Holo Cyan 2 + Topping Centaurus Review & Measurements
Introduction
The Holo Cyan 2 is a recent release from Holo Audio that comes in at a substantially lower price ($1200) than the Holo May or Spring 3, whilst keeping the impressive objective performance that Holo is known for in the realm of R2R DACs. However, Topping has also recently released the Centaurus in collaboration with Holo Audio, a DAC using Holo’s R2R module from the Cyan 2, but available for $200 less; $999 at the time of writing. Given as these two DACs are using the same DAC module, I wanted to provide measurements for each of these devices in a single post so that they can be directly compared. And to ensure that results are a fair and direct comparison of each product, I have swapped the same DAC module between each unit to take measurements with the same DAC module itself to ensure that the only differences remaining will be differences between the two products themselves, rather than any unit variation that may be present in the R2R ladder.
If you’d like to watch the video reviews for each of these devices you can do so below!
Design and I/O
The Holo Cyan 2 and Topping Centaurus are similar in many ways besides sharing a common DAC module. Each devices offers balanced and single ended outputs, SPDIF, AES, I2S and USB inputs, though the Centaurus also adds bluetooth input, and a 12V trigger in and out which will allow you to integrate the device with other 12V trigger compatible devices and turn your system on/off with a single button.
Topping Centaurus rear IO
Holo Cyan 2 rear IO
More obvious differences are found on the front of each device. The Cyan 2 forgoes a display and controls, instead having a series of LED indicators for both input sample rate and the active input source, whereas the Centaurus features an LCD display and menu which allows the user to adjust various settings such as digital volume, oversampling, input selection, and also functions as an FFT visualiser whilst playing music.
The Cyan 2’s inputs are not manually switchable, instead utilizing an automatic system that switches the selected input depending on what it detects as actively supplying audio data. In my usage I found that this generally did work well for AES/SPDIF/I2S inputs, however if a USB source was used, the DAC would lock to this input and switching away from it required physically disconnecting the USB cable. This will be an annoyance for some users, although for anyone that tends to connect only a single source including myself, it’s not much of a concern.
Internally, both DACs feature the same Holo Audio R2R DAC module, which contains both an R2R ladder for decoding of PCM audio, and a discrete 1-bit DSD converter, allowing each DAC to convert and play both PCM and DSD audio natively.
Holo Audio R2R Module
Holo Cyan 2 Internal
Topping Centaurus Internal
Aside from the R2R module itself, each product has a fairly drastically different design. The Cyan 2 has a larger I/V and output stage than the centaurus in addition to a fairly substantial linear power supply. In contrast the Centaurus uses a switching power supply, and also has some additional components to facilitate internal oversampling, EQ, bluetooth connectivity and the main display.
I personally was quite excited upon hearing about the Centaurus, given as one of the major omissions from the Cyan 2 was internal oversampling; it is a NOS (non-oversampling) only DAC, and therefore users preferring the sound of oversampling, which avoids the upper treble rolloff induced by NOS, need to use external tools such as Foobar, Audirvana, Roon or HQPlayer to do that digital processing.
However, unfortunately it seems the Centaurus’ internal oversampling may not be the solution I had hoped, as will be shown in the measurements below.
So how do these two products stack up against each other and against other DACs?
Measurements
Measurements Include:
- THD+N (SINAD)
- THD+N vs Output Level
- THD+N vs Frequency
- IMD vs Output Level
- Low Level Signal Output
- Jitter
- Linearity
- Output Impedance
- Volume Matching
- Noise
- Intermodulation Distortion (IMD)
- Crosstalk
- Multitone
Test Setup
- Audio Precision APx555 B-Series analyzer
- Measurement setup and device under test are running on an AudioQuest Niagara 3000 power conditioner.
- DACs were connected via USB unless specified otherwise.
- DACs were warmed up for 6 hours prior to testing
- Unless otherwise specified, Centaurus was set to ‘Best OS’ mode (192khz upsampling)
- Unless otherwise specified, Cyan 2 was fed with 192khz data to provide direct comparison to the Centaurus
- Tests shown are with a 200kΩ input impedance configured on the analyzer.
- Exact analyzer/filter configurations for each measurement are detailed in the full reports
- CH1 (Blue) = Left, CH2 (Red) = Right
Results
Dynamic Range (AES17): 125.4dB (Centaurus), 126.4dB (Cyan 2)
SNR: 125.5dB (Centaurus), 127.1dB (Cyan 2)
IMD SMPTE: -96dB (Centaurus), -100dB (Cyan 2)
Noise Level RMS 20hz-20khz: 2.1uVrms (Centaurus), 2.1uVrms (Cyan 2)
DC Offset: 320uV (Centaurus), 1mV (Cyan 2)
Susceptible to intersample overs: Both DACs inherently immune in NOS mode, Centaurus is susceptible to intersample overs when using internal oversampling.
Output Impedance: 100 Ohm XLR/50 Ohm RCA (Centaurus), 40 Ohm XLR/20 Ohm RCA (Cyan 2)
THD+N / SINAD
Disclaimer: SINAD should not be used as a figure to judge the overall performance of a product, it is a figure that combines several aspects of performance into one number whilst not including many others. It should be used as one factor among many when evaluating the overall objective performance of a device, and the type/structure of distortion can often be more important than the level of distortion/noise itself.
1khz 0dBfs Sine - Holo Cyan 2:
Starting with the Cyan 2, 111dB for an R2R DAC is extremely impressive. Currently the majority of R2R DACs seem to get around -95 to -100dB, and it seems Holo has been able to keep most of their impressive performance even in a smaller, more affordable device. This puts the Cyan 2 as one of the lowest distortion R2R DACs on the market, beaten only by its big brothers the May and Spring 3, and the Schiit Yggdrasil MIB.
1khz 0dBfs Sine - Topping Centaurus (4V):
Centaurus performance is nearly identical to the Cyan 2, and again can be placed among the lowest distortion R2R DACs available at any price.
1khz 0dBfs Sine - Topping Centaurus (5V):
The Centaurus does have the option to be set to 5V output, however performance does seem to drop noticeably. I changed it back to 4V output for the remainder of the tests.
1khz 0dBfs Sine - Topping Centaurus (4V - NOS mode 44.1khz):
For the previous measurements, the DAC was in the default ‘Best OS’ mode, which sets the DAC to 192khz upsampling. Unfortunately, setting it to NOS mode seems to cause some fairly significant performance issues if you then feed it 44.1khz content (which is the vast majority of music).
The same behaviour does not occur if you feed it 48khz, and also does not occur on the Cyan 2 at all, so I’m not entirely sure what’s happening here, but this unfortunately does mean that the NOS mode on the Centaurus is effectively only properly usable for 48khz content. In addition, when in NOS mode and feeding the device 44.1khz content, the two channels were no longer exactly in phase:
I did test a second unit of the Centaurus to see if this was an issue confined to a single bad unit, but unfortunately both showed the same behaviour. I reached out to Topping for comment, but they responded saying that 44.1khz is expected to behave ‘slightly worse’ than 48khz. Whether or not you’d consider a 20dB drop in performance to quality as ‘slightly’ is up to you, but this does seem like a fairly notable issue that hopefully can be addressed. It is particularly unusual given as the same DAC module moved into the Cyan 2 behaves absolutely fine. And as a final verification I also tried using the 2nd DAC module as well (in case there was some hidden difference between the Cyan 2 and Centaurus variant) but the issues persisted.
1khz 0dBfs Sine - Topping Centaurus (4V - NOS mode 192khz input):
Given as one reason a fairly large number of people seek out NOS R2R DACs is to be able to take full advantage of external high performance upsampling tools like PGGB and HQPlayer, and in fact is also why the Holo DACs support 1.536Mhz PCM and DSD1024 input, I was curious to see whether using HQPlayer or other oversampling tools would work well with the centaurus. Unfortunately it seems to have some even more substantial issues than what we saw with 44.1khz input, and so this was pretty much a non-option.
Hopefully this is simply a software issue and Topping may be able to address it with a firmware update at a later date.
1khz 0dBfs Sine - Topping Centaurus (4V - 16x OS mode):
Unfortunately again it seems even using the Centaurus’ internal oversampling, which can be set to 2x, 4x, 8x or 16x, does not solve the high sample rate issue. When set to 2x or 4x, performance was usually ok (although I saw occasional issues at 4x, despite the fact that ‘best OS’ mode which also resamples to 4x rates seems to behave fine), but at 8x or 16x, performance once again dropped substantially, and this time there were some unusual jitter/phase noise products visible around the bottom of the fundamental 1khz tone.
1khz -0.5dBfs Sine - Holo Cyan 2 DSD:
The Cyan 2’s native DSD converter means you effectively have two DACs in one box. And it seems the performance of the DSD converter is rather excellent as well, with -116dB THD+N. The quality of the end result with DSD is heavily dependent on the quality of the modulator used to create the DSD stream, but HQPlayer’s advanced modulators like ASDM5EC-Super are extremely high performance, and PGGB is also an excellent option that now has a free version available.
THD+N vs Output Level
Cyan 2
Centaurus
THD+N vs Output level on both DACs is effectively identical, as is somewhat expected given the DAC itself is likely the limiting factor here.
THD+N vs Frequency
Cyan 2
Centaurus
THD+N vs frequency is overall very flat on both devices. The Cyan 2 seems to show slightly better results overall, but the profile of distortion vs frequency is effectively the same for both products.
Using NOS mode at 44.1khz/48khz sample rates will cause a rise in THD at high frequencies, but this is just due to the inherent nature of NOS not filtering out ultrasonic aliased products from the output, and is not a hardware limitation.
Frequency Response
Frequency response behaviour was identical for both products, and was primarily dependent on sample rate. When running NOS, both DACs have a roughly -3dB drop by 20khz, which is the mathematically expected result for any NOS DAC. At higher sample rates, the frequency response is much closer to flat.
IMD vs Output Level
Cyan 2
Centaurus
IMD does change a fair bit depending on output level, as is typical with most R2R DACs. Both DACs again showed nearly identical behaviour with the same DAC module, with the exception of the very upper areas where the cyan 2 shows a little more rise, though this is mostly due to the topping being set to 4V and thus the graph is shifted right slightly, cutting off the rise on the right hand side. When the Centaurus was set to 5V, the result was identical.
Low Level Signal Output (96khz bandwidth)
This test plays a very low level signal through the device to check for any unusual behaviour.
-90.31dBfs 24 bit dithered sine - Cyan 2
-90.31dBfs 16 bit undithered sine - Cyan 2
-90.31dBfs 16 bit dithered sine - Cyan 2
I’m just showing the Cyan 2 here as neither DAC showed anything unexpected, and looking at these tests in NOS without the Centaurus’ ‘best OS’ mode dithering gives us a bit clearer a view into the behaviour. For 16 bit undithered signals, we can clearly see the three levels with no overshoot, and no imbalance in level between positive and negative. Everything is behaving nicely and both DACs handle small signals quite well.
Jitter
Cyan 2
Jitter performance on the Cyan 2 is effectively perfect. Combining this with the built in USB galvanic isolation, it means you can just hook the DAC up directly to your PC, never have to worry about noise, and you’ll be getting perfect jitter performance. You do not need a DDC or Isolator for this DAC, and in fact a DDC will in most cases not be able to match this performance.
Centaurus
The Centaurus still has quite good jitter performance, though some deterministic jitter products are visible.
There are however some other interesting aspects of jitter performance for each product. The Cyan 2 seemingly does not use a PLL for its non-USB inputs, meaning it is running directly off the clock signal you feed to the DAC over the SPDIF/AES/I2S connection. As a result, the quality of the DDC/Streamer/Source you use will directly impact the resulting jitter performance of the DAC, any jitter on the incoming clock signal will just be passed through, but a high quality low jitter source will yield good performance. However again, given the excellent USB implementation and inherent clocking, USB is probably the best option for most as there’s no need to spend money on a DDC if you already have isolation and basically perfect clocking. The Centaurus’ jitter was seemingly very dependent on the configuration. The above result was in ‘best OS’ mode, but setting the internal oversampling to 8x or 16x caused jitter to increase substantially, and setting it to ‘NOS’ mode resulted in…..well I’m not sure how to describe this:
Bandpass Filtered Linearity
This test plays a 1khz tone through the device, and steps down the level in small increments. Linearity measures how accurately the actual output of the device reflects the intended output. For example, if the level was stepped down by 20dB, but the output of the device was 19.7dB quieter, there is a 0.3dB nonlinearity, which would show as a 0.3dB rise on the graph.
Cyan 2
Centaurus
Both the Centaurus and Cyan 2 had excellent linearity performance. Interestingly the Cyan 2 did seem to do slightly better here. Given as R2R DACs are extremely reliant on a quality, low impedance PSU (hence why many high end R2R DACs have what seem to be overkill PSU designs), this could be the result of the difference in PSU design, though that is just my personal speculation.
Noise
Cyan 2
Centaurus
Noise levels are overall very low. Both products are exceptionally quiet, and also do a fantastic job of avoiding any mains leakage from the power supply.
Crosstalk
Cyan 2
Centaurus
Crosstalk performance is quite good, remaining below -120dB on both devices at all frequencies. The Cyan 2 slightly outperforms the Centaurus here, keeping a few dB lower up to around 1khz, and around 20dB lower at high frequencies.
Multitone
Cyan 2
Centaurus
The multitone results are once again, extremely impressive for an R2R design in both instances here. I was a little surprised to see the Cyan 2 getting 6dB or so better performance particularly at low frequencies, this is possibly down to the I/V stage design?
Conclusion
The Cyan 2 and Centaurus are two of the best performing R2R DACs available at any price, beaten only by the Holo May, Holo Spring 3, and Schiit Yggdrasil MIB. However both come with caveats. The Cyan 2 is NOS only, and so for users wanting to use oversampling, they’ll need to do so with an external solution. And the Centaurus whilst being nearly identical to the Cyan 2 in overall performance for a couple hundred dollars less, unfortunately suffers from some major issues if you do anything but keep it in ‘best OS’ mode, preventing you from being able to use NOS or higher oversampling rates without significant performance penalties.
As discussed in my subjective review, the Cyan 2 is also extraordinarily good from a subjective standpoint, and I am happy to recommend it as one of the best options anywhere close to this pricepoint, with the note that you may need to be willing to enable oversampling in your music player if the more laid back sound of non-oversampling is not your preference.
The Centaurus unfortunately seems to be a somewhat unfinished product, and for the time being, I would not recommend purchasing one. However I hope that the current issues can be resolved with a future firmware update, and if so, it will also become an extremely high performance option for the money.
