Topping Centaurus R2R DAC Review

[peng]

As a side note,disregarding Centaurus,what differences one expects comparing a good 118dB SINAD Holo May R2R with a 120dB SINAD whatever?
I could design a scenario where playing levels along with recorded levels can bring some DAC's flaws to audible regions with special signals but that's it.
Good luck discriminating them with music.

For 118 dB vs 120 dB, I would have to bet no as well. People might say that's just one metric, and it is only for one point, such as tested with a 1 kHz signal so at lower/higher frequencies the differences may be much more significant, but 118 -120 dB is clearly well below the threshold of audibility so while it may be much worse in the full band, it would till be low enough for people with no hearing loss and ears trained.

 

[Mnyb]

My point was impulse response might be the advantage that NOS fans were clinging to, considering that all the other measurements are worse.

But it’s not really an advantage it’s just looks as advantage if you don’t understand these measurements but yes I think we align here .

 

[Loomynarty]

Topping must have gone to great pains in building these, precision of the parts in the resistor network is important to performance of R2R dacs. But, why bother? One can get a "conventional" DAC with even greater linearity for much less money.

I would love to see results of a well-designed double blind listening test to see if ANYONE can actually hear differences between the various high-quality DACs that are available these days. And I wonder if the only DACs which people could discern as sounding different might be the ones with inferior measured performance - i.e., Audio-GD DACs etc. They might introduce enough artifact that some people could hear them as being different, but I wonder if even inferior DACs sound different enough for humans to be able to discriminate. We might not even be able to hear the difference between one of the low-achieving DACs on Amir's ranking and the top-ranking ones.

This might be of interest to you.

 

[SMen]

Where is the f…ing difference between the Topping D90 III Discrete (aviable scince August 2024) and this device?

This what I find interesting with regard to what Topping are doing here... taking these different converter technologies, and doing their best to make 'statements' engineering wise for each. There is also the Sabre D90 which knocked out all others at the time measurement wise if I remember correctly.

Why do it? I suppose you could say for the money, but I think it's for the engineering challenge - to make statements ... and they are competing for kudos in the chinese market never mind other markets.

On the amplifier front they brought out the B100 which is class B as far as I understand. What other 'mainstream' producer brings out a class B for audiophiles?

Class AB, class B, class D ... spot the difference between those in your blind tests ...

Sabre, 1 bit discrete, R2R ... spot the difference between those in your blind tests ...

after we have engineered them ... Kind of cool really.

the roll off might be audible mind you ... but (and I am rambling on here beyond any reply) ... just think of how many modern speakers might actually benefit from this!

Way you go Topping I say...

 

[SoundsGood2Me]

Thanks Amirm!

And Thank You Topping for doing us Audiophools proper with so many bleeding edge products! I love seeing such killer measurements time after time, and then you throw affordability in and why would you go anywhere else? Welp except for power amps, I wish they'd build something capable of 500W like my Benchmark HAB2s bootstrapped mono.

I really appreciate how well organized all your reviews are and every device is measured identically for easy comparison. I've read Stereophile for 40+ years and while I appreciate the reviews I really don't trust a pair of rented ears discussing Plankton and Harmonic Holism when John Atkinson's well thought out measurements are right there telling the truth. What is even the point of any audio mag that doesn't actually measure the DUT? It's cracked me up the several times that a reviewer (who gets to keep the gear PLUS a paycheck) wrote glowing praise for something that later measured like garbage.

It's a true blessing to be in the Best Times for Audio, EVER. Now if only we could get some modern replacements for Zeppelin and Floyd and Yes recorded on -120dB S/N systems! Pic not related

 

[VVR]

Amir, can we see the multitone graph as well?

 

[MaxwellsEq]

R2R NOS DACs were never on my radar before, they were just a curiosity until the Centaurus came out. I started reading a few reviews and measurements and got curious.
The latest fave rave seems to be the Holo Audio May, which throws digital audio theory out with the bathwater in NOS mode, but many love it.

I'm struggling to understand what they're trying to do, and three things are often wrapped together, which doesn't seem right to me. Am I misunderstanding?

1. R2R DACs use a switched ladder network of filters to implement the conversion, which is an old, simple and expensive implementation.
2. Most DACs use digital filters to perform over-sampling, but NOS DACs don't do this, they convert at the native rate. This avoids pre-ringing and post-ringing.
3. NOS DACs seem to necessarily avoid using the low pass reconstruction filter, or at least one that only filters 44.1kHz

These three things are often considered collectively, but I'm sure they're separate.

1. R2R DACs are just a different architecture for performing the DA conversion. It's not very linear, and maybe people are just nostalgic? I don't care.
2. Digital filters have different frequency and time domain responses, depending on whether they're linear or minimum phase, fast or slow. The ringing is usually at 22kHz and shouldn't matter, but some people have preferences for one or another. I think NOS DACs don't use digital filters (or they're just switchable) and they don't have ringing, just lots of HF distortion instead.
3. NOS DACs don't use reconstruction filters, you just get an un-smoothed analogue output and live with that, hoping that the steps are inaudible.

John Atkinson sometimes measures the time domain effect of the different filters. Here are the Holo May in OS and NOS modes, and the MBL N31 with slow filter respectively:
View attachment 413929View attachment 413927View attachment 413928
The NOS time domain response to an impulse is just that - an impulse, without any pre- or post-ringing from the DAC (though the frequency domain response is compromised).
My point is that the NOS impulse response is really very similar to the MBL slow filter response above - not the same but very similar.
Is that where the benefit in a NOS DAC comes from? But can't you achieve much the same thing with a conventional DAC and a slow filter instead?

Unfortunately, most what you wrote is incorrect because you are confusing decoding and filtering by mixing them together.

Let's start with reconstruction filters. You MUST have them if you have any hope of reproducing the original recording. This is a baked-in consequence of digital sampling theory as invented in the middle of the last century by Whittaker, Nyquist and Shannon. It does EXACTLY the same thing whether you have a current SOTA DeltaSigma, an original Phillips 4x oversampler, a ladder or R2R DAC. The filter's job is to allow everything through without any impact for up to half the source sampling rate and then to plunge precipitously down to block everything above half the sampling frequency. Sure, with certain test signals, filters ring, but good ones don't when playing music.

Then let's look at the DAC itself. Assume we have a mathematical system where 1 = the quietest breath of a gentle breeze and where 256 = the loudest bit of an explosion. This can be encoded in an 8-bit digital word. When all the bits are 11111111 = 256; when all the bits are 00000000 =1. The challenge is converting these bits into a voltage which can make speakers move. Suppose we set the first bit flipping on to be 0.01V, when the second bit turns on, it has be be double that at 0.02V. The third bit flipping on is double again - 0.04V. This 2x mathematics with each bit needs to be extraordinarily accurate. The simplest architectures to understand are ladders and R2R DACs where each step is done with precision resistors. It's easy to understand, but spectacularly hard to get right. And if you do get it right once, it will drift with temperature and with time.

The result is - even the very best R2R and Ladder DACs struggle to get close to the near-perfection achieved by DeltaSigma. Where they are at their worst is not at the very quiet or very loud part, but slap bang where the music is. They add distortion. It's probably not possible to hear this distortion, however. BUT this has NOTHING to do with filters.

 

[totti1965]

Actually for R2R the holo audio spring and may are quite good dacs (using spring 2 for years since it's release and no complaint)

Even the new Holo Audio Cyan 2 seems to be an extremely good R2R DAC.

 

[phoenixdogfan]

Thanks for this review, Amir. I'm going to vote "Not Terrible" because it costs more for being an RTR DAC than a comparable Sabre unit, AND it has measurably inferior performance. However, the decrement in performance is probably inaudible if set properly, but the big problem is there exist settings on the unit, which if employed, could actually lead to damaged speakers. So it's a dangerous toy that should not be put in the hands of newbies or children. I could probably have voted "Fine" if not for the possibility of a blown tweeter for the naive user.

 

[VientoB]

I blame YouTube reviewers for all these R2R DACs.

 

[MaxwellsEq]

you know that today with devices of this kind you lose practically nothing on an objective level

That's not the case. Look very hard at the intermodulation chart. It's good at 'silent' and it's good at 'so-loud-your-ears-bleed'. It's poorer where the bulk of music is. This is common with ladder and R2R DACs.

The image below makes it look artificially better for extremes than it is for normal music.

Meanwhile this image below explains why. Notice the top-left and bottom-right of the chart are good (which is why the figures in the above image are good), but the music happens in the middle of this slope which should be flat. This is where distortion modifies the music:

A good DAC will measure well at -20dBFS, -30dBFS, -40dBFS and have a smooth slope. This DAC would show poorer results than a DeltaSigma (even an ESS with the IMD hump).

 

[amirm]

Amir, can we see the multitone graph as well?

Oops. Just added to the review:

 

[amirm]

What about testing the PEQ filter(s) performance?

Filters modify the input signal and with it, change the nature of measurements generating results that are not meaningful. I can arrange where they don't impact the measurement which is what I did with the Topping D50 III: https://www.audiosciencereview.com/...ng-d50-iii-balanced-dac-with-eq-review.53856/

 

[Chillazar]

Topping is weird. They announced the coming of the A and D900 so long ago, released some other products that were announced later and now they bring this without any announcement. Also, I think they should have given this another design. This looks way too much like a tool instead of the usual audiophile jewelry that people who buy R2R expect. SMSLs R2R DAC may have worse performance but it looks better and I can understand people who buy it for that reason.

 

[totti1965]

Would be interesting to see this against the FiiO K11 R2R ( which I think is about the cheapest R2R implementation ) - though FiiO own rated THD+N for the K11 R2R was 0.025%

According to Sengpiel this is - 72.04 dB

Klirrfaktor Klirrdaempfung Dezibel dB % Prozent berechnen Klirrfaktor Audio THD TDH+N Klirr Faktor Signal Verzerrung Verzerrungen umrechnen k2 k3 Dezibel dB Prozent % - sengpielaudio Sengpiel Berlin

Klirrfaktor Prozent % Dezibel dB Klirrdämpfung berechnen umrechnen Audio THD TDH+N Prozent Klirr Verzerrungen Signal Rechner Berechnung k2 k3 k4 k5 Dezibel - Eberhard Sengpiel sengpielaudio

sengpielaudio.com

 

[VientoB]

Topping is weird. They announced the coming of the A and D900 so long ago, released some other products that were announced later and now they bring this without any announcement. Also, I think they should have given this another design. This looks way too much like a tool instead of the usual audiophile jewelry that people who buy R2R expect. SMSLs R2R DAC may have worse performance but it looks better and I can understand people who buy it for that reason.

Yeah they should have borrowed some copper buttons from Holo Audio as well. The YouTube reviewers would be all over it.

 

[AndreaT]

One hell of an academic project for Topping. Inferior measurements to current SOTA, excellent results for an R2R DAC. Now I look forward to a magnetodynamic MM pick-up needle for the micro groove LPs and maybe even moving coil…Not for me, thanks! I am still basking in the glow of my Okto Stereo DAC!

 

[totti1965]

If only holo audio cyan 2 get measured too...

SINAD of the Holo Audio Cyan 2 is about - 106 dB according to Topping.
But yes, IM-Distortion and Distortions at - 20 db or - 30 dB would be also very very interesting.

The first Generation of the Cyan was already tested long ago. It performed with SINAD - 101.5 dB.
@amirm was quite happy!

Review and Measurements of Holo Audio Cyan DSD DAC

This is a review and detailed measurements of Kitsune Hifi's Holo Audio Cyan DAC with DSD option. It is on kind loan from a member. The Cyan comes either in PCM R2R with headphone amp or just DSD (you can also get both). The cost for the DSD version is USD $1099. I am assuming you have to...

www.audiosciencereview.com

His remarkings in March 2019 were something like: „If it must be R2R, then go with a device that measures well - like the Holo Audio Cyan.“

 

[totti1965]

Holo Audio Cyan (March 2019) vs Topping Centaurus R2R DAC - and the winner is:

But on the other hand - The 192 kHz Sampling Frequency noise flor and IM Distortions of the 2019 Holo Audio Cyan is on par with that of the Topping Centaurus!!!

 

[staticV3]

But on the other hand - The 192 kHz Sampling Frequency noise flor and IM Distortions of the 2019 Holo Audio Cyan is 20 dB lower than that of the Topping Centaurus!!!

It's not, look again