Ares II doesn't get a good reivew on ASR. Is it because R2R usually doesn't measure well comparing to chip-based DAC?
It is a multifaceted problem where the issue is simply ‘educate the consumer’.
Amir noted going back about eight years ago, when the first example of ‘tuned for spec sheet’ equipment floated across his workbench, that ‘ASR may have created a ‘monster’.
heck camera industry made the error of heavily mentioning ‘megapixel’ when launching new models, and
made it look like megapixel mattered MOST (not true at all, lens and processing are equally important, and ‘with regards to megapixel’, a 4mp fovean sensor is ~ 16mp xtrans sensor ~ 24mp CMOS sensor equivalent- naturally the market is filled with high number CMOS sensors as ‘the numbers LOOK more impressive’ and “yellow belt consumers can READ a spec sheet”..)
The camera industry soon realised they had made a beast that would take them down.. (the electronic companies started to make ‘technically capable’ (yet aesthetically useless) cameras).
many folded knowing that consumers would be ‘too hard to educate’.
That is what has happened with hifi.
We can read spec sheets, so ‘spec sheet warefare’ allows us to feel we are using our education to aid in making a purchase (helps makes us feel good in the process)
the problem is the same as Amir identified nearly a decade ago-
once manufacturers learn to tailor for the ‘tests’, they can build parts that look impressive on paper but may not be..
The best ‘easy’ example that comes to mind in a parallel field (so we don’t all get our knickers in a twist defending our recent ‘hifi’ purchases) is Samsung Solid State drives.
For years they would beat Intel drives ‘at the spec sheet level’ (tailored to perform the technical tests better), and yet in the ‘real world’ the Intel drives, that were optimised for situations users would actually find themselves in, would easily trump the ‘better spec’d’ Samsung parts.
Samsung are one of those ‘masters of spec sheet manipulation’ companies.. (they made this apart of their business model years ago)
With regards to Amirs’ recognition of ‘a manufacturer having realised his testing process’, could ‘tailor a product’ to suit. In that ‘first’ recognition of a piece of kit on his test bench that was objectively worse for targetting specs’ that didn’t necessarily make it a better product, he, and the thread that followed made quips about how their webpage may have been responsible for ‘shaping the market’.
Spin on a decade and ‘just about all amplification stages’ are feedback designs (better benchmark results) etc.. (yet may impact audio negatively)
When a manufacturer tailors for ‘audio’ reproduction, they typically give up ‘absolute’ benchmarking status.
Sadly the nature that is Ladder DACs, a lot of electronics can get in the way of ‘perfect benchmarking’, and so on the ‘spec sheet’ level of warefare, they are ‘not very good’.
The good news is that our head and hearts doesn’t actually net much from the numbers, so much as we DO GET from the reality of ‘the product as a whole’.
A denefrips Ares Enyo/12th
smokes the equivalent pricepoint multibit and D/S offerings, typically.. especially when up against ‘fly by night products’ that exist for spec sheet warefare.
Sabre chips are a great example of DAC chips that were tailor made from the ground up to ‘do well with benchmarking’, and yet when we look at how difficult it was to implement the early designs and get great results from them (they were interference prone beasts that could easily be ‘tripped up’), but for all their difficulty to implement (and cost), they DID yield some great numbers. (and most buying them are not Conductors with ear training who know that not all recreated soundwaves are ‘equal’)
I have recently posted on headfi, referring to a six moons interview with a Burson engineer, who informed of why they chose the Sabre chip over a Wolfson 8741 (the other ‘most expensive’ single chip option), and they found the eight channel ESS9018 benchmarked better in channels paired mode (the typical implementation coming to market by a few hifi companies), but that even though it benchmarked worse as a two channel (non dual diff config) DAC (I am guessing this leaves parts of the chip not engaged), the actual audio quality output was ‘more musical’.
Whether a device is built from ground up to benchmark well (ie a Topping D90), or perform well (musical DACs) is a decision that the designer/engineer makes.
If you are an average hifi company (or SSD maker or camera maker)- you target spec sheet absolutes and let uninformed consumers THINK their spec sheet sales data is ‘all that’..
If you have the chops to stand on your own two feet, and let your skill and talents ‘sell’; then you can forgo spec sheet warfare and ‘let people enjoy’ a nice product.
Sadly the world has many more spec sheet researchers (as Amir noted) that will be misled by ‘better benchmarking’ and there is no short/easy answer.
Teach a man to fish?
Sure- educate the consumer..
Science has helped audio- sure… but I wouldn’t let a spec sheet determine my next hifi purchase (or generate my ‘short list’)
Every time I have upgraded my hifi to ‘lesser spec sheet’ capability (but from higher price points/‘equipment tiers’) I have netted vastly more
musical results.
Having a Denafrips Ares is ‘the top of shortlist’ for ‘budget DACs’ -
it is near equal priced to a Topping D90
The Topping D90 is probably the worst (dedicated) DAC in my house.. (certainly NOT by the spec sheet, which indicates it is ‘the best’)
Fast switching vs an iFi Diablo (Diablo wins), or a Burson V2+ (easily), the D90 often doesn’t reveal anything empirically worse.. (sometimes a collapsed soundstage depth makes a recording sound ‘funny’ vs EVERY OTHER DAC I have ever heard)..
but on great recordings…
hmmm
Amusingly- I am running my (Denafrips owning’) mates old DAC, an Audio-Gd nfb 3.1 (possibly AudioGds WORST EVER DAC) (I put internal photos up a couple of days ago here on headfi;
https://www.head-fi.org/showcase/audio-gd-nfb-2-and-nfb-3.13105/reviews )
interestingly the NFB in the DAC parts product name (NFB 3.1) stands for ‘non feedback’ design- it has many rails of class A and outputs ‘proper’ audio… (dynamically so, and ‘very revealing’)(with actual soundstage depth too)…
Would I choose a ten year old DAC using outdated topology over a ‘flavour of the moment’ piece of ‘fad-fi’?
Yes.
Absolutely.
IIS, and balanced and … all these things that become ‘flavour of the moment’ to push us ‘ever forwards’ (generally finding synergy with other equally designed/ similarly built parts) do not help anyone except ‘second hand market’ dwellers who can score high quality parts for peanuts.
I’d judge equipment, against my scientific mind, by pricepoint and weight, and then brand, and then listening output/spec sheet..
giving least weight in my comparison to ‘spec sheet’ (and more weighing to ‘weight’).
Re: a Denafrips benchmarking poorly- that is the nature of the beast for components in a R2R design.
is it an issue.. Not at all. (nor is it something you’d be able to notice)
edited to add a hyperlink to ‘photos’