Denafrips ARES II USB R2R DAC Review

[Frank Dernie]

So what means? 0,001% vs 0,0001% THD?

No.
IME if the distortion is below 0.1% (-60dB) and the noise below -80dB it is transparent to me, most electronics has been transparent for decades and differences imagined, unlike cartridges, speakers and headphones few of which approach this level of performance.
After all all my record players sound pretty good and they are nowhere near as good as 0.1% distortion and all have more noise that -80dB, even the best, very, very expensive one.
I love my Verum 1 headphones.
Dicking about with electronics is an exercise in imagination and belief in myths IMO.
The transducers are where the weaknesses are and it takes a spectacularly bad bit of engineering to make something genuinely audibly different on the electronic side, not enough power or inability to drive a complex load are about it unless you go for nostalgic vintage stuff.
I have been dicking around with this for over 50 years now and it is sad to see the death of expertise permeating almost every walk of life

 

[YSC]

It is worth the designer measuring his designs to make sure he hasn't cocked up or getting hum loops etc but it takes spectacularly bad engineering to make electronics audibly different.
When I have auditioned DACs carefully level matched all the properly engineered ones sounded the same, yes, though there were some reconstruction filter choices which sounded different. The most expensive was £14,000 the least £1100 fwiw.

Most transducer designers don't do much proper engineering, sadly, maybe that is why so many of them are not very good?
All my headphones sound different to each other and I like most of them in one way or another, the only ones I end up listening to only a short time before changing are Kingsound electrostatics.

For transducers my favourite is the old Hifiman he500 planar headphones, really liked how it produces sound and bass

 

[Frank Dernie]

I'm jealous

I am sure the differences are mainly just frequency response variations. I deleted my last post since I was just teasing/trolling and it is not fair.

 

[Marlowe]

BTW, the review appears to understate the USD price a bit. According to the website of its US distributor (Vinshine Audio) it's currently $768 USD, apparently including free shipping--and an order lead time of approximately 6-8 weeks, which I believe applies to all Denafrips DACs. (I'm seriously considering getting one later this year, but since I'm retired on a fixed income and don't use credit, it takes a few months to save up.)

 

[Robbo99999]

There are some interesting components including that toroidal transformer.

lol, looks like 1950's combined with the 1970's or something! I'm only new to the audiophile scene, so my knowledge is limited, but I did buy the SoundBlasterX G6 DAC and combined it with JDS Labs Atom Amp +(AKG K702)...with the first two being bought on the strength of the reviews on this website! But yeah, this DAC doesn't seem to measure well...seems to be so many overpriced DACs that are not measuring well....I think they're just trying to take people's money on the perception that increased price and increased weight equals higher sound quality, ha! Thank God for this website!

 

[Spocko]

And better me ;-) I could no hear a difference either.

To be fair ADI is much more functional vs Modi 3. It has remote, it has volume control, it has balanced out, it has digital EQ and a nice display. But all I remember that both sounded the same on a speaker system (Schiit Aegir to B&W 702 S2). Maybe now I would hear a difference? Do not know.

And now I listen more on headphones. After playing with Gungnir and Bitfrost 2, I realized I can’t stand Modi 3 sound... I just can’t listen more than 10-15 min with it. But with either Bitfrost 2 of Gungnir I can listen and listen. Vocals and strings - they sound so much better. I do not know why and I will not use silly words to describe it.

Ohh and I learned that I also like Rogue Audio The Pharaoh - that has tube pre-amp and class D power amp. I always though that these are two things that I should not like - tubes add distortion and that class D will never sound as good as A or A/B... Something must be really wrong with me - but how do you measure that ;-)?

Some people prefer the taste of fried tofu to BBQ baby back ribs, it’s all good.

 

[Spocko]

Then it is useless to test any solid stateelectronics - everything is transparent.

You nailed it. It’s more about engineering quality and getting your money’s worth of effort from design and parts. It’s like a quartz Timex has been a great timekeeper for decades but people still pay extra for a good automatic. Many have sung the praises of the DAC $20 dongle. It’s why I listen to IEMs occasionally - to get out that last inch of transparency.

 

[majingotan]

It’s why I listen to IEMs occasionally - to get out that last inch of transparency.

To my subjective opinion I’ve never heard an IEM that can even compete with my Yamaha HS7 in transparency and extracting the finest details and nuances in music reproduction. Not even the 64-Audio U18Tzar, Tia Fourte, Sony IER Z1R, CA Solaris and my Andromeda IEM even match my Yamaha HS7 in that resolution regard not to mention the most lifelike tonality that no IEM can ever reproduce even with EQ. Though speakers are inherently better than headphones and IEMs in tonality to my own opinion.

 

[BogdanR]

...
If you like pumping ultrasonic noise into your amplifier and speaker, by all means, use the NOS mode.
....

Ahh! That explains the bats randomly hitting my windows at night when I play music!

 

[Spocko]

To my subjective opinion I’ve never heard an IEM that can even compete with my Yamaha HS7 in transparency and extracting the finest details and nuances in music reproduction. Not even the 64-Audio U18Tzar, Tia Fourte, Sony IER Z1R, CA Solaris and my Andromeda IEM even match my Yamaha HS7 in that resolution regard not to mention the most lifelike tonality that no IEM can ever reproduce even with EQ. Though speakers are inherently better than headphones and IEMs in tonality to my own opinion.

Impressive!

 

[majingotan]

Impressive!

Thanks! Only drawback is that you must have no louder than a 30 dB room background noise and listen on-axis at 2 feet away from tweeters. 77-80 dB SPL average listening volume are required to do that. Otherwise, you won’t be able to hear the orchestra track in Pink Floyd’s Eclipse track at that stated listening volume for example. In retrospect, you also hear the finest details in those aforementioned IEMs but the difference to my subjective opinion is that the finest details are “easier” to hear through HS7 speakers due to much larger sound stage to my opinion. Others might find the opposite otherwise.

 

[restorer-john]

I'm curious whether the test is able to differentiate the aliasing products that will result from removing the reconstruction filter. The issue with no filter is twofold - lots of ultrasonic junk that can lead to pathological behaviour of some amplifiers, and aliasing of signals in the passband.

And this again is the crux if it. NOS = non over sampling, which does not imply no low pass filter. That is the province of silly "tweakers" and "modders".

By design, most of the non-Philips 1st generation CD players were NOS out of the box, but they also contained perfectly designed (and expensive) multi element LPFs, so, as @Frank Dernie pointed out, the spectral splash up the spectrum simply wasn't there. Neither was aliasing in the audible band.

If anyone doesn't believe that, I have several 1st generation machines here...

And not all 14bit D/A converters used an oversampling digital filter either.

It's amazing that 37 years after the fact, people are still arguing about the superiority of FIR and IIR filters and OS vs NOS.

 

[Shoaibexpert]

Let's call this DAC audibly transparent in OS mode.
Curious to see if Denafrips at least solved the treble roll-off in lower bitrates in NOS and how the slow filter in OS performs.
Those wanting to spend considerably less on a DAC with similar performance (or better) have plenty to choose from but you can do much less for a lot more money as well.

Thats true...

 

[Baron Slater]

Since this is called audio science review it is amusing to see how many unscientific comments are being made about delta sigma and r2r dacs being the same. Here is a link to a presentation by no other than Burr Brown stating clearly that R2R is superior but more expensive. The only reason for the existence of delta sigma is cost reduction.

http://www.jitter.de/pdfextern/DesignSem5.pdf

 

[majingotan]

Since this is called audio science review it is amusing to see how many unscientific comments are being made about delta sigma and r2r dacs being the same. Here is a link to a presentation by no other than Burr Brown stating clearly that R2R is superior but more expensive. The only reason for the existence of delta sigma is cost reduction.

http://www.jitter.de/pdfextern/DesignSem5.pdf

Your ppt slides is quite obsolete BTW. Have you explored enough measurements on this site before making inaccurate statements?

 

[BDWoody]

Since this is called audio science review it is amusing to see how many unscientific comments are being made about delta sigma and r2r dacs being the same.

Ummm...who said they were the same? They clearly have different basic architecture...

A very good one is likely indistinguishable from most any competent D/S under any kind of controlled listening. Is that what you are taking issue with?

Also, first post seems a bit on the needlessly provocative side... Do we have a newly minted troll?

 

[Baron Slater]

I should have explained myself better.
There are many comments like "R2R can measure quite well too", "Not bad for R2R" or "I fail to understand the appeal of R2R" implying that DS is superior or R2R is at best the same as DS.

I see no mention of the difference between R2R DAC chips with laser trimmed resistors and R2R using discrete resistor ladders where 0.1% tolerance is clearly insufficient.

Indeed the two function in a fundamentally different way.

A DS dac with only a few bits needs loads of signal processing and feedback to increase the apparent resolution. The resolution of a ladder DAC is intrinsic to its design and does need correction.

Of course both can sound good or bad.

 

[Veri]

A DS dac with only a few bits needs loads of signal processing and feedback to increase the apparent resolution. The resolution of a ladder DAC is intrinsic to its design and does need correction.

You realise any R2R DAC performing at 'modern' D-S performance levels, reaches it only via lots and lots of FPGA-programmed feedback, to offset the resistor errors throughout the conversion?

Correction, feedback, in any case you'd want the cleanest output..

 

[YSC]

I should have explained myself better.
There are many comments like "R2R can measure quite well too", "Not bad for R2R" or "I fail to understand the appeal of R2R" implying that DS is superior or R2R is at best the same as DS.

I see no mention of the difference between R2R DAC chips with laser trimmed resistors and R2R using discrete resistor ladders where 0.1% tolerance is clearly insufficient.

Indeed the two function in a fundamentally different way.

A DS dac with only a few bits needs loads of signal processing and feedback to increase the apparent resolution. The resolution of a ladder DAC is intrinsic to its design and does need correction.

Of course both can sound good or bad.

But unless it’s a NOS r2r, those sample processes of oversampling are applied to both DS and R2R, nothing really pure as you’ve stated

 

[BDWoody]

I should have explained myself better.
There are many comments like "R2R can measure quite well too", "Not bad for R2R" or "I fail to understand the appeal of R2R" implying that DS is superior or R2R is at best the same as DS.

They typically haven't had very good execution...leading to very poor measured results.

Since the job of the DAC is to be transparent, and the rest is details, I am one of those who wonders what people are expecting when they buy one of these. It doesn't bother me...i just wonder what made them believe it should be 'better' somehow.

It isn't about sounding good or bad... it's unlikely anyone on the planet could distinguish one from the other (assuming they are both well executed designs) under controlled conditions.