How much have DACs improved over the past 10 years?

[Solan]

Which technologies have progressed, and which have stood still.

In the chip corner, I hear little news about anything new and revolutionising, and even “older” model from a few years back remain top models. TI’s best chip is a Burr-Brown from … 2008, I think, and improvements from other companies have been marginal and measurable rather than audible, as far as I can discern.

R2R has improved lately, though, and people go euphonic over (for instance) the Denafripp sound, and the Multibit variety of R2R (like in the Schiit models) also get shining reviews reviews. So something has happened in the R2R field. (And if you have money, the dCS ring DAC has happened, as its own variety of R2R.)

But is R2R really any better than chips, once we’re past a minimal price level? Or are we still just choosing what to sacrifice: clarity, or musicality?

Finally: Is it rather than D->A itself the amplification stage inside a DAC that distinguishes the bosses from the wannabes? That is, a technology that is essentially the same these days as it was 10 years ago.

[Polygonhell]

There are a lot of very good even 15+ year old DACs.
The big changes are better USB if the older DAC’s have it at all higher resolution inputs 192/24 starts around 2004 ish prior to that you won’t find anything above 96/24.
A high end DAC from 15 years ago is still a high end DAC today.

IME at the higher end of the DAC space it’s less about R2R vs specific Chips, and more about implementation, the quality of the output section, and the power supply.
Lampizator moved from R2R to delta sigma designs in their high end DAC’s recently, and the general consensus seems to be they still sound like Lampizator DAC’s.
But there are lots of great DS DAC’s and R2R DAC’s.

The Ring DAC isn’t R2R, it’s basically a 5 Bit Delta Sigma design with proprietary upsampling and noise shaping. With some additional tricks to make the 5 bits VERY accurate.

[Lars]

I just took delivery on a Chord DAVE and it is one amazing DAC; Chord does it their own way and it works great…

[jnorris]

I keep swapping between the original Peachtree DAC-IT (ESS9022), the SMSL SU8 (ESS9038), Schiit Modi Multibit (R2R), and Topping E30 (AKM4493). And damned if I can hear a clear and repeatable difference when switching back and forth. I get impressions, but nothing more concrete than than. The thing is, the Peachtree is a design from 2011. So I have to say, there isn’t much dramatic improvement in the last 10 years

[Doug_Dame]

According to Google Search, the Peachtree DAC-IT cost $999 in 2011 dollars.

The Schiit Modi Multibit is $249 and the Topping E-30 is $149, in 2021 dollars.

If the audible performance is the same, for 1/4 or less the price, that seems like quite remarkable improvement to me.

And the number of people able & willing to buy a $250 or less DAC is VASTLY larger than the number of people who will buy a $1000 DAC. From an affordability point of view, we’re talking the difference between audiophile elite (+/-) vs close to mass market.

[generic]

@jnorris

So I have to say, there isn’t much dramatic improvement in the last 10 years

@Doug_Dame

If the audible performance is the same, for 1/4 or less the price, that seems like quite remarkable improvement to me.

This comparison needs to focus on best-of-breed for each chipset or type of technology. It’s fair to compare a modern (2021) Chord, a Schiit Yggdrasil, a Mytek Brooklyn, and an RME ADI 2 fs to the best DACs from the same vendors as far back as possible (e.g., 2011). It’s not fair to compare $1,000 or $5,000 products to $150 products from different eras to assess potential and audible improvements. Once one compares the best of each generation, one can then shift to value for money comparisons.

---

Regarding hearing DAC differences, I compared my AKM4490, Burr-Brown, and an older ESS Sabre DAC against the Bifrost 2 (Multibit). I also have very bad but clear memories of a now gone TI PCM5102 DAC. There are major, obvious differences to my ears. This comes in two forms:

• Brightness and high frequency artifacts. My Delta Sigma models generate piercing whines to a greater or lesser extent. This is easy to detect as they cause tinnitus or rapid fatigue. The Zen DAC is the best of my Delta Sigma DACs, as it rolls of the highs. However, additional but piercing high frequency content can add to ‘air’ and assist in identifying differences between amps and headphones/speakers too. So, there can also be real differences between ‘analytical’ and ‘laid back’ DACs.

• Nuance in room atmospherics, reverberations, plus vocal or instrument timbre. The Bifrost 2 greatly outclasses the others, with distinct layers and just “more stuff” there. The Delta Sigma DACs are not even close. The Zen DAC has similar weighting and tone (mid emphasis and richness), but is relatively one dimensional and blurry.

Was it fair for me to put modern $30 or $129 or $250 products against a $699 one? Nope. See above using the best available for each generation.

[Solan]

@generic

My Delta Sigma models generate piercing whines to a greater or lesser extent. This is easy to detect as they cause tinnitus or rapid fatigue.

Now this is interesting in its own right! Do you have any references to this? Asking because I myself have a high-frequency tinnitus. Only audible when I try to sleep, or (ha ha) everything around me is quiet, but annoying nonetheless. If a change of DACs could remove that, I’m all in!

[jnorris]

The DAC-IT was $499 back then. I bought it as a demo in 2015 for $300, and soon after they came out with the DAC-ITX which again was $499 (but used the ESS9023 chip).

[generic]

@Solan

Do you have any references to this?

References…no one understands tinnitus very well. For me I think it follows from neural fatigue – see the Do your ears ring? thread. Also see my and other comments about the need for magnesium as a neural transmitter.

For fatigue I rank order my DACs as follows:

1. Bifrost 2 – Minimal fatigue but limited ‘air.’ It’s not warm so much as rich and layered in the middle. It becomes faintly grainy and rough but doesn’t whine.
2. Burr-Brown – Tuned for mid range richness, but has a bit of “lightbulb buzz” that gets worse over time.
3. AKM4490 – Lots of smoothed air, but some setups routinely whine.
4. ESS Sabre – Ouch. Faint but constant whines.
5. TI PCM5102 – Useless. Loud whines all the time.

[jnorris]

I have no doubt you think this is true, but after 6 years of listening intently to different DAC over Martin Logan electrostatics I have never heard what you describe. Have you had your hearing checked? Perhaps you are oversensitive in this range or have overly bright speakers. If, indeed, these artifacts exist I would suspect that the low end DAC market would dry up and die instead of flourishing with multiple positive tests and reviews.

[generic]

Individual hearing sensitivity varies greatly. Just read the polarized opinions of the Focal Utopia on this site – I fall in the middle on that debate. Many people also suffer from permanent high range hearing loss these days due to rock concerts, race cars, guns, and more. Yes, I’ve had my hearing checked and it was normal per ‘objective measurements.’

Also, Martin Logan electrostatics may or may not offer the best way to compare DACs. I hear more high range resonances/artifacts and differences on dynamic drivers than on planars.

[jnorris]

I feel the opposite. The Martin Logans exhibit a level of transparency and accuracy that box speakers can only dream of. As for resonances and artifacts, the ubiquitous 1" dome tweeter in many of its iterations is far more guilty of those aberrations. Then there’s the boxiness and chestiness of the lower mids to confound things.

[generic]

This is why scientists developed controlled experimental methods. Each type of technology has different potentials, limitations, and quirks. My planar headphones generally don’t result in the artifacts I describe, as they are inherently more diffuse.

I’ve heard Martin Logan setups multiple times, and none of them struck me as particularly technical (e.g., KEF level). My cheap Sennheiser HD-600 also communicates all sorts of noise and hiss and other wonkiness. It responds well to amp, DAC, and cable changes. This is useful for detecting upstream differences, even if not accurate or transparent.

[tjg]

@generic - have you used EQ to roll off highs with various DACs and amps?

[generic]

Yes. That works when possible, but not all signal chains allow EQ. DACs become less distinct and less identifiable with rolled off treble. Amps respond differently, as some are center-focused, some lack bass extension, and some are intentionally distorted.

There’s technical a game of chicken going on: Obtain maximum detail with minimal harshness or noise.

Formal signal detection theory: Some DACs perform poorly on both noise and resolution (straight line) while others do better on one or both dimensions (d’=3 or d’=4).

[Lars]

I had a Hugo 2 that I was using as a preamp into my Quicksilver Tube Headphone amp, and it was good, but I later got the RME one for about half the price of the Hugo and I liked the sound from the RME better and sold the Hugo. Haven’t even done any of those sound processing adjustments on the RME. The Chord DAVE I just got is another story altogether, but at its price, it SHOULD be!

[Pharmaboy]

I can only talk about my own DAC journey & how that resolved my longstanding sonic issues with delta-sigma DACs:

1 - My first 2 or 3 DACs (starting ~2005) were delta-sigma. The system they were in changed over time, but no matter which speakers, amp, and cables were in use, these DACs sounded pretty miserable–in essence, continuing the sonic nasties I’ve always heard from digital. By “sonic nasties” I mean 2-dimensional notes (not much if any air around the notes); accentuated transients, particularly on things like percussion, cymbals, etc; a bright/edgy quality; and a thin-ness of the harmonic content of various instruments that really have it (violins, other stringed instruments)

2 - I got my first non-delta/sigma DAC about 6 years ago, the Audio GD NOS 19, a very nice Chinese non-oversampling multibit design. And overnight, my entire relationship with digital changed: the nasties were gone; music sounded more real & present; and I was very happy to trade a slight roll-off of upper treble and dimininuation of soundstage for that harmonically richer, non-edgy sound

3 - I sold that one and got Audio GD’s multibit, the DAC-19, the sound of which is pretty close to their NOS design

4 - Last year I decided to go NOS again and got the MHDT Labs Orchid, a NOS DAC built around the semi-ancient Philips chip, the TDA1541A + a tube output buffer stage. Though this DAC’s stats can’t compete w/more recent D/S designs, the sound is glorious, perfect for me–very musical, natural, organic, relaxed, yet revealing most of the detail I know to be in familiar music

It’s always possible that my sonic preferences and hearing totally suck–maybe I’m missing high treble content that everybody but me can hear…maybe this DAC sounds like shit & I’m the only one who doesn’t know it…? Anything is possible.

But from my perspective, NOS & multibit sound way more like real music than any previous digital I’ve heard. I’ve heard these DACs through revealing ATC speakers, among others–the sonic goodies are always there no matter how the rest of the system is configured.

These days I can crank my speakers or headphones to moderately loud levels on occasion and not need NSAIDs and a cold compress on the forehead afterward.

PS: The best of the d/s DACs I’ve heard is the humble Peachtree Audio DAC iTx. I had 2 and gifted one to my brother. I use is strictly for backup; it doesn’t sound as good as the Orchid, but it commits fewer d/s sonic misdemeanors than other d/s DACs I’ve had here.

[goldilocks]

I too have tinnitus and no one understands it when I say something sound harsh. As if having tinnitus means having extra sensitive detector of harsh sound. Interesting you mention “air” because I find that DAC/AMPS with airy sound is often harsh - to me airy sound is adding multiple higher harmonics of the fundamental frequency - filling the room, making it “full”. The opposite of this is the sound with “dark background” which often sound intimate and narrow (soundstage). I do not get fatigued by this narrow sound as much. While Bifrost 2 may be less fatiguing because it has no air (narrow soundstage), it can still sound harsh if the timbre is grainy. To me, timbre being grainy/smooth is an additional dimension of timbre that is different from “air” and “intimacy”. So in my experience, a gear with smooth and intimate timbre yields the most non-fatiguing sound. To my ears at least.

[pennstac]

Welcome. I suspect different meanings of “air”. @generic, do you have either the frequency chart or a pointer to it that’s been posted here a few times? The one that relates certain high frequencies to air?

Some of the confusion surely comes from a non-audio meaning of “airy”,which sounds to me related to your usage, @goldilocks

[generic]

@pennstac – I think you are talking about the DIY Audio Heaven chart:

It’s a bit cluttered – there’s a nice thread on frequency characteristics at Head-Fi with that chart and more:

What are the main 'enhanceable' sound characteristics and their ranges?

This one gets to the point about air versus sibilance:

@goldilocks

As if having tinnitus means having extra sensitive detector of harsh sound. Interesting you mention “air” because I find that DAC/AMPS with airy sound is often harsh - to me airy sound is adding multiple higher harmonics of the fundamental frequency - filling the room, making it “full”. The opposite of this is the sound with “dark background” which often sound intimate and narrow (soundstage).

I don’t disagree. Tinnitus, be it from neural quirks or mild hearing damage (e.g., routine exposure to rock concerts, motorcycles, guns, heavy machinery, etc.), may shift your frequency sensitivity and perception around or magnify what should be in the background.

To my ears the ‘air’ frequencies are so very high that they don’t give me issues – test yourself with the frequency sweep tool at Audionotch. To my ears there’s an intense perceived volume bump around 2500 to 3000 (e.g., threat and danger calls: scream/cry frequencies), then it moves to mosquitoes or fingernails-on-the-chalkboard between 5K and 10K. Air lies beyond those.

What bothers me the most seems to be randomness in the sibilant but still clearly tonal frequencies. The original BF2 certainly has defects, and defects might be heard differently by you than me – and probably are heard differently. I’ve found that the DACs/amps/headphones easiest on my ears generate clean, pure, non-random sounds. To my ears anything generating a random treble spray (e.g., some Beyerdynamic and Grado heapdhones; Koss Porta Pro; the Sennheiser HD-600 on some setups) mixes the air frequencies with the sibilant frequencies. The Focal Elex does something different, as it sounds to my ears as if DAC/amp randomness is transformed into very loud and unpleasant tonal bands. To my ears the air per se is painless and just floats about as whooshing, breath, and wind.