I don’t know what technical/analytical market is.
RME makes the best DAC with built in equalizer for the money… its a bargain if you understand what it does.
Patents???
RME – these and similar products follow from the studio production market and seek neutrality and accuracy over comfort or “listenability.” Many user-oriented audiophile brands and every tube DAC rolls back precision in favor of other qualities. Some people find pure technical performance to be too cold, analytical, and fatiguing.
You might read the Chord threads such as this to learn about their technology:
Obviously you have Chords and like to throw out names and terms like XLR…OOOOOOh.
Your reply is only an opinion of 2 Chord fans.
1 Like
WHAT? I HATE Chord products!
We have a rule here about tone and how you conduct yourself.
You’ve been here 8 hours and are already on the wrong side of that.
If you want to be antagonistic this is the wrong place for you.
Cut it out.
8 Likes
The Chinese have recently stepped up their game on their higher end offerings. Case in point: The Topping D900 is reference level DAC that more than holds it’s own against other Delta/Sigma DAC’s such as the Chord Qutest/TT2.
Having previously owned the Chord DAC’s I actually prefer the D900 performance and sonic signature. Until recently, would not have thought it possible.
2 Likes
Thank you Ian.
I was about to pipe up similarly. I’d never accuse the pink frog of liking Chord or not fully understanding what he’s hearing. That way is danger.
It was staring to sound like the Wizard of Oz in here. Sn(ears) from the Woodsman and certain other aspects from the Scarecrow.
1 Like
Other than the Schiit, which other dacs DO you love?
To my ears, DACs mainly fall on a continuum between precise and prone to edginess versus relaxed and prone to excessive smoothing. Plus, the cheaper ones (i.e., budget ESS, Cirrus Logic, or AKM delta sigma implementations; budget iFi) more often than not have piercing treble, roughness, flatness, hollowness or perceived nonlinearity, or other artifacts. Some expensive ones also have artifacts too. See @donjoe0 above for some of the common explanations.
Once you get past the budget tier, my main reaction is whether DACs push precision or smoothing. Smooth DACs do take away harshness, but often make all music sound inappropriately calm and similar. I’ve demoed Chord, dCS, Holo Audio, Border Patrol, LTA, the Mola Mola Tambaqui, etc. There are often TINY differences between mid and high priced DACs, and the differences may be made irrelevant by other factors. (I don’t love all Schiit DACs, nor have I heard every Schiit model.)
While I routinely don’t like Chord and dCS (both always sound too smooth), the question of “DACs you love” comes down to system synergy. Will you be running the Focal Utopia or KEF speakers where excess treble is a risk? Use a more relaxed DAC. Will you be running with a tube amp? Then, probably don’t bother with a tube DAC. Are you running an overdriven tube amp that colors everything? I’d stick with precision. Is this a speaker setup where the room is dead and where more definition is desired? If so, try something with precision. Is this an odd shaped room or are you running “encompassing” speakers without a great soundstage? Well, small differences between DACs may not be hearable at all. Etc.
4 Likes
You bring up valid points, especially with synergy. I’ve had gear/headphones that received excellent reviews, yet the sonic result was lees than satisfactory when connected together as a chain.
I would argue that it’s primarily the FILTERING involved with the DAC chip that makes or breaks the sound. My Canever ZeroUno has a 9018 chip, but sounds nothing like any other ESS chip. Mario wrote his own code, uses a somewhat unique I/V conversion, has a special clock and power supply setup, Lundhal transformers, etc. To me, the DAC gets the sound right, in that it has the tonal accuracy of Delta/Sigma, whilst making the sound very organic, extolling the traits people associate with R2R. Definitely an exception to the norm set of assumptions.
Personally, NOS DACs just don’t sound right to me. I know others love NOS R2R. We all hear and focus on different aspects of sound reproduction.
No special point to this … just some observations …
If the transformers are in the output stage, they’re probably responsible for more of the “sound” of the thing than any other significant element.
Filters, unless incompetently done, or using a completely different filtering model (e.g., linear phase vs. minimum phase), tend to have most of their deltas in the region above 19 kHz … where there is very little musical information in the first place (not that most people could hear it, even if there was). Though, of course, where the filter ringing occurs, and its nature, has an effect at all frequencies.
True NOS DACs will have FR roll-off (typically ~3-4 dB within the upper audible band) and aliasing, which contributes to a lot of the perceived differences.
And then I think the Topping D900 (from an earlier post) is a 1-bit DAC. The first thing it does with any incoming signal is to convert it to DSD256 (an informationally lossless, but not bit-perfect conversion), and then applies enormous amounts of noise-shaping to a) recover the >99% of amplitude information that the DSD conversion throws away and b) eliminate the huge amounts of ultrasonic noise that is generated as a result.
In theory, even if not filtered out, since that noise is ultrasonic it shouldn’t be directly audible. In practice, it can cause all sorts of (potentially damaging) issues in downstream electronics and transducers. So it is important that the DAC filters that properly; and some simply don’t.
4 Likes
Appreciate the feedback. Frankly, not sure what Mario has coded his filtering algorithms to work with the unit. He keeps that close to the vest. I do know he developed his own MQA filters, which are not of much use unless I’m listening to Artie Shaw and his orchestra on TIDAL, which still shows MQA applied, along with some other legacy recordings. 
I would reckon, based on feedback from Rob Watts on his filters, that the entire frequency spectrum would be impacted. Regarding the output transformers from the DAC, which are prior to the output stage, that would likely indeed have an impact on the sonics. The output design is what he calls super linear, where both even and odd harmonics are suppressed. A neat trick not normally seen in SS output stages.
As for the D900, I would argue that Topping got it right. I’m using it with a Bryston BHA-1, and the sound is excellent. The DAC has brought new life to headphones, making them sound better than before. It’s several steps up from my old DACs, ( PS Audio DSD MK1, or the RME ADI-2). It’s a DAC that seems to have both the linearity of Delta/Sigma and the open dynamics of the best R2R DAC’s without the shortfalls of either topology.
The filters operate across the entire spectrum.
If you look at the outputs of such filters, be it purely as a mathematical function (where you can see the theoretical behavior of a perfect implementation - which is where Rob says he derives his measurements from) or measure the actual electrical output of a practical implementation (where real world concerns apply), you’ll see that the bulk of the differences between filters are at higher frequencies.
The remainder will be in impulse response which has more to do with the nature of the filter than the specific implementation of that type of filter.
The how of that might be special; but at the end of the day effectively suppressing harmonic distortion (odd and even) is not really a “neat trick” … it’s belt-and-braces stuff - there’s a litany of budget DACs (~$100 or so) that do this down to -116 dB or better (0.00015% THD).
Never heard the thing, but I am always skeptical about “heroic measures” being taken in speciality designs - especially where more than half the time (not necessarily here) those measures are only required because of iffy design decisions elsewhere in the unit.
They might well have. Their technical implementations tend to measure very well, which means it’s an engineering/objectively led design.
It’s just that, for me, going through all the trouble to convert from PCM to 1-bit DSD - losing the principal benefit to DSD in the process, never made much sense. It’s all extra work and expense for no specific benefit. Which doesn’t mean it can’t sound good (though every DAC I’ve heard that does conversion to DSD as part of its core D/A has been a little transiently softened vs. the better multi-bit and D/S converters (assuming, say, a 5-bit modulator implementation). At least when playing non-native DSD source material.
But, again, never heard the thing, and likely never will … I just find, more and more, that some of the superlatives and “marketing” claims that get made for a lot of this stuff don’t make sense.
The D900, for example, performs objectively worse than Topping’s own D50 Mk3 … but costs 8x more. Again, doesn’t necessarily say anything about how their output is perceived, just interesting from a marketing claims vs. actual performance vs. cost perspective.
2 Likes
I should add/clarify … I’m just talking about design and implementation ideas, models, theories and so on. I’m not specifically commenting on anyones choice of DAC, their perceptions of it, preferences or anything else.
Just the tech, approach, theory and results (objective - since by definition subjective, i.e., listening, results vary based on many other factors).
1 Like
Mend-It-Mark got attacked by the designer of this 25,000 pound pre-amp…video taken down…big controversy…check out the flimsy interior construction around 7:00…
1 Like
I’ll have to watch that properly … but a quick scan brought back memories of the first time I saw inside a Lampizator DAC, and later a Wells Audio amp. Now, I’m not drawing comparisons on the relative performance of any of them … just the first impression I got (which was not positive) on seeing their respective guts.
I think the Taiko Audio servers are my favorite examples of massively heroic implementations (always strikes me as being $15-20K in casework and PSU, infested with a tweaky PC) - that didn’t result in any detectable difference in data at the input to, nor analog output of the DAC, nor any bit-level delta or data difference via the network.
Compared to a baseline macMini … the realized performance at the DAC’s output was unchanged.
2 Likes
If you look at the white paper, one can see that the unit is built like a brickhouse. The output stage has zero negative feedback as well, so the measurements are pretty good all things considered. Full disclosure: I got this second hand for short $$$. Turned out to be a good buy at the second hand price. Your points regarding special claims with a given piece of hardware are well taken. I think many of us have learned this the hard way over time.
Regarding the D900. one could argue that R2R is more expensive to make, and generally does not measure as well as D/S DACs. To me, it definitely sounds better than any of the other Topping DACs to date..
I’ve got no issue with the build, nor its performance (given its architecture and electrical design ethos). And I’ve not heard one, so I’m not commenting on how it sounds - generally I prefer lower/no feedback designs to those with massive open-loop gain and epic levels of negative feedback.
The white paper illustrates what I’m talking about, though. Not in a bad way, just confirming my thought. But it explicitly calls out the issues with the choice for an NFB implementation, which then necessitates compensation (the “super linear” piece), and then also makes the point that this is not something new.
It’s a bit like Chord DACs, where they have a filter with a very high tap count. It’s not needed for the raw sample rates it’ll be fed. It is needed for the highly-upsampled result from the WTA1 and WTA2 steps. If you were not doing such extreme upsampling, a vastly shorter filter could be used to the same effect (about 1,000 taps is common for 44.1 kHz material in most chip-DACs).
Which doesn’t mean that way of doing things isn’t beneficial - just that some of the implementation is dictated by upstream choices and is solving for other design decisions rather than something innate to the process.
I like interesting, novel and clever designs. I’m just less keen about the way they are often marketed - and charged for - as something special, when they’re really just there to fix other issues inherent to the design and serve no other purpose.
Nothing specific to the CanEVER products - lots (maybe most) manufacturers do it.
3 Likes
I understand where you coming from regarding the marketing of these products, and I am in agreement. The Canever caught my attention, as it was quite different in design from most DACs, plus it had a cool built in headphone amp. The asking price was quite reasonable, since most audiophiles aren’t familiar with the brand. Figured I could easily flip it if I didn’t like it. Turns out it really does sound very good, definitely a keeper. I thought the white paper was pretty informative, without being over the top with outrageous claims.
Regarding the comments about Chord, the issue come down to how the units sound with their approach. I’ve had the Chord TT, Hugo-2, and TT2 over time. They do have a different sound signature from most D/S DAC’s (RME-ADI2 or most of the Topping products) as an example. To me, it comes down to personal preferences. I liked the Chord DAC’s, but I prefer the current DAC’s I have on had.
Caveat emptor.
2 Likes