optical, Lock Range Wide, 100ns)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results show visible sidebands at 10kHz and 14kHz, at a significant -80dBrA. This shows that the Wide Lock Range setting on the CD 50n should not be used with the optical input if the source is prone to jitter.

J-Test (optical, Lock Range Narrow)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output of the CD 50n using the Narrow Lock Range setting. The result is identical to the Wide Lock Range setting for the coax input.

J-Test (optical, Lock Range Narrow, 100ns)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The result is very poor, with a significant increase in the noise floor. As with the coaxial input, the Narrow Lock Range setting should not be used with sources that may be prone to jitter.

J-Test (optical, Lock Range Medium)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output of the CD 50n using the Medium Lock Range setting. The result is identical to the wide Lock Range setting for the coax input.

J-Test (optical, Lock Range Medium, 100ns)

The plot above shows the results of the J-Test test for the optical digital input measured at the unbalanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz without perfect jitter immunity. The results show visible sidebands at 10kHz and 14kHz, but at an extremely low -145dBrA. This is evidence of the CD 50n’s very strong jitter immunity using the Medium Lock Range setting. For the optical input, this should be the preferred setting for sources that are prone to jitter.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (coaxial input, Filter 1)

The plot above shows a fast Fourier transform (FFT) of the CD 50n’s unbalanced-line level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green), using the Filter 1 filter. There is a soft roll-off above 20kHz as the white-noise spectrum shows. There are low-level aliasing artifacts in the audioband at -120dBrA at 6 and 13kHz. The primary aliasing signal at 25kHz is barely suppressed at -10dBrA.

Wideband FFT spectrum of white noise and 19.1kHz sine-wave tone (coaxial input, Filter 2)

The plot above shows a fast Fourier transform (FFT) of the CD 50n’s unbalanced line-level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green), using the Filter 2 filter. There is a steep roll-off above 20kHz in the white-noise spectrum due to the brick-wall filter. There are absolutely no imaged aliasing artifacts in the audioband above the -135dBrA noise floor. The primary aliasing signal at 25kHz is at -80dBrA.

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

The chart above shows THD ratios at the unbalanced line-level output into 100k ohms (blue/red) and 600 ohms (purple/green) as a function of frequency for a 24/96 dithered 1kHz signal at the coaxial input. The 100k and 600 ohms data are very close throughout the audioband, with only a 5dB increase in THD at 20kHz into the 600-ohm load. The right channel outperformed the left channel by roughly 5dB throughout. THD ratios into 100k ohms (right channel) ranged from 0.0003% from 20Hz to 5kHz, then up to 0.0005% at 20kHz.

THD ratio (unweighted) vs. frequency vs. sample rate (16/44.1, 24/96)

The chart above shows THD ratios at the unbalanced line-level output into 100k ohms as a function of frequency for a 16/44.1 (blue/red) and a 24/96 (purple/green) dithered 1kHz signal at the coaxial input. THD ratios are roughly equivalent between the 16/44.1 and 24/96 data, roughly between 0.0003 and 0.0005%,  with the same 5dB increase in THD between right and left channels seen in the previous graph.

THD ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD ratios measured at the unbalanced output as a function of output voltage for the coaxial input into 100k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). The 24/96 outperformed the 16/44.1 data at lower levels due to the increased noise floor with the lower 16-bit depth data (the analyzer cannot assign a THD ratio for peaks that do not manifest above the measured noise floor). For the 16/44.1 data, THD ratios ranged from 2% at 200uVrms, down to just below 0.0005% at the maximum output voltage of 2.34Vrms. The 24/96 THD ratios ranged from 0.1% at 200uVrms, down to the same 0.0005% at the maximum output voltage.

THD+N ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD+N ratios measured at the unbalanced output as a function of output voltage for the coaxial input into 100k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). The 24/96 outperformed the 16/44.1 data due to the increase noise floor with the lower 16-bit depth data. For the 16/44.1 data, THD+N ratios ranged from 20% at 200 uVrms, down to just 0.002% at the maximum output voltage of 2.34Vrms. The 24/96 THD+N ratios ranged from 1% at 200uVrms, down to 0.0005% at the maximum output voltage.

Intermodulation distortion vs. generator level (SMPTE, 60Hz:4kHz, 4:1, 16/44.1, 24/96)

The chart above shows intermodulation distortion (IMD) ratios measured at unbalanced output for 16/44.1 (blue/red) and 24/96 input data (purple/green), from -60dBFS to 0dBFS, for the coaxial input. 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 16/44.1 data yields IMD ratios from 2% down to 0.002% at 0dBFS. The 24/96 data yields IMD ratios from 0.1% down to 0.0005% from -10 to -5dBFS, then up to 0.001% at 0dBFS.

FFT spectrum – 1kHz, 16/44.1 at 0dBFS

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohms for the coaxial digital input, sampled at 16/44.1. We see signal harmonics dominate at the second (2kHz) and third (3kHz) position at roughly -110dBrA, or 0.0003%, but also visible at lower levels up to and beyond 20kHz. There is only one small power-supply noise peak—the left channel at 120Hz at -130dBFS, or 0.00003%.

FFT spectrum – 1kHz, 24/96 at 0dBFS

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohm for the coaxial digital input, sampled at 24/96. Due to the increased bit depth, the noise floor is much lower compared to the 16/44.1 FFT at a very low -150 to -160dBrA. We see signal harmonics ranging from -110dBrA to -150dBrA, or 0.0003% to 0.000003%, all the way to 20kHz (and beyond). Here again, the second (2kHz) and third (3kHz) signal harmonics dominate at roughly -110dBrA. We find power-supply-related noise peaks at the second (120Hz), fourth (240Hz), and eighth (480Hz) harmonics, at -130dBrA to -140dBrA, or 0.00003% to 0.00001%.

FFT spectrum – 1kHz, 16/44.1 at -90dBFS

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohm for the coaxial digital input, sampled at 16/44.1 at -90dBFS. We see the signal peak at the correct amplitude, and no signal harmonics above the noise floor within the audioband.

FFT spectrum – 1kHz, 24/96 at -90dBFS

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at -90dBFS. We see the signal peak at the correct amplitude, no signal harmonics, and the same power-supply-related noise peaks as seen in the 24/96 0dBFS FFT above.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 16/44.1)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the unbalanced output into 100k ohms for the coaxial input at 16/44.1. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 2.34Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz) is at -125dBrA, or 0.00006%, and the third-order modulation products, at 17kHz and 20kHz, are at -130dBrA, or 0.00003%.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the unbalanced output into 100k ohms for the coaxial input at 24/96. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 2.34Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is at -120dBrA, or 0.0001%, and the third-order modulation products, at 17kHz and 20kHz, are at -125dBrA, or 0.00006%.

Intermodulation distortion FFT (coaxial input, APx 32 tone, 24/96)

Shown above is the FFT of the unbalanced line-level output of the CD 50n with the APx 32-tone signal applied to the input. The combined amplitude of the 32 tones is the 0dBFS reference, and corresponds to 2.34Vrms into 100k ohms. The intermodulation products—i.e., the “grass” between the test tones—are distortion products from the amplifier and below the -140dBrA, or 0.00001%, level. This is a very clean IMD result.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Digital-to-Analog Converter Measurements

Created: 01 April 2024

Link: reviewed by Phil Gold on SoundStage! Ultra on April 1, 2025

General Information

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

The EMM Labs DV2i was evaluated as a DAC and conditioned for 30 minutes at 0dBFS (volume set to maximum) into 200k ohms before any measurements were taken.

The DV2i is marketed as an integrated DAC because it includes a digital volume control and streamer section. The volume-control knob is on the front panel. The DV2i offers six digital inputs: coaxial S/PDIF (RCA), optical S/PDIF (TosLink), AES-EBU (XLR), USB, a proprietary EMM Labs input, and a network (ethernet) digital input. There are two line-level outputs: balanced (XLR) and unbalanced (RCA). Comparisons were made between unbalanced and balanced line level outputs, where no appreciable differences were seen other than the extra 6dB of signal over the balanced outputs. 1kHz FFTs are nonetheless provided for both balanced and unbalanced outputs.

There are two technologies to make note of: the D2Vi’s proprietary “adaptive” filter, intended to give ideal time-domain or frequency-domain response depending on the characteristics of the incoming signal; and the single-bit (AKA DSD) digital-to-analog converter technology, once again proprietary to EMM Labs. These provide unique performance characteristics.

The analyzer’s input bandwidth filter was set to 10Hz-22.4kHz for all measurements, except for frequency response (DC to 1 MHz), FFTs (10Hz to 90kHz), and THD vs. frequency (10Hz to 90kHz). The latter was to capture the second and third harmonic of the 20kHz output signal.

The DV2i’s digital volume control offers 100 volume steps (the user can also select “dB” instead of “1” to “100” on the display). Volume steps 0 through 12 offer 2dB increments, levels 13 through 40 yield 1dB, and 41 through 100 yield 0.5dB resolution. Channel-to-channel deviation was very good, at around 0.006-0.008dB throughout the range.

Volume-control accuracy (measured at line-level outputs): left-right channel tracking

Published specifications vs. our primary measurements

The table below summarizes the measurements published by EMM Labs for the DV2i compared directly against our own. The published specifications are sourced from EMM Labs’s website, either directly or from the manual available for download, or a combination thereof. Assume, unless otherwise stated, 1kHz at 0dBFS into 200k ohms and a measurement input bandwidth of 10Hz to 22.4kHz:

Our primary measurements revealed the following using the coaxial input and the balanced line-level outputs (unless specified, assume a 1kHz sinewave at 0dBFS, 200k ohms loading, 10Hz to 22.4kHz bandwidth):

Frequency response vs. sample rate (16/44.1, 24/96, 24/192)

The plot above shows the DV2i’s frequency response as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 24/96 dithered digital input signal from 5Hz to 48kHz, and finally orange/pink represents 24/192 from 5Hz to 96kHz. The behavior at low frequencies is the same for all sample rates—perfectly flat down to 5Hz. The behavior at high frequencies for all three digital sample rates is soft filtering at 24/96 and 24/192, and closer to brickwall-type filtering at 16/44.1. For all three sample rates, the responses are at -0.25dB at 20kHz. The -3dB points for each sample rate are roughly 21kHz, 41kHz, and 70kHz, respectively.

Digital linearity (16/44.1 and 24/96 data)

The graph above shows the results of a linearity test for the coaxial digital input (the optical input performed identically) for both 16/44.1 (blue/red) and 24/96 (purple/green) input data, measured at the balanced line-level output of the DV2i. For this test, the digital input was swept with a dithered 1kHz input signal from -120dBFS to 0dBFS, and the output was 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 24/96 input data is essentially perfect down to -120dBFS, while the 16/44.1 input data performed well, only over-responding by 3dB (left/right) at -120dBFS. The 24/96 data yielded such superb results that we extended the sweep down to . . .

. . . -140dBFS. Above we see that even at -140dBFS, the DV2i is only overshooting by 2-5 dB between -140 and -130dBFS. This is an exemplary linearity-test result.

Impulse response

The graph above shows the impulse responses for a looped 24/44.1 test file that moves from digital silence to full 0dBFS (all “1”s), for one sample period then back to digital silence, fed to the coaxial digital input, measured at the balanced outputs, into a 200k ohm load for the left channel only. We can see that DV2i yields an impulse response with essentially no pre- or post-ringing behaviour, or one that emulates a non-oversampling DAC. This is a surprising result given the extended and near brickwall-type frequency response for a 16/44.1 input signal. The DV2i’s adaptive filters appear to, at least according to our tests, provide the best of both worlds: near perfect time-domain and frequency-domain response.

J-Test (coaxial input)

The plot above shows the results of the J-Test test for the coaxial digital input measured at the balanced line level output of the DV2i. J-Test was developed by Julian Dunn the 1990s. It is a test signal—specifically,  a -3dBFS undithered 12kHz squarewave sampled (in this case) at 48kHz (24 bits). Since even the first odd harmonic (i.e., 36kHz) of the 12kHz squarewave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sinewave (the main peak). In addition, an undithered 250Hz squarewave 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 sinewave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The coaxial input shows a strong J-Test result, with peaks visible, but only below the -140dBrA level centered around the main signal peak.

J-Test (optical input)

The optical input shows effectively the same result as the coaxial input.

J-Test (AES-EBU input)

The AES-EBU balanced input shows basically the same result as the coaxial and optical inputs.

J-Test (coaxial input, 2kHz sinewave jitter at 100ns)

The plot above shows the results of the J-Test test for the coaxial digital input (the optical and AES-EBU inputs behaved the same) measured at the balanced line-level output, with an additional 100ns of 2kHz sinewave jitter injected by the APx555, which would manifest as sideband peaks at 10kHz and 14kHz, without perfect jitter immunity. The results show no visible sidebands. This is further evidence of the DV2i’s strong jitter immunity.

Wideband FFT spectrum of white noise and 19.1kHz sinewave tone

The plot above shows a fast Fourier transform (FFT) of the DV2i’s balanced-line level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green). There is a soft roll-off above 20kHz in the white-noise spectrum. This contradicts the brickwall-type results we found in our frequency response plots at 16/44.1. The DV2i, with its adaptive filters, appears to behave differently depending on the type of signal fed to it (we assume that’s the point). Most importantly, there are absolutely no imaged aliasing artifacts in the audioband above the -135dBrA noise floor. The primary aliasing signal at 25kHz is heavily suppressed at -120dBrA. The general rise in the noise floor above 20kHz is likely due to the DV2i’s DSD processing.

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

The chart above shows THD ratios at the balanced line-level output into 200k ohms (blue/red) and 600 ohms (purple/green) as a function of frequency for a 24/96 dithered 1kHz signal at the coaxial input. The 200k and 600 ohms data are identical throughout the audioband, which is an indication that the DV2i’s outputs are robust and can handle loads below 1k ohms with no difficulty. There was an evident difference in THD ratios between the left and right channels, with the left channel outperforming the right by about 5dB from 20Hz to 2kHz. THD ratios (left channel) ranged from 0.00009% from 20Hz to 1.5kHz, then up to roughly 0.001% at 4.5kHz through to 20kHz. The higher THD ratios at higher frequencies were seen in all plots and are due to the rising high-frequency noise floor due to the DSD processing (the analyzer cannot assign a THD value below the noise floor).

THD ratio (unweighted) vs. frequency vs. sample rate (16/44.1, 24/96)

The chart above shows THD ratios at the balanced line-level output into 200k ohms as a function of frequency for a 16/44.1 (blue/red) and a 24/96 (purple/green) dithered 1kHz signal at the coaxial input. The 24/96 data (left channel) consistently outperformed the 16/44.1 data by 10dB from 20Hz to about 2kHz due to the inherently higher noise floor at 16 bits (i.e., the analyzer cannot assign a THD value below the noise floor). THD ratios with 16/44.1 data ranged from 0.0003% from 20Hz to 2.5kHz, then up to roughly 0.001% at 4.5kHz through to 20kHz. The higher THD ratios at higher frequencies were seen in all plots and are due to the rising high-frequency noise floor because of the DSD-type processing.

THD ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD ratios measured at the balanced outputs as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green), with the volume set to maximum. Once again, the 24/96 outperformed the 16/44.1 data, with a THD range from 1% at 300uVrms to 0.0002% at 4Vrms, while the 16/44.1 ranged from 3% down to 0.0003% at the maximum output voltage of 4Vrms.

THD+N ratio (unweighted) vs. output (16/44.1, 24/96)

The chart above shows THD+N ratios measured at the balanced outputs as a function of output voltage for the coaxial input into 200k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green), with the volume set to maximum. The 24/96 outperformed the 16/44.1 data, with a THD+N range from 5% down to 0.0005% at 4Vrms, while the 16/44.1 ranged from 30% down to 0.002% at the maximum output voltage of 4Vrms.

Intermodulation distortion vs. generator level (SMPTE, 60Hz:4kHz, 4:1; 16/44.1, 24/96)

The chart above shows intermodulation distortion (IMD) ratios measured at balanced output for 16/44.1 (blue/red) input data and 24/96 input data (purple/green), from -60dBFS to 0dBFS. 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 16/44.1 data yielded IMD ratios from 2% down to 0.002% at 0dBFS. The 24/96 data yielded IMD ratios from 0.3% down to 0.0006% from -5 to 0dBFS.

FFT spectrum – 1kHz (digital input, 16/44.1 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1. We see signal harmonics at 2 and 3kHz. The second (2kHz) harmonic is at -125/-130dBrA, or 0.00006/0.00003%, and the third harmonic (3kHz) is at -130dBrA (right visible only), or 0.00003%. There are also no power-supply noise peaks to speak of to the left of the main signal peak.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohms for the coaxial digital input, sampled at 24/96. Due to the increased bit-depth, the noise floor is much lower compared to the 16/44.1 FFT at a low -150dBrA. We see signal harmonics ranging from -120dBrA to -140dBrA, or 0.0001% to 0.00001%, at 2/3/4/5kHz. Here also, there are no power-supply noise peaks to speak of to the left of the main signal peak. The rise in the noise floor above 20kHz is due to the DSD processing.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS, unbalanced output)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohm for the coaxial digital input, sampled at 24/96. It is essentially identical to the FFT above using the balanced inputs.

FFT spectrum – 1kHz (digital input, 16/44.1 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 16/44.1 at -90dBFS. We see the signal peak at the correct amplitude, and no signal harmonics or noise peaks.

FFT spectrum – 1kHz (digital input, 24/96 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at -90dBFS. We see the signal peak at the correct amplitude, and essentially no signal harmonics or noise peaks.

FFT spectrum – 1kHz (digital input, 24/96 data at -80dBFS)

This is not a test we typically do, but because the DV2i has an onboard volume control, we wanted to show if there were differences between lowering the signal level of the analyzer versus using the volume control at its lowest level (-80dBFS). Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at -80dBFS with the volume set to maximum but with the analyzer’s signal level reduced to output the correct level. We see the signal peak at the correct amplitude, and essentially no signal harmonics or noise peaks.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS with volume control set to -80dBFS, its lowest level)

This is not a test we typically do, but because the DV2i has an onboard volume control, we wanted to show if there were differences between lowering the signal level of the analyzer versus using the volume control. Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the coaxial digital input, sampled at 24/961 at 0dBFS with the volume set to minimum (-80dB). We see essentially the same FFT as above with the volume at maximum and the input signal at -80dBFS. This is good evidence for EmmLabs’s claims of a transparent digital volume control.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 16/44.1)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 16/44.1. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 4Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is not visible above the noise floor at -135dBrA, or 0.00002%, and the third-order modulation products, at 17kHz and 20kHz, are at -135dBrA, or 0.00002%. The signals to the right of the 18kHz + 19kHz summed sinewave are presumably the result of aliasing artifacts due to the nature of the adaptive digital filter.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced output into 200k ohms for the coaxial input at 24/96. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 4Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz) is at -135/-125dBrA (left/right), or 0.00002/0.00006%, and the third-order modulation products, at 17kHz and 20kHz, are at -130dBrA, or 0.00003%.

Diego Estan
Electronics Measurement Specialist

Details

Parent Category: Products

Category: Digital-to-Analog Converter Measurements

Created: 01 March 2024

Link: reviewed by Roger Kanno on SoundStage! Hi-Fi on March 1, 2024

General Information

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

The Magnetar UDP900 was evaluated as a DAC and was conditioned for 30 minutes at 0dBFS (1.95Vrms RCA out) into 100k ohms before any measurements were taken. But as mentioned below, the headphone output was also measured.

The UDP900 is a universal 4k UHD Blu-ray player. It offers one digital input (asynchronous USB) allowing for its evaluation as a DAC. It is important to note that the Audio Precision (AP) analyzer does not have a dedicated digital-audio output over USB. Audio data over USB to the device under test (DUT) is achieved via a computer (in our case a Lenovo ThinkPad X1 laptop running Windows 11) running the APx software controlling the AP analyzer. The dedicated Magnetar Windows USB driver for the UDP900 was downloaded from the Magnetar website and installed on our laptop. The driver control panel allows for the selection of 16-bit or 24-bit two-channel data. This, along with the APx software controlling the sample rate, allowed for true 16-bit/44.1kHz, 24/96, and 24/192 audio data to be sent to the UDP900 DAC. The UDP900 has seven user-selectable digital-filter options, but for the digital measurements below, the default filter, labeled Brick Wall, was used.

The UDP900 has both balanced (XLR) and unbalanced (RCA) line-level analog outputs. Typically, we find very little performance difference between both types of outputs (other than an extra 6dB of gain over balanced). In the case of the UDP900, however, noticeably more THD was measured over the balanced outputs. Further, the balanced outputs yielded 7.5dB (as opposed to the typical 6dB) more gain than the unbalanced outputs. At 0dBFS (1.95Vrms over RCA and 4.6Vrms over XLR), the balanced outputs yielded 5dB more THD at 1kHz, and a very significant 20dB more at 20kHz (graphs included in this report) than the balanced ouptuts. For this reason, unless otherwise stated, the unbalanced (RCA) analog outputs were used. The UDP900 also offers a ¼ ″ TRS headphone output, which was also evaluated.

The analyzer’s input bandwidth filter was set to 10Hz–22.4kHz for all measurements, except for frequency response (DC to 1 MHz), FFTs (10Hz to 90kHz), and THD vs frequency (10Hz to 90kHz). The latter to capture the second and third harmonics of the 20kHz output signal.

Published specifications vs. our primary measurements

The table below summarizes the measurements published by Magnetar for the UDP900 compared directly against our own. The published specifications are sourced from Magnetar’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, 1.95Vrms output (RCA) into 100k ohms and a measurement input bandwidth of 10Hz to 22.4kHz, and the worst-case measured result between the left and right channels.

Our primary measurements revealed the following using the USB input and the unbalanced line-level outputs (unless specified, assume a 1kHz sinewave at 0dBFS, 100k ohms loading, 10Hz to 22.4kHz bandwidth):

Our primary measurements revealed the following using the USB input and the headphone output (unless specified, assume a 1kHz sinewave at 0dBFS, 300 ohms loading, 10Hz to 22.4kHz bandwidth):

Frequency response (16/44.1, 24/96, 24/192 with Brick Wall filter)

The plot above shows the UDP900’s frequency-response (relative to 1kHz) as a function of sample rate. The blue/red traces are for a 16-bit/44.1kHz dithered digital input signal from 5Hz to 22kHz, the purple/green traces are for a 24/96 dithered digital input signal from 5Hz to 48kHz, and finally orange/pink represents 24/192 from 5Hz to 96kHz. The behavior at low frequencies is the same for all digital sample rates—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 20.3, 44 and 81kHz respectively. The ripples (about +/- 0.2dB) in the frequency responses at higher frequencies are real—confirmed with steady-state measurements.  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.

Digital linearity (16/44.1 and 24/96 data)

The graph above shows the results of a linearity test for the digital input for both 16/44.1 (blue/red) and 24/96 (purple/green) input data, measured at the line-level unbalanced outputs of the UDP900. The digital input was swept with a dithered 1kHz input signal from -120dBFS to 0dBFS, and the output was 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 24/96 input data is perfect down to -120dBFS, while the 16/44.1 input data began to over-shoot significantly below -90dBFS. The 24/96 data yielded such superb results that we extended the sweep down. . .

Digital linearity (16/44.1 and 24/96 data)

. . . to -140dBFS. Above we see that even at -140dBFS, the UDP900 is only overshooting by 1dB. This is an exemplary linearity test result for the 24/96 data, but somewhat poor for 16/44.1 data (a good result would be flat down to -100dBFS).

Impulse response

The graph above shows the impulse responses for a looped 24/44.1 test file that moves from digital silence to full 0dBFS (all “1”s) for one sample period then back to digital silence, fed to the digital input, measured at the unbalanced analog outputs, for the left channel only. We can see that UDP900 DAC reconstruction filter exhibits symmetrical pre/post ringing as seen in a typical sinc function.

J-Test (USB input)

The plot above shows the results of the J-Test test for the USB input measured at the unbalanced line-level output of the UDP900. J-Test was developed by Julian Dunn the 1990s. It is a test signal—specifically, a -3dBFS undithered 12kHz squarewave sampled (in this case) at 48kHz (24 bits). Since even the first odd harmonic (i.e., 36kHz) of the 12kHz squarewave is removed by the bandwidth limitation of the sampling rate, we are left with a 12kHz sinewave (the main peak). In addition, an undithered 250Hz squarewave 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 sinewave jitter by the Audio Precision, to also show how well the DAC rejects jitter.

The USB digital input shows an average J-Test result, with a few peaks at the -125dBrA level and below, clearly visible both near the primary 12kHz signal peak, and below 1kHz. This is an indication that the UDP900 may be sensitive to jitter.

Wideband FFT spectrum of white noise and 19.1kHz sinewave tone (Brick Wall filter)

The plot above shows a fast Fourier transform (FFT) of the UDP900’s line-level output with white noise at -4 dBFS (blue/red) and a 19.1 kHz sinewave at 0dBFS fed to the coaxial digital input, sampled at 16/44.1 (purple/green). There is a steep roll-off above 20kHz in the white-noise spectrum, characteristic of a brickwall-type filter. There are no imaged aliasing artifacts in the audioband above the -130dBrA noise floor. The primary aliasing signal at 25kHz is at -100dBrA.

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

The chart above shows THD ratios at the line-level output into 100k ohms (blue/red) and 600 ohms (purple/green) as a function of frequency for a 24/96 dithered 1kHz signal at the USB input. Also shown are THD ratios for the balanced output (pink/orange) into a 200k ohm load. The 100k and 600 ohms data are identical throughout the audioband, which is in indication that UDP900’s outputs are robust and can handle loads below 1k ohms with no difficultly. THD ratios from the unbalanced outputs into 100k ohms ranged from 0.005% from 20Hz to 800Hz, then up to 0.01% at 3kHz, then down to 0.006%at 20kHz. The balanced outputs yielded THD ratios from 0.005% from 20Hz to 100Hz, then a steady rise to 0.06% at 20kHz.

THD ratio (unweighted) vs. frequency vs. sample rate (16/44.1 and 24/96)

The chart above shows THD ratios at the line-level output into 100k ohms as a function of frequency for a 16/44.1 (blue/red) and a 24/96 (purple/green) dithered 1kHz signal at the optical input. THD ratios were identical and ranged from 0.005% from 20Hz to 800Hz, then up to 0.01% at 3kHz, then down to 0.006% at 20kHz.

THD ratio (unweighted) vs. output (16/44.1 and 24/96)

The chart above shows THD ratios measured at the line-level unbalanced output as a function of output voltage for the USB input into 100k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). Also shown are THD ratios for the balanced output (pink/orange) into a 200k ohm load. For the unbalanced output, the 24/96 data outperformed the 16/44.1 data at low output voltages by roughly 20-30dB, with a THD range from 0.1% at 200uVrms to 0.0002% at 0.25Vrms, then up to 0.005% at the maximum output voltage of 1.95Vrms. The 16/44.1 data ranged from 10% down to 0.002% at 0.5 to 1Vrms, then up to 0.005% at 1.95Vrms. The balanced output (at 24/96) yielded slightly higher THD ratios (2-3dB) than the unbalanced output up to about 0.1Vrms. From 0.1 to 1Vrms, THD ratios over the balanced output were as much as 15dB higher than the unbalanced output.

THD+N ratio (unweighted) vs. output (16/44.1 and 24/96)

The chart above shows THD+N ratios measured at the line-level unbalanced output as a function of output voltage for the USB input into 100k ohms from -90dBFS to 0dBFS at 16/44.1 (blue/red) and 24/96 (purple/green). Also shown are THD+N ratios for the balanced output (pink/orange) into a 200k ohm load. For the unbalanced output, the 24/96 data outperformed the 16/44.1 data at low output voltages by roughly 15dB, with a THD+N range from 1% at 200uVrms to 0.0015% at 0.5Vrms, then up to 0.006% at the maximum output voltage of 1.95Vrms. The 16/44.1 data ranged from 10% down to 0.003% at 1Vrms, then up to 0.006% at 1.95Vrms. The balanced output (at 24/96) yielded slightly higher THD+N ratios (2-3dB) than the unbalanced output up to about 0.2Vrms. From 0.2 to 1Vrms, THD+N ratios over the balanced output were as much as 5dB higher than the unbalanced output.

FFT spectrum – 1kHz (digital input, 16/44.1 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the line-level unbalanced output into 100k ohm for the USB digital input, sampled at 16/44.1. The third (3kHz) signal harmonic dominates at -85dBrA, or 0.006%. There are also a multitude of low levels peaks from 100Hz to 20kHz just below the -120dBrA, or 0.001%, level.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the unbalanced output into 100k ohm for the USB digital input, sampled at 24/96. We see the same signal harmonic dominate at 3kHz at -85dBrA, or 0.006%, as is seen in the 16/44.1 FFT above. We find a power-supply-related noise peak at 120Hz at -130dBrA, or 0.00003%.

FFT spectrum – 1kHz (digital input, 24/96 data at 0dBFS, balanced output)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the balanced output into 200k ohm for the USB digital input, sampled at 24/96. Compared to the FFT above with the unbalanced output, here we see clearly visible higher odd-ordered signal harmonics (5/7/9/11kHz, etc.) from -95dBRa, or 0.002%, to -130dBrA, or 0.00003%, at 20kHz.

FFT spectrum – 1kHz (digital input, 16/44.1 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the line-level output into 100k ohms for the USB digital input, sampled at 16/44.1 at -90dBFS. We see the signal peak at the correct amplitude, with significant odd-ordered signal harmonics at -100dBrA, or 0.001% and below.

FFT spectrum – 1kHz (digital input, 24/96 data at -90dBFS)

Shown above is the fast Fourier transform (FFT) for a 1kHz input sinewave stimulus, measured at the line-level output into 100k ohms for the USB digital input, sampled at 24/96 at -90dBFS. We see the signal peak at the correct amplitude, no signal harmonics, and even-ordered power-supply-related noise (120/240/360Hz) dominated by the 120Hz peak at -125dBrA, or 0.00006%.

Intermodulation distortion vs. generator level (SMPTE, 60Hz:4kHz, 4:1, 16/44.1, 24/96)

The chart above shows intermodulation distortion (IMD) ratios measured at the line-level unbalanced output for 16/44.1 (blue/red) input data and 24/96 input data (purple/green), from -60dBFS to 0dBFS. 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 16/44.1 data yields IMD ratios from 1% down to 0.005% at -10dBFS, then up to 0.03% at 0dBFS. The 24/96 data yields IMD ratios from 0.1% down to 0.002% at -15dBFS, then up to 0.03% at 0dBFS.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 16/44.1)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the unbalanced line-level output into 100k ohms for the USB input at 16/44.1. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 1.95Vrms (0dBrA) at the output. We find it difficult to identify the second-order modulation product (i.e., the difference signal of 1kHz) amongst the array of noise peaks just below the -120dBrA, or 0.0001%, level, while the third-order modulation products, at 17kHz and 20kHz, are at -95dBrA, or 0.002%.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the line-level output into 100k ohms for the USB input at 24/96. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 1.95Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is at -135dBrA (right channel only), or 0.00002%, and the third-order modulation products, at 17kHz and 20kHz, are at -95dBrA, or 0.002%.

Intermodulation distortion FFT (18kHz + 19kHz summed stimulus, 24/96, balanced output)

Shown above is an FFT of the intermodulation distortion (IMD) products for an 18kHz + 19kHz summed sinewave stimulus tone measured at the balanced line-level output into 200k ohms for the USB input at 24/96. The input dBFS values are set at -6.02dBFS so that, if summed for a mean frequency of 18.5kHz, would yield 4.6Vrms (0dBrA) at the output. We find that the second-order modulation product (i.e., the difference signal of 1kHz), is at -130dBrA, or 0.00003%, and the third-order modulation products, at 17kHz and 20kHz, are at -70dBrA, or 0.03%. Again, the result for the balanced output is much worse than for the unbalanced output.

Intermodulation distortion FFT (24/96 input, APx 32 tone)

Shown above is the FFT of the unbalanced outputs of the UDP900 with the APx 32-tone signal applied to the input. The combined amplitude of the 32 tones is the 0dBrA reference, and corresponds to 1.95Vrms. The intermodulation products—i.e., the “grass” between the test tones—are distortion products and are around the extremely low -120dBrA, or 0.0001%, level.

Diego Estan
Electronics Measurement Specialist