Topping Centaurus R2R DAC Review

[fatoldgit]

It's controlled by an FPGA

so is that saying the FPGA is in the signal path (more than say just changing relays to reroute a signal)?

Or is the issue because the FPGA change caused a selection of what what would normally be an illegal combination of sub-systems?

There is no mention of PEQ in the review so I could understand that PEQ is done via software and so a change there could be a vector for unintended consequences.

Just checked and it does have PEQ so that will be done in software and be a vector for unintended consequences.

 

[fatoldgit]

Look I dont know where you got all that from my very brief posts. This sounds like something more you can ask Topping themselves or experts in the field I'm no engineer just a guy who like Audio gear. I'm just putting out what I have experienced myself by swapping back firmware.
There is some kind of logic obviously controlling the meters and settings.

Sure but it leads to a hopefully interesting discussion about how [some] DAC's are "glued" together by software and how that might lead to issues if testing is inadequate after a software change.

Peter

 

[NTTY]

Explain this to me.

This isn't an FPGA DAC... so the basic DAC function is hard set (in this case resistors).

And the mentioned error was when in NOS mode noting that obviously (?) switching between NOS and non-NOS also toggles in/out another hard set bit of logic, baked into silicon that cant be changed via a software update.

So how do two "set in silicon" functions get screwed by a software update?

Or is the NOS/no-NOS mode written in software?

And whether it is or isnt, where was the QA testing before they released the new update? The DAC has a limited set of functions so it isnt hard to cycle thru the various combinations on the test bench.

Peter

It seems you mixed the messages of @keenerz and @binaryblaze who experienced two different issues.

And you also seem confused about the logic of the conversion. Digital (oversampling) filtering happens before conversion, in the FGPA. So the logic can be changed here. The final process of converting digital words into voltage happens in the R2R network in all cases.

 

[SMen]

Some people associate digital's ability to produce flat treble response as a "harshness." The bit of roll off may help these people.

It has been a long time since I heard a DAC sound harsh. I sometimes wonder if memories from the 'dawn of digital time' don't colour these viewpoints - a memory bias kind of thing.

Got the point, like saying it rolls of just like the good old vinyl. But it is obvious that this is not an argument made from a high fidelity standpoint.

Vinyl replay was often quite the opposite of roll off ... there were (and are often still) rising responses above 12 khz, moving coil or magnet. The whole low tip mass in higher end replay (MM especially) was to get these peaks above 20khz and beyond ... I love my old boron tube technics ... fortunately the Jico SAS peak doesn't much trouble me ... I'm still good to just below where it starts to be a problem... maybe I'll use the PEQ to get a nice 16khz bump

 

[oleg87]

so is that saying the FPGA is in the signal path (more than say just changing relays to reroute a signal)?

Or is the issue because the FPGA change caused a selection of what what would normally be an illegal combination of sub-systems?

There is no mention of PEQ in the review so I could understand that PEQ is done via software and so a change there could be a vector for unintended consequences.

Just checked and it does have PEQ so that will be done in software and be a vector for unintended consequences.

The Holo DAC module Topping is using has some Altera CPLD on it. I assume this is dedicated to handling the actual DAC business, and Topping has some other processor handling USB/PEQ/system control duties.

 

[Sokel]

The Holo DAC module Topping is using has some Altera CPLD on it. I assume this is dedicated to handling the actual DAC business, and Topping has some other processor handling USB/PEQ/system control duties.

PEQ at the first Topping implementation is done in XMOS,don't know about this one but it seems the same.

 

[binaryblaze]

Oops. Either Firmware issue or test files. Can you try with REW’s generator?
And no, it’s not clipping else odd harmonics would be higher than the rest, it’s high level harmonic distortion.
With -0.1dBFS, you’re nearly back to what it was, so I’d really suspect something wrong in digital domain, either with the interpolator of the Centaurus (after firmware update) or when creating the test file at that level.
Alternatively to Adobe, you can create test tones with Audacity’s generator.

The same thing happens at v0.14 as well

 

[boxerfan88]

The 1kHz spray of harmonics seem similar to the measurement charts of Soekris dac1541. The low level spray of harmonics seems to be inherent in R2R designs.

Having said that, this Topping Centaurus noise floor & harmonic spray is at a much lower level compared to dac1541.

 

[raif71]

 

[BeeKay]

Typical YT reviewer audio buzzword bingo. Will we ever get rid of these guys and those keeping them afloat by patreon?

 

[binaryblaze]

Are there any honest, objectivist reviewers around? There's Julian Krause but he mostly reviews audio interfaces

 

[JSmith]

Are there any honest, objectivist reviewers around?

In effect, there can't be by definition really... as they are just listening to things without controls or any kind of proper comparison. Even if they believe they're being honest, they're really not being so.


JSmith

 

[Mnyb]

In effect, there can't be by definition really... as they are just listening to things without controls or any kind of proper comparison. Even if they believe they're being honest, they're really not being so.


JSmith

Some few are honest but what they say is simply not information , they are not even wrong ? due the reasons you say, no controls no blinding + undefined jargon that only has meaning to the person uttering it "buttery midrange" etc

 

[SMen]

Are there any honest, objectivist reviewers around? There's Julian Krause but he mostly reviews audio interfaces

Archimago does extensive tests, and supports them with subjective listening tests that any appreciator of music would appreciate.

I find some of these quite useful ... but only after ignoring the hyperbole, and preferably with the balance of a proper objective review such as we find here. I'm happy with my purchases this past year with this balance.

Erin's Audio Corner is another one. There are a few.

 

[amirm]

There are a few.

We are seriously outnumbered compared to these fluff, imagination-only youtubers....

 

[raif71]

For me, it doesn't have to be we vs them... just a difference of opinion / POV

 

[NoteMakoti]

Indeed.
Pushing a delta-sigma DAC to new levels of even more inaudible purity is a celebration of engineering and technical prowess.
Doing the same with R2R DACs is just audiophoolery.
Makes perfect sense to me ...

I'm of two minds here; I like to see really overengineered pointless DACs like the ones Topping and SMSL make in the high-end, but I wish these reviews focused more on features nowadays. Features sell a DAC way more than sound quality nowadays; it's one of the reasons manufacturers are even pushing R2R.

 

[SMen]

We are seriously outnumbered compared to these fluff, imagination-only youtubers....

Yes, it's interesting. I really think you are cutting through though, despite being outnumbered. And I'd say encourage the few ...

I've even heard some of them:

• matching volume levels (or trying to) ...
• acknowledging bias
• admitting sometimes: "heck, I can't hear any difference at all between these"!

but I do find myself having to find out for myself, with my best efforts at controlled listening tests, and within my pretty small budget.

You got me here with your Aiyima A07 review many moons ago, because it affirmed my surprise at how it sounded so good ... 'santé' as they say over here in France, your good health.

 

[BeeKay]

Features sell a DAC way more than sound quality nowadays; it's one of the reasons manufacturers are even pushing R2R.

This fits more or less the Vinyl Renaissance. The safe, warm feeling of some tangible technology perhaps? Even if R2R and vinyl do not represent the state of what is technically feasible, a trend is recognizable among some buyers. Why not? As long as it is advertised correctly without slipping into the Snake-Oil territory, I agree.

 

[SMen]

Quite a lot of people here seem to respect this product, I certainly do. It's impressive!

But to answer your questions
"...If a resistor network is capable of passively decoding a 16/44 signal..." "...and the rest of the product is competently designed..."

The challenge is a result of requiring perfectly precise resistors. Such things don't exist. But let's say you find a set of perfect resistors. They'll only be perfect today, by tomorrow they will be slightly less perfect. In a year's time, even less perfect. If your room is cooler in the winter, they will be less perfect; if it's hotter in the summer, they will be less perfect.
This is why manufacturers moved away from resistor-based dividers - too difficult, to do well, consistently over time and temperature, economically.

I'm curious about the degree of imperfection.
Are these changes or fluctuations going to impact on frequency response over time?

My speakers sure seem to fluctuate quite a lot between summer and winter. 30+ degrees centigrade in summer, and the bass seems much fuller, sound warmer. But I don't think that is a temperature bias in my head because it is indeed hotter!

Or is the measured performance of this unit such, that these changes in perfection will or should remain below audible: is the engineering here as good as it gets currently? Good enough to make these imperfections over time small enough not to worry about? I survive with my speaker fluctuations.