Good measuring DACs vs.

I recently owned and compared a number of DACs in my system and was particularly interested in the sound of two "perfect measuring" DACs, the Mola Mola Tambaqui and the Benchmark DAC3 HGC. With either of those, it seemed every note came out clearly, cleanly, and accurately, without a hint of distortion. Both have been reviewed by Stereophile, and John Atkinson concluded his review measurements with,

"The Mola Mola Tambaqui offers state-of-the-digital-art measured performance. I am not surprised HR liked its sound."


"Benchmark’s DAC3 HGC offers state-of-the-art measured performance. All I can say is "Wow!"

So, why is it that neither of these two objectively perfect DACs seem to emotionally engage me to the same level as my Mojo Audio Mystique EVO Pro, which is an R2R design using (basically antique) AD1862 "Z" chips? How can I not perceive the same levels of body, tone, or dimensionality from two DACs which exhibit "state-of-the-digital-art measured performance" and that really do nothing wrong?


Aha.  Because the term "measured performance" means nothing.  I can measure a ford truck and it is longer than a Ferrari.  Does that make it faster?  Just because numbers are published doesn't mean they sound better.  That hack website that makes ridiculous claims to get clicks proves that.  I won't name them.

For engaging look at a Lampzator, any of the DHT models.  


I find the Stereophile review of DAC 3 HGC, to put it gently, weird. The reviewer was comparing it to PS Audio DSD by having PS Audio DSD connected to the Benchmark analog in, and stated he heard no difference between the two DACs. Well no 💩!!!

I owned two Benchmark DACs, the DAC 1 years ago, and DAC 3 HGC fairly recently. I will say the DAC 3 HGC is slightly more musically engaging than the original DAC 1, but that’s not saying much. I sold it because it sounded like a studio monitoring tool, cool and dry, unable to convey emotion. 
I haven’t heard the Tambaqui, but would love to. So I can’t speak to it but will take your word for it.

Perfect measurements don’t translate to good sound. I stumbled upon an ASR review of the Marantz SA-10 SACD player on YouTube and watched it just for 💩s and giggles. The ASR crowd in the comments was hailing Amir for saving them $7400 because the $99 Chinese Toping DAC (his reference) measured much better. Well guess which unit sounds better and not just marginally…

PS.  I'm not talking down either of the DACs you mentioned.  I'm just saying measuremets aren't what makes them good.

A lot of people like the sound of the Tambaqui and to clarify, I found it to be a more enjoyable sounding DAC than the Benchmark but neither provided the level of enjoyment for me in my system that my Mystique EVO Pro provides. Even the SMc DAC-2 is more musical. It is hard to verbalize since the Tambaqui basically had no audible flaws and had I not had other DACs here I probably could have been happy with it. Even the Benchmark is respectable for what it does, which is clear, clean, absolutely distortion-free sound, but for whatever reason, I am not hearing the level of emotion that I need in my reproduced music. Regardless of whatever flaws they may have, the EVO Pro and SMc DAC-2 both display a strong level of emotion in the music, IMO of course.

m0ns List at sonusapparatus, under DACs, provides a description that is in alignment with what I hear from the Mojo Audio Mystique EVO Pro.

"A very fun, organic, and rich sort of DAC with both fantastic micro and macro ability, great impact and slam, great control and grip, and organic presentation with a bit thicker weighting, and pretty great stage too, more forward in things, although sometimes strangely picky. Does what it does really well, super fun and engaging without being offputting, limits of coloration for me."

Advertising dollars tend to make things sound better than they actually are. 😁

The Benchmark DAC3B (same as the HGC) is a little hot on the top. However, that is sometimes needed. I am using it today with a slightly warm speaker and preamp. The DAC3B seems to level off that warmth a bit for my tastes. 

I loved my RAAL VM-1a tube headphone amp with the DAC3B. 

Measurements are grossly inadequate as they can only approximate dynamic conditions.

Most commercial Dacs measure well, but they sound grossly different. 

As far as I know, below a certain level of distortion and frequency response flatness, there’s very little correlation between better measurements and better listener preferences.

Just like with amps. Below around 0.1% distortion, the specs that we commonly use rarely help us know what will sound "better."

In other words, almost all DAC’s have very good measurements, especially compared to what was out in the 1990’s. What sounds best to listeners has never been proven to correlate, though there are successful designers like Nelson Pass who argue strongly for certain types of distortion being better than no distortion, and his commercial success speaks for itself.

Personally I seem drawn to DACs with AKM chips... but I certainly can't say it's true always.  I used to really like Burr Brown based DACs as well.  Ages ago.  No measurement to point to and explain it.

I have tried a dozen or so DACs at wide price levels, wide technology differences, bad measurements, good measurements, no measurements. I don't know why this is but I have had highly rated DACs sound not as good to my ears as ones costing 1/4 price no one cares to even write about.  But while some may not sound as good in one system when I swapped it to another system in the house, with same source but different amp and speakers, and it sounded great. That led me to trying different ones in different systems (I have 4). I finally thought I had it figured out and so bought a higher/better/pricier model of one of the DACsfrom same company for one system and it sounded worse than their cheaper less capable model. Room/amp/speaker dependent? These DACs have included ones with various filterings, two chips, one chip, different chips in otherwise same model, R2R, NOS, and so on. I finally stopped when I thought stuff sounded good. It is like balancing a fulcrum. When you get it slowly back away and don't touch anything.

I think that one premise that a lot of people seem to accept is that distortion always sounds bad. Any change from the input signal is distortion. So all the things that studios do to make recorded instruments sound “better” are distorting the signal. Noise is another thing, but I don’t think the human ear is that sensitive to noise if it is constant. Listen to Take Five on vinyl. Loaded with tape hiss, but still enjoyable.


DAC measurements are the least reliable of all, especially since they tend to concentrate in the D rather than the A side of the DAC. As per Erik Squires, most modern DAC chips measure excellently in terms of their linearity, so most of the current measurements don't give any idea of sound quality of the finished product.

Charyo hits the nail on this one. Not all distortion is bad sounding. Look at those 300B amps, they sound good, but distort a lot. Some distortion is pleasant sounding, so if it has a high price tag, then it's good.

Post removed 

Do people understand the brain and aesthetic preference so well that they can capture these in various acoustic measurements and then have those measurements standardized for deployment in the audio industry?

Of course not, but since numbers are "serious" they can be seriously misused.

Some numbers mean something related to aesthetic taste.

But aesthetic taste in audio is as complex a phenomenon as "marriage" or "friendship" is. So why would we assume it can be reduced and parlayed, thus? Rhetorical question.

@koestner 1+

Some manufacturers intentionally add harmonic distortion to amplifiers because many audiophiles like it. Not to mention that most audiophiles are trying to judge equipment through speaker/rooms that have insane amounts of distortion. It is like trying to evaluate a photograph through a foggy window. Even with the very best systems one has to be careful.  Then there are the effects of price and appearance on audiophile psychology, expectation bias and so forth.  Put in a properly blinded situation I doubt any of us could reliably identify any of these DACs. What any piece of audio electronics sounds like, with the exception of amplifiers, is more a matter of psychology than anything else. 

If measurements are meaningless, an audiophile's opinion in regards to sound quality is worthless...except mine.


Because it's all subjective. I'm sure that as many people that love the Mola Mola & Benchmark there are equal amount of people (or more) that don't care for it. Doesn't make either person right or wrong. 

You could almost drop it into 2 different classes of people. It would be oversimplifying it but we have those that like the utmost transparency and others that prefer the sound of a little distortion. Not noise generated by power supplies or internals but distortion in the form of harmonics..which a little bit of certain order harmonics don't sound like distortion in the way many might think. Still distortion though. 

I do feel that this is where measurements can be useful. Actually I'm of the belief that all data and measurements are useful.. it also keeps manufacturers honest or helps us know which one's make BS claims. Trust your ears as they say, but also don't discredit measurements and data..they can tell you a lot about a product. You don't have to be of the mindset that perfect measurements mean perfect sound. Well to some people it does but it's all subjective at that point, so whatever. 

Personally I like a dac with upwards of 20bits of dynamic range or "state-of-the-art measured performance". Clean and transparent. I can add some flavor or distortion somewhere else down the signal line. 

I sent back a few of the aforementioned great measuring DACs and kept the ones that measured worse. Another lesson in learning that listening matters, a lot.

@carlsbad2 Ford truck can definitely pull heavier cargo than Ferrari. You can’t compare toy to a tool. Same thing in audio. Ferrari is fast toy for a couple of people for entertainment purposes only, while Ford truck is a tool designed for work with various pulling or hauling tasks depending on model besides taking you from point A to point B. According to traffic rules and regulations, Ferrari will make it from Miami to Houston approximately as fast as Ford F150. Tools usually have more genuine and descriptive measurement results posted on data sheet while audio toys for baby boomers rich and foolish children usually post "advertised version" of measurements.

Bottom line everyone hears differently. Goes with what sounds best to you not with something someone else tells you should sound good or best (or measures good).

czarivey: A Ferrari will get you a lot more girls than that Ford truck though!

I wasn't aware that rich children were foolish 🤔

@ellajeanelle not here in South Dakota. Girls will look at you and say who is that stupid city boy?

According to some Top Gear show releases, Lambor attracted lots of BOYS while hot-roded Jeep attracted GIRLS so go figure...

Rich children grown to everything "padded" and "softened" can’t be smarter than street rat by all possible means.  

By an insanely wide margin, the worst testing DACs would be those from Audio Note.  The images of waveforms are totally unrecognizable as sine waves.  Find one to audition.  That audition will convince you of the utter meaninglessness of measurements.

@koestner That is highly unlikely, but I’d be curious to hear if there are any. 300B are directly heated triodes, which tend to emit tons of RFI and EMI. Placing their circuits near a digital signal almost always results in noise leakage / interference. Typically any circuits of that sort have to be kept a good 2-3ft away from the digital circuit. 

@koestner Check out Lampizator. They have a few models with DHT tubes and at one time employed the 300B.  They may still.

I also compared Mojo and Benchmark. Agree that Mojo is more emotionally engaging. Sold the Benchmark. I also had a Border Patrol dac and really liked it. I have 300B amps. One might conclude that I like distortion, but what I listen for is dimensionality (presence, immediacy, life size weight), tonality, and a naturally put together soundstage. When listening, I prefer to think I am at a performance in front of me instead of being in the seat of the sound engineer. Benchmark is a good analytical tool but not something I could relax to for extended sessions.

Is it along the same lines as why listening to "The Doors -break on through to the other side" is more engaging in the car than on my hifi? Love the heck out of it on a long drive. Such a let down at home; and 24/192 remaster is no better.

mclinnguy: The Doors was one of those countless bands that never put out a descent sounding recording.  Probably sounds better in your car because your car's audio system is less resolving than you your home system...

A number of DACs have come through my system (most not owned by me) and I couldn’t tell much if any difference between them. The two I have, use chips nearly 20 years apart in design and I can’t make out predictable difference. I know, that’ll likely get me banned 😉

I chalk it up to DACs do their job and any “flavor” through distortion or anything else must be pretty darned finite (again, if it’s there) for the device to be working as it should. I would also suspect it’s as simple as many speakers, rooms, or speaker x room interactions just exact more influence on overall sound than kit further upstream. Could be wrong, but doesn’t make much sense otherwise given how sound waves and reflections (vs. electrical signals) work. Amps and speakers have both had serious (albeit model-limited) investigation on the correlation between certain measurements and listener preference. I know of no such thing for DACs, disappointing.

Obviously measurements are only useful for this sort of kit if they can be demonstrably correlated with listener preference (and on a meaningful scale). Using uncorroborated measurements to predict/determine suitability of a device is about as well-informed and free of bias as expecting one DAC to sound better than another because of chassis color. And here’s a real pitfall: if one tests/studies measurements (that you understand) before listening to the device, irreversible bias has been introduced. Measurements and consumer pref studies are like turkey and dressing - different groups might prefer one or the other, but relatively few will argue they’re gonna go better together.

@czarivey thanks for agreeing with my post, even though you apparently (or purposely) don't understand it.


It does seems more like you apparently (or purposely) don't understand your own post and I kinda clarified that in details.


Do you buy a painting based on the purity of the oils used??

That's an interesting quote, and I'm not sure how to take it. My first thought is that it's like asking if you buy a piece of equipment just because it uses expensive components. There is something to that. It might be more reliable and last longer. But it won't necessarily measure or sound any better. But it could be that the pallet of oils you are referring to are the measured results themselves. It could be that the DAC measures as it should in every parameter, as the creator intended to give the desired perceptual effect. Or it could be that the measurements that look bad weren't intentional, but were deemed acceptable and were necessarily compromised to get something else right that typically isn't measured because it's more difficult to measure. I suspect the former is more common. Intentional distortion that has a good overall perceptual effect on a lot of listeners with systems at a particular level. 

I’m one of a minority around here who finds that poor measuring gear is typically less engaging. My only hypothesis is that it’s because I’ve lived with systems, (speakers through to the source) that produced practically zero audible distortion the bass frequencies. 




I really didn't discover DACs until I started ripping CDs to a NAS.  The improvement vs using the CD player wasn't subtle.  First DAC I bought was an Emotiva.  Impressive thing with a hefty remote control, sounded fine.  Then I bought a Modi Multbit (older version), and there was a difference, particularly in the highs.  Then a Modi so I could run it off the USB power of a NUC.  It had better SINAD based on ASR's reviews, so I thought it would be a cheap way to see/hear if SINAD mattered.  The difference was subtle, but getting rid of a wall wart power supply was worth it.  But I missed the remote control of the Emotiva, and bought a SMSL SU-9.  Again, a subtle but definite improvement based on blind testing.  In each case, improvement was heard in a better measuring DAC.  The spend wasn't necessarily linear with the perceived improvement, but I wonder at what point it has more to do with voicing than accuracy....