Headphone Output
Sound Quality& Comparisons
The sound quality of the 1 watt discrete class A headphone amplifier with its smooth and clean sound signature is very impressive. Yulong refined DA8 class A headphone amplifier and implemented it on DA8II. My measurement shows that the output power is actually around 1.4 watt at 32 ohms, higher than the 1 watt specification. Before I received my Geek Pulse XFi in late June this year, for more than 1.5 years, my main desktop DACs were Yulong DA8 and Mytek Stereo192-DSD. Most of the time, I would say 95% of my total use of the headphone output, I used the Yulong DA8 headphone output. I find Mytek Stereo192-DSD headphone output is thin sounding and less engaging. DA8 headphone output is simply sounds better, fuller, with more dynamic than Mytek 192-DSD headphone output. I'm glad that Yulong has improved the DA8 headphone output even further on DA8II. DA8II headphone output sound signature is more natural than the slightly warm DA8 headphone output. It retains the smooth grain free signature, and improves on the transparency, speed, dynamic, and detail. Sound signature is neutral, not analytic and not warm either. It has good bass and midrange body, not thin, with good detail and dynamic.
Quality of DA8II headphone amplifier is comparable to a good desktop headphone amplifier. With around 6.8 Vrms max voltage, and 1.4 watt at 32 ohms, it practically drives almost anything from low impedance to 600 ohms headphones, except ‘hard to drive’ planar like my Hifiman HE-6. I have Violectric HPA V200 desktop amplifier that is connected to one of my desktop DAC most of the time, and difference of the sound quality between using the V200 and connecting directly to the build-in headphone output of DA8II is not night and day. Sound from V200 connected to DA8II line output has slightly higher in clarity (but also slightly more grainy), slightly better drive, dynamic and detail, but only on the level of slightly better, and DA8II headphone output quality follows pretty close. V200 also sounds dryer than DA8II headphone amplifier, sounds like the decay is tad shorter. For some recordings, I do feel Yulong DA8II headphone amp decay sounds more natural. Generally both sound great with all the headphones I tested. Unlike my Mytek Stereo192-DSD which headphone amplifier is just like an additional feature of the DAC, Yulong DA8II headphone amplifier can be considered one of the main features. Yulong DA8II is probably one of the best one box solutions of DAC + Amp at this price range.
Before moving forward with headphone output comparisons and measurement, there are two things that I would like to mention here:
1. Tight headphone socket.
I had a chance to check 3 demo units of Yulong DA8ii before taking a unit for review, and I found that on most units, the headphone socket grips rather too tight, tighter than DA8 headphone socket. Sometime it is too tight that I have to hold DA8ii chassis to unplug the headphone jack. I hope Yulong take note on this, and make some adjustment for the headphone socket, to make it less tight, but not too loose either.
2. Distortion check.
I found one of the demo units has the headphone amplifier that is slightly distorted on one channel. The distortion was only audible when using IEMs. When using headphones the distortion was not audible, the distorted headphone output was only sounded more analytic on headphones, but was not distorted. So my advice when getting Yulong DA8, or actually any DAC & headphone amplifier, check the headphone output with IEMs. Small distortion is more likely to be more audible on sensitive IEMs.
Comparison with Yulong DA8 headphone output
Compared to the older Yulong DA8 headphone output, Overall improvement that I can hear from the DA8II headphone output are transparency, detail, speed, and tightness. Since there is no way to compare only the headphone amplifier section, it is more of an overall improvement of the whole unit and not merely the headphone amplifier section alone. In general DA8II is slightly livelier than DA8. Transparency and clarity are also improved. DA8 headphone output is slightly warmer, bass is slightly fatter, and also slightly less in perceived details. DA8 sonic signature is a touch warm, while DA8II is more towards neutral. More obvious differences is on the bass, where I can hear better punch, speed, texture and tightness on DA8II, where DA8 bass is generally fatter and less textured. But the overall differences are actually quite mild, and not always obvious on some recordings. For some brighter sounding headphones, DA8 warm and fatter signature is actually can be an advantage, to tame and smoothen treble brightness. While for warmer headphones, DA8II shines. I would like to state again, that the improvement is more in level of refinements, and it doesn’t shift DA8II signature too much from DA8 signature. Even if the detail, clarity, transparency, and speed have been improved, it doesn’t make Yulong DA8II sounds analytic. I’ve been listening to Beyerdynamic T1 from Yulong DA8II for hours without any fatigue, indicating that DA8II is not analytical sounding. Otherwise, I won’t be able to enjoy my T1 and HD800 with Yulong DA8II. In my opinion both DA8 and DA8II have excellent headphone amplifier section, but overall DA8II has slightly better sound quality.
Comparison with Geek Pulse XFi headphone output
Observing the headphone output sound quality of both, despite of having different sound signature, Yulong DA8II and Geek Pulse XFi are about in the same class, and in my opinion, better than Mytek 192-DSD headphone output. Most obvious is in the treble signature where Geek Pulse XFi treble sound slightly softer. Geek Pulse XFi also sounds a tad fuller around the midrange and bass area, with stronger bass presence. On the other hand, DA8II sounds slightly smoother, less dry, and more transparent, while still having good body on midrange and bass. Treble extension seems to be slightly better on DA8II, while XFi has more powerful bass. DA8II also perceived as sounding a little more spacious. Both headphone outputs are quite powerful, with XFi has higher output power. Dynamic and drive ability are quite comparable between the two, with DA8II sounds a tad faster in transient, and also slightly more forward in presentation. At the end it is more about headphone matching and personal preferences. For HD800 and T1 for example, I do prefer XFi stronger bass signature. But for Shure SRH1540 and ATH-R70x, I prefer the more transparent Yulong DA8II.
Geek Pulse XFi wins for the feature of balanced headphone output. Powerful balanced headphone output is really useful to drive ‘hard to drive headphone’ such as my Hifiman HE-6. Hopefully the next generation of DA8, probably DA9, will be equipped with a powerful balanced headphone output.
Comparison with Mytek Stereo192-DSD headphone output
Mytek headphone amplifier section, that probably is using TPA6102A2 chip family, with 5 ohms output impedance in the headphone output circuit, though sounds pretty decent and quite transparent, is not really up to the level of the DA8, DA8II and Geek Pulse XFi headphone amplifier. Mytek headphone output as mentioned before, sounds thinner in comparison, lacking of fullness and dynamic around bass to low midrange area. Overall it is more analytical and less musically engaging. Yulong DA8II headphone output is clearly better in comparison to Mytek Stereo192-DSD headphone output.
In summary, Yulong DA8II, DA8, and Geek Pulse XFi headphone output are comparable, while feature wise Geek Pulse XFi wins due to the balanced headphone output. Yulong DA8 headphone output quality is actually pretty close to DA8II, with DA8II is slightly better in clarity, detail, dynamic, and overall sounds livelier. While Mytek Stereo192-DSD headphone output is not really in the same league as the others, lacking fullness, dynamic, and the noisiest for IEM use.
On-Off Popping Noise
There is mild to moderate popping noise when switching the power on or off with headphone plugged in. Only mild popping noise on higher impedance headphones, but can be moderately loud on sensitive IEMs. Not up to the level that the popping sound might spoil the headphone or IEM, but especially for IEM, it is better to switch on or off without any IEM connected to the headphone output, especially when switching on. Using external desktop amplifier with balanced connection reduces the On-Off popping noise. In comparison with other DACs popping noise, DA8II headphone output on-off popping noise is actually softer than Mytek 192-DSD and Geek Pulse XFi, only slightly louder than Yulong DA8. I recorded the popping noise of the headphone output using Creative SB E5. I power cycle each DAC for 3 times starting from switching to On then Off a few seconds later. Starting with Yulong DA8II (blue spikes), Yulong DA8 (Yellow spikes), Geek Pulse XFi (Green spikes), and Mytek Stereo192-DSD (Red spikes) at the last. We can see from the screenshot below, the height of the pulses, Geek Pulse XFi and Mytek 192-DSD on-off popping noise are louder than Yulong DA8II.
Noise Floor & IEM
Headphone output noise floor observation is practically more relevant to IEM application, and often heard as hiss noise when player is silent or playing soft passages of a recording. Every IEM has different sensitivity to hiss noise, and it is also related to human hearing equal loudness curve. Meaning, the perceived loudness of the hiss noise is depending on the harmonic frequency of the noise floor. If most of the harmonic frequency of the noise is outside human sensitive frequency region, it might not be audible even if it is measured rather high.
Honestly, noise floor is not an easy subject to discuss, simply because with measurement equipment that I have, what is measured is not always related to what is heard. What is measured with the highest noise floor rms value, might be heard as the quieter amp. I don't have precision audio analyzer, and my USB oscilloscope (Owon VDS3102) is not the best instrument for noise measurement. So what I did is simply observing what I can see from the oscilloscope window to estimate the Vrms & Vpp value, and FFT window observing the harmonic frequencies. From the 4 desktop DACs tested, LH Geek Pulse XFi is showing the highest noise floor value, but it is also the quietest, with lowest audible hiss noise when tested with the most sensitive IEM I have. Yulong DA8 and DA8II on the other hand, showing much lower noise floor value on oscilloscope than Geek Pulse XFi, but they have louder perceived hiss noise. Probably this phenomenon might be explained by the FFT graph. Geek Pulse XFi FFT graph shows a very different pattern than other DACs. The harmonic frequency of the noise floor is evenly distributed without any particular peaks in the 0 to 50 kHz region, like a very soft white noise, but it has pretty high of ultra-high frequency noise in the range of megahertz, which probably explains why the noise floor sounds softer even though it is shown quite high on the oscilloscope. While other DACs though showing lower noise value, their noise floor harmonic frequencies has some peaks on certain frequencies, and not evenly distributed like Geek Pulse XFi white noise kind of pattern. The few frequency peaks are probably what cause the noise floor to be more around audible as compared to soft white noise kind of noise. But this is just my observation. Beside some googling, I haven't studied much about noise floor behaviour.
From all the IEMs that I have, 1964 Ears V3 and DUNU DN-2000 are some of the most sensitive IEMs to amplifier hiss noise. So observation of the hiss noise perceived loudness is based on what I heard using 1964 Ears V3 and DN-2000.
Test setup:
32 ohms resistor was used as dummy load. Some amplifiers are not stable without load.
Owon VDS3102 100 MHz oscilloscope, set to 20 MHz bandwidth.
Oscilloscope Phosphorus screen digital persistence mode was enabled.
Output volume was set between 500 mV - 540 mV (rms) when playing 100 Hz 0 dBFS. But actually for digital volume control, I observed noise floor level is quite consistent across the volume level, and doesn’t change with volume.
Noise floor is observed while playing silent track.
Digital oscilloscope input channels are set to DC coupled to show the level of DC offset.
Owon VDS3102 is used to estimated noise floor RMS value and harmonic frequency.
DUNU -DN-2000 is used for subjective observation of hiss noise perceived loudness.
In summary, all DACs headphone amplifier has typical noise floor level of desktop amp. Geek Pulse XFi has the lowest perceived hiss noise, that is practically almost ignore-able, but on oscilloscope it is shown to have highest level of noise floor. Geek Pulse XFi noise when observed using 1964 Ears V3, I can hear a kind of noise pattern, not totally even hiss noise. But using DN-2000 the noise pattern is less obvious, more just like even sounding soft hiss noise. The other 3 DACs, Yulong DA8, Yulong DA8II, and Mytek Stereo192-DSD have approximately the same level of perceived hiss noise, with only little differences between them. Mytek Stereo192-DSD is slightly noisier, therefore it has the highest perceived hiss noise among the 4 DACs. In practical, I would say the level of headphone output hiss noise from all the 4 DACs tested here are low enough for general less sensitive IEMs, practically almost ignore-able. But for very sensitive IEMs can be a little annoying.
Below is the 'subjective' hiss noise perceived loudness, from the softest to the loudest:
Geek Pulse XFi (softest)
Yulong DA8 & Yulong DA8II (about the same)
Mytek Stereo192-DSD (loudest)
While the Vpp and Vrms fluctuation value of the noise floor observed from the digital oscilloscope, is as the following:
Yulong DA8: Vpp ≈4-5 mV ; Vrms ≈12-14 mV
Yulong DA8II: Vpp ≈5-9 mV ; Vrms ≈6-10 mV
Mytek Stereo192-DSD: Vpp ≈7-13 mV ; Vrms ≈6-9 mV
Geek Pulse XFi: Vpp ≈33-45 mV ; Vrms ≈23-28 mV
Oscilloscope base noise floor with cable and 32 ohms load connected:
Yulong DA8:
Yulong DA8II:
Mytek Stereo192-DSD:
Geek Pulse XFi:
Yulong DA8 FFT screenshots:
Yulong DA8II FFT screenshots:
Mytek Stereo192-DSD FFT screenshots:
Geek Pulse XFi FFT screenshots:
Geek Pulse XFi ultra-high frequency harmonics FFT screenshots:
I have no idea why my Geek Pulse XFi headphone output is showing pretty high noise value, but in practical it has the lowest perceived hiss noise. From a YouTube video that showing the board layout of the Geek Pulse XFi, the position of the dual clock chips are just right behind the headphone output section. So the high frequency noise on the headphone output (none on the line output) could be from the clock chips. But that’s just my guess. The video shows that Larry has placed some slots on the PCB to create some degree of separation between the clock section to the headphone output section. So, please take the above measurement with a pinch of salt, as I mentioned earlier it is not a proper noise floor measurement, just showing what I have observed on my Owon VDS3102 oscilloscope (bandwidth set at 20 MHz).
https://www.youtube.com/watch?v=HIAqtblSmjQ
Low Level Signal & Noise
When using sensitive IEMs, typically my listening level is around 100 mV rms or less. At this low level signal, noise floor may start to become an issue when using sensitive IEMs. Therefore I include observation of low level signal. I played 100 Hz sine wave, 24-194 resolution, to all the DACs, set the headphone output to be at around 100 mV rms on 32 ohms, and take the oscilloscope screenshot of the headphone output.
Yulong DA8II, volume -36.0 dB:
Yulong DA8, volume -36.0 dB:
Mytek Stereo192-DSD, digital volume -33.0 dB:
Mytek Stereo192-DSD, analogue volume -33.0 dB:
Geek Pulse XFi, (High gain) volume -35.5 dB:
Geek Pulse XFi, (Medium gain) volume -28.0 dB:
Geek Pulse XFi, (Low gain) volume -24.0 dB:
Though Geek Pulse XFi low level signal looks noisy, somehow I didn’t hear the noise is audibly affecting the performance of the headphone output. It is something that I cannot explain. Most probably is because the noise spectrum is more on the ultra-high frequency that is beyond human hearing capability. But I do actually expect the XFi headphone output to be as clean as other DACs headphone output.
Power Output & Output Impedance
Power output is measured using 15 ohms load for maximum output current measurement, and 600 ohms load for maximum output voltage measurement. My simple digital oscilloscope doesn't do THD measurement, but it has FFT feature. Maximum output power criteria is maximum output at less than 1% THD. To estimate the maximum output voltage before the waveform get distorted, I visually monitor the waveform on the oscilloscope, and monitor the FFT window to check that the harmonic distortion is less than 40 dB (100 times) from the main frequency.
Headphone output with 15 ohms load, at volume -6.5 dB:
Headphone output with 15 ohms load, at volume -6.0 dB:
Maximum Output Voltage
6.8 Vrms at 600 ohms load
Maximum Output Current
3.2 Vrms at 15 ohms load
Max output current: 213 mA
Calculated maximum headphone power output:
1445 mW at 32 ohms
925 mW at 50 ohms
154 mW at 300 ohms
77 mW at 600 ohms
Headphone Amplifier Output Impedance: ≈0.26 ohms
Temperature
With class A headphone amplifier, the area around headphone output gets pretty warm after sometime, and I notice DA8II headphone output area is generally warmer than DA8. Just for comparison, operating in room temperature in Singapore (tropical country) which was around 29.5 degree Celsius during the time of measurement, I measured DA8II headphone socket was around 42.4 degree Celsius, while DA8 headphone socket was around 39.2 degree Celsius. Pretty warm, but not up to alarming hot. To place it in open area with some air circulation is recommended.
Channel Imbalance
Balance between Left and Right channels is very good on Yulong DA8II. Channel imbalance of headphone output around hearing level is very little to almost zero. And it is rises slightly only near maximum volume, but kept low below 0.2 dB. Channel imbalance measurement is using oscilloscope and 600 ohms load.
Brand / Model
| Vol. Position
| Hp Output (Vpp)
| Imbalanced
(dB)
| |
Left
| Right
| | | |
Yulong DA8II
| -60 dB
| 22.400 mV
| 22.800 mV
| -0.15
|
-50 dB
| 64.000 mV
| 64.800 mV
| -0.11
| |
-40 dB
| 0.196
| 0.196
| 0.00
| |
-30 dB
| 0.632
| 0.632
| 0.00
| |
-20 dB
| 1.960
| 1.960
| 0.00
| |
-10 dB
| 6.160
| 6.160
| 0.00
| |
-1.0 dB
| 17.200
| 17.400
| -0.10
| |
0 dB
| 19.200
| 19.600
| -0.18
| |
RMAA Test of Headphone Output
Please take note:
RMAA test is only as good as the quality of the audio interface used for the measurement. And in most cases, only useful for verification purpose of the audio quality within the 20Hz to 20 kHz range. For example, most audio interface line input only have linear frequency response up to around 20 kHz, or probably close to 50 kHz at higher sampling rate. So if I measure an amplifier with flat frequency response up to 100 kHz (which is common), RMAA test result using the 50 kHz audio interface will only shows frequency response up to 50 kHz. So in this case RMAA test result doesn't reflect the frequency response of the amplifier under test, but the frequency response of the line input of the audio interface. The same principle applies for the noise measurement. RMAA test result won’t be able to show noise measurement lower than the line input interface noise floor. For example, my HRT LineStreamer+ SNR is around 104 dB, so we can’t expect to see noise measurement result that is much lower than -104 dB (with a few dB margin) even if the line output of Yulong has 135 dB SNR. Once again, please take note, RMAA test is only for verification purpose that the audio equipment is working within the common reasonable specification, and not accurately reflecting the real specification of the equipment.
RMAA measurement in this review is done using RMAA 6.4.1 in 24 bit - 96 kHz, using HRT LineStreamer+ as the line input interface (ADC).
Specification of HRT LineStreamer+ :
Full Scale Sensitivity: 2.25 Volts RMS
Input Impedance: 5k Ohm
Frequency Response: (20 Hz/20 kHz) +0 / -.4 dB
S/N Ratio (DC to 30 kHz): 104 dB
THD+N (1 kHz Full Scale): .003%
Due to unknown reason, my Yulong DA8 RMAA result is showing abnormal level of distortion that in my opinion is not a valid test result to be shown. My DA8 so far is working fine, and showing good measurement result on oscilloscope, but somehow RMAA test failed for unknown reason. Therefore, I will only show RMAA result of Yulong DA8II, Mytek Stereo192-DSD, and LH Geek Pulse XFi.
RMAA 6.4.1 Test Result, comparison of:
Yulong DA8II Headphone Output
Mytek 192-dsd Headphone Output
LH Geek Pulse XFi Headphone Output
Summary
Test
| Yulong DA8II Headphone Output
| Mytek 192-dsd Headphone Output
| LH Geek Pulse XFi Headphone Output
|
Frequency response (from 40 Hz to 15 kHz), dB:
| +0.04, -0.13
| +0.01, -0.05
| -0.09, -0.21
|
Noise level, dB (A):
| -106.5
| -101.5
| -103.0
|
Dynamic range, dB (A):
| 106.5
| 101.4
| 102.7
|
THD, %:
| 0.0059
| 0.0051
| 0.0058
|
IMD + Noise, %:
| 0.0060
| 0.0060
| 0.0065
|
Stereo crosstalk, dB:
| -84.4
| -82.8
| -85.8
|
Noise level
Dynamic range
THD + Noise (at -3 dB FS)
