R2R RIP or Resurrection?

[100Hz]

Ok so I’m a late starter in hi-fi and my audio experience is limited, including tech…so I’ve been reading around and looking to tweak my existing humble set-up. In fact my old Hifiman DAP is on its way out which inspired me to catch-up on latest DAC trends

This rabbit hole led me to the Delta-Sigma (DS) v R2R debate -well rehearsed on forums I know, so excuse me for bringing it up here…but I’m interested on the science take to cut through the metaphoric analogies that have been gaining traction. I’m not saying they don’t have there place, and indeed some I found helpful, eg comparing a telephoto optical lens to digital zoom when talking about oversampling..But when descriptions start gushing with “natural” and “organic” I get concerned that what actually might be lower resolution sound is being romanticised. Forgive my own analogy, but I’m of the generation that grew-up with film photography and although digital was initially a poor rival, I quickly came to accept that the finest film grain was never going to compete with digital, ultimately. Subjectively I still love the characteristics of film but in resolution and dynamic range digital is undisputed. (I realise that SD v R2R is not a digi v analogue question but the holy grail seems to be that “anogolue” sound, natural and organic). Similarly there maybe characteristics of R2R that are sought for, but I chuckle when a negative attribute to one listener is spun (corrected?) into a positive one by another. Eg, for the same R2R DAC: “imaging has diffuse borders” (a weakness) whereas from another perspective, this is interpreted as, “softer rounder edges, more natural”. Without live experience I’m not qualified to judge, but I feel disappointed this can’t be expressed more neutrally leaving the reader to decide if they want that kind of characteristic/attribute - in this case I’m having to guess the sound has softer resolution (?).

And so going back to the science, I get the impression that R2R, albeit an older technology, has been making significant advances to rival mainstream DS DACs, but is it a technology, in theory at least, more limited to compete long-term in the arms-race with DS? Eg, something inherent in its physics? (correct me if I’m wrong, but Moore’s Law would apply equally to both right?). Or perhaps just purely on economic and consumer led grounds, eg I rarely use any larger format camera than my iPhone these days as the algorithms have become sophisticated enough to be a respectable replacement - ok, we’re not their yet for purists, and still lots of other variables to consider -but in terms of extrapolating this technology to the future, not hard to imagine I’d say

I’m not qualified to explore this further in a useful way, but would appreciate thoughts/contributions from those of you who are! ..and my humble thanks for your time to have read this far

 

[dunring]

For the science of it, check the Audiosciencereview of the latest and very expensive Topping R2R model:
https://www.audiosciencereview.com/forum/index.php?threads/topping-centaurus-r2r-dac-review.59283/

For those, it's not about performance, the SINAD on them is the same as a Soundblaster G6 for $50. I've had a few hundred DACs for my business and the differences between them are really hard to discern. My world is value for money. The best best is a mid range SMSL DAC that has color and filter options. Color settings introduce distortion in the 2nd and 3rd harmonic, and filters chance the amount of pre and post ringing on a note. Between the two you can dial in your favorite sound fairly well when combined. The Topping or SMSL DACs with the tube or crystal settings are fairly close to a tube hybrid, but none have the analog warmth of a real one. I'd only recommend a R2R DAC if you've heard one, and really find it's worth buying.
I use a Topping A70Pro with an E70 9028Pro because it's flawless in performance, and the A70Pro uses new technology so it doesn't sound the same as their other models. I've had much more expensive stuff come through, but that pair is just personally the one I'll never upgrade. The SMSL H400 amp with the Gustard X30 DAC/streamer was mindblowing, but spendy for a DAC as a daily driver. It was the best thing I've ever heard though, but the Topping stack is much more practical.

 

[100Hz]

Thanks for the link, I’ve come across Audiosciencereview and respect their rigorous approach

 

[peterinvan]

Ok so I’m a late starter in hi-fi and my audio experience is limited, including tech…so I’ve been reading around and looking to tweak my existing humble set-up. In fact my old Hifiman DAP is on its way out which inspired me to catch-up on latest DAC trends

This rabbit hole led me to the Delta-Sigma (DS) v R2R debate -well rehearsed on forums I know, so excuse me for bringing it up here…but I’m interested on the science take to cut through the metaphoric analogies that have been gaining traction. I’m not saying they don’t have there place, and indeed some I found helpful, eg comparing a telephoto optical lens to digital zoom when talking about oversampling..But when descriptions start gushing with “natural” and “organic” I get concerned that what actually might be lower resolution sound is being romanticised. Forgive my own analogy, but I’m of the generation that grew-up with film photography and although digital was initially a poor rival, I quickly came to accept that the finest film grain was never going to compete with digital, ultimately. Subjectively I still love the characteristics of film but in resolution and dynamic range digital is undisputed. (I realise that SD v R2R is not a digi v analogue question but the holy grail seems to be that “anogolue” sound, natural and organic). Similarly there maybe characteristics of R2R that are sought for, but I chuckle when a negative attribute to one listener is spun (corrected?) into a positive one by another. Eg, for the same R2R DAC: “imaging has diffuse borders” (a weakness) whereas from another perspective, this is interpreted as, “softer rounder edges, more natural”. Without live experience I’m not qualified to judge, but I feel disappointed this can’t be expressed more neutrally leaving the reader to decide if they want that kind of characteristic/attribute - in this case I’m having to guess the sound has softer resolution (?).

And so going back to the science, I get the impression that R2R, albeit an older technology, has been making significant advances to rival mainstream DS DACs, but is it a technology, in theory at least, more limited to compete long-term in the arms-race with DS? Eg, something inherent in its physics? (correct me if I’m wrong, but Moore’s Law would apply equally to both right?). Or perhaps just purely on economic and consumer led grounds, eg I rarely use any larger format camera than my iPhone these days as the algorithms have become sophisticated enough to be a respectable replacement - ok, we’re not their yet for purists, and still lots of other variables to consider -but in terms of extrapolating this technology to the future, not hard to imagine I’d say

I’m not qualified to explore this further in a useful way, but would appreciate thoughts/contributions from those of you who are! ..and my humble thanks for your time to have read this far

I am in the "value for money" group. I have enjoyed both SABRE and AKM dongle DACs and desktop DACs (see sig. below).

I have settled on the FiiO K11R2R in OS mode driving a Schiit Vali3 amp.
With the Loki+ analogue tone controls I can tweak it for every headphone (and every album) to suit my old ears;

I am also in the "trust your ears" group and encourage you to test a relatively inexpensive R2R DAC against a similarly priced DS.

 

[100Hz]

What a cute set-up
Must post a pic of mine when I have time (Hifiman601 DAP > xduoo TA26s > B&O Beolab3 speakers and/or Grado60 headphones)
I did consider the V3 you have before going for the TA26

There is in fact a place in London that will have a good enough selection for audition, so on my list - but TBH I’m not even looking for purchasing advice right now, more interested in discussion and views -so thank you for sharing!

 

[Diet Kokaine]

If you like graphs, go with a chip DAC. Graphically they all sound the same, or so I'm told.
If you like music, go with a good R2R DAC.

 

[100Hz]

I’ve mostly seen flat frequency response graphs for the former.. and I’ve also read about being “true to the original recording”

By “music” do you mean a digital conversion that has a certain character, eg deeper bass, less treble (not only)? I personally can’t hear beyond 15kHz anyway..

Can you expand on what a musical sound means to you? Its gonna be subjective right? I’ve always enjoyed how music sounds from my HM601 which has an ancient chip by today’s standard

 

[gregorio]

In fact my old Hifiman DAP is on its way out which inspired me to catch-up on latest DAC trends …

I presume you mean audiophile “DAC trends”? In which case I’m a bit confused because you also state you’re interested in the actual facts/science and unfortunately, the two are at least somewhat mutually exclusive.

I get concerned that what actually might be lower resolution sound is being romanticised.

You’re right to be concerned. Typically an R2R DAC would be lower resolution than a DS DAC, the very best R2R DACs achieve a resolution that approaches the better DS DACs but none of this is about “lower resolution sound”, it’s about a lower resolution analogue output signal. In other words, the measurable differences in the analogue output signal of DACs are too small to be reproduced by transducers as sound (at a reasonable listening level), so there is no “lower resolution sound”, it’s the same resolution. There are a few exceptions, there are a tiny number of R2R NOS DACs and also one or two employing tubes on the analogue output stage which distort the output so much that the difference can not only be resolved into sound but actually audible sound in one or two cases.

And so going back to the science, I get the impression that R2R, albeit an older technology, has been making significant advances to rival mainstream DS DACs, but is it a technology, in theory at least, more limited to compete long-term in the arms-race with DS? Eg, something inherent in its physics? (correct me if I’m wrong, but Moore’s Law would apply equally to both right?).

R2R is an older, superseded and now defunct technology, it relies on very accurate matching of resistors in a ladder arrangement (a “resistor ladder”), which makes them much more expensive, more unstable (performance can vary with temperature for example), less reliable and less accurate than DS DACs. This is why I clarified “audiophile DAC trends” above, as opposed to just “DAC trends”: There is no trend towards DACs that are simultaneously more expensive, less accurate, less stable and less reliable except in the audiophile community. There are no professional or even prosumer R2R ADCs or DACs on the market, the last of them died out 20 or so years ago and most were gone around 30 years ago.

The reason R2R still exists in the audiophile community is two fold; it justifies a higher price and audiophiles can be convinced by marketing (incentivised reviews, shills, advertising) that it’s weaknesses are strengths, that inaudible lower performance/resolution is somehow audibly better or in those few cases where the lower accuracy is actually audible, that it’s somehow better fidelity. Watch out for that marketing, they’ve invented a lot of it over the years. For example the “trust your ears group” are typified by those who only trust their perception of their sight, biases and ears combined but flat out refuse to trust their ears. The filter pre/post ringing nonsense, that’s been proven to not even register in the brain’s auditory cortex, let alone be audible. The graphs/measurements vs music marketing is also nonsense, a DAC has no idea it’s even outputting music, it just converts digital data into an analogue signal, that’s why it’s called a DAC rather than a DMC (Digital to Music Converter).

Moore’s Law applies to the number of transistors in an integrated circuit (and therefore roughly to computing power), it’s not applicable to resistors or R2R DACs. It’s not really applicable to DS DACs either though, because the computing power necessary to convert digital to analogue is trivial, it was trivial for processors around 30 years ago, let alone today’s processors.

Graphically they all sound the same, or so I'm told.

Who has told you that? I’m guessing you deliberately made-up that falsehood because it doesn’t even make sense, how can something sound graphically the same, when a graphic is a visual (not aural) representation? In addition, the various graphs representing measurements of DAC performance demonstrate that DACs all perform differently. The graphs do NOT measure/represent sound at all, let alone that it’s the same sound as other DACs!

G

 

[2leftears]

Eg, for the same R2R DAC: “imaging has diffuse borders” (a weakness) whereas from another perspective, this is interpreted as, “softer rounder edges, more natural”. Without live experience I’m not qualified to judge, but I feel disappointed this can’t be expressed more neutrally leaving the reader to decide if they want that kind of characteristic/attribute - in this case I’m having to guess the sound has softer resolution (?)

These descriptions have no sensible meaning in context of sound, and are thus highly subjective interpretations. Everyone is guessing at what it means.
(EDIT: I often wonder if the people using these terms actually have any idea what it means themselves. I suspect a good number of them are simply parroting stuff they have heard other 'audiophiles' say).

If you like graphs, go with a chip DAC. Graphically they all sound the same, or so I'm told.
If you like music, go with a good R2R DAC.

Simply go with a good transparent DAC, which can be DS, or R2R, or hybrids, and doesn't have to be expensive.

If you like music, spend most of your time & money discovering & listening to music (new artists preferably, they deserve their turn and a chance at getting noticed, even if they do suck at their own marketing ).

 

[100Hz]

I presume you mean audiophile “DAC trends”? In which case I’m a bit confused because you also state you’re interested in the actual facts/science and unfortunately, the two are at least somewhat mutually exclusive.

You’re right to be concerned. Typically an R2R DAC would be lower resolution than a DS DAC, the very best R2R DACs achieve a resolution that approaches the better DS DACs but none of this is about “lower resolution sound”, it’s about a lower resolution analogue output signal. In other words, the measurable differences in the analogue output signal of DACs are too small to be reproduced by transducers as sound (at a reasonable listening level), so there is no “lower resolution sound”, it’s the same resolution. There are a few exceptions, there are a tiny number of R2R NOS DACs and also one or two employing tubes on the analogue output stage which distort the output so much that the difference can not only be resolved into sound but actually audible sound in one or two cases.

R2R is an older, superseded and now defunct technology, it relies on very accurate matching of resistors in a ladder arrangement (a “resistor ladder”), which makes them much more expensive, more unstable (performance can vary with temperature for example), less reliable and less accurate than DS DACs. This is why I clarified “audiophile DAC trends” above, as opposed to just “DAC trends”: There is no trend towards DACs that are simultaneously more expensive, less accurate, less stable and less reliable except in the audiophile community. There are no professional or even prosumer R2R ADCs or DACs on the market, the last of them died out 20 or so years ago and most were gone around 30 years ago.

The reason R2R still exists in the audiophile community is two fold; it justifies a higher price and audiophiles can be convinced by marketing (incentivised reviews, shills, advertising) that it’s weaknesses are strengths, that inaudible lower performance/resolution is somehow audibly better or in those few cases where the lower accuracy is actually audible, that it’s somehow better fidelity. Watch out for that marketing, they’ve invented a lot of it over the years. For example the “trust your ears group” are typified by those who only trust their perception of their sight, biases and ears combined but flat out refuse to trust their ears. The filter pre/post ringing nonsense, that’s been proven to not even register in the brain’s auditory cortex, let alone be audible. The graphs/measurements vs music marketing is also nonsense, a DAC has no idea it’s even outputting music, it just converts digital data into an analogue signal, that’s why it’s called a DAC rather than a DMC (Digital to Music Converter) and obviously, a DAC

Moore’s Law applies to the number of transistors in an integrated circuit (and therefore roughly to computing power), it’s not applicable to resistors or R2R DACs. It’s not really applicable to DS DACs either though, because the computing power necessary to convert digital to analogue is trivial, it was trivial for processors around 30 years ago, let alone today’s processors.

Who has told you that? I’m guessing you deliberately made-up that falsehood because it doesn’t even make sense, how can something sound graphically the same, when a graphic is a visual (not aural) representation? In addition, the various graphs representing measurements of DAC performance demonstrate that DACs all perform differently. The graphs do NOT measure/represent sound at all, let alone that it’s the same sound as other DACs!

G

Thank you so much, I value being corrected to improve my understanding (eg my so-called “sound resolution”) - very helpful. I did mean trends to broadly capture real technological developments but also what’s popular according to peoples’ current tastes (I think there is some correlation but obviously a subjective one, and I still find this interesting)

Also I admit to feeling the seduction of marketing, and also eloquent (some might say “poetic”) reviews, eg:

- and that’s exactly why I posted here, to get back some grounding

@2leftears - I’m indeed getting back into more music listening, and enjoying some smaller bands on the blue-grass scene. Have found https://archive.org/ to be a great resource for live music recordings as well as discovering artists off-the-beaten track..

Gearwise I’m going to be satisfied with replacing my failing 601 with another DAP for now - but still remain interested in following tech developments, improving my understanding of the science, as well as evolving tastes of younger generations

 

[XTF1]

Also I admit to feeling the seduction of marketing, and also eloquent (some might say “poetic”) reviews,

This is perfectly fine as long as you understand it is just that—marketing and poetic reviews (not measurement-based)…
If you think owning a R2R is cool and makes you happier, by all means, go for it !!! I suspect most of the R2R DACs currently available are good enough to be audibly transparent.
The point is to not trying to rationalize your choice with some mix of pseudo-science BS and subjective perception and then impose it as the absolute truth.

 

[71 dB]

If you like music, spend most of your time & money discovering & listening to music

Some people spent ridiculous amount of their money, time and energy on marginal at best improvements in sound quality when the joy and excitement of discovering new music you like is much more rewarding, at least for me.

(new artists preferably, they deserve their turn and a chance at getting noticed, even if they do suck at their own marketing ).

In my opinion music shouldn't be marketed like shampoo. That's too cynical for cultural products and it also leads to the situation where already big artists can afford much bigger marketing compared to smaller artists. Instead, (good) music should be promoted and advocated by people who have taste, understanding and knowledge about music.

If you think owning a R2R is cool and makes you happier, by all means, go for it !!!

If you happen to be American and planning to buy imported stuff, especially stuff made in China, better do it ASAP, because the prices will skyrocket very soon!

 

[castleofargh]

Ok so I’m a late starter in hi-fi and my audio experience is limited, including tech…so I’ve been reading around and looking to tweak my existing humble set-up. In fact my old Hifiman DAP is on its way out which inspired me to catch-up on latest DAC trends

This rabbit hole led me to the Delta-Sigma (DS) v R2R debate -well rehearsed on forums I know, so excuse me for bringing it up here…but I’m interested on the science take to cut through the metaphoric analogies that have been gaining traction. I’m not saying they don’t have there place, and indeed some I found helpful, eg comparing a telephoto optical lens to digital zoom when talking about oversampling..But when descriptions start gushing with “natural” and “organic” I get concerned that what actually might be lower resolution sound is being romanticised. Forgive my own analogy, but I’m of the generation that grew-up with film photography and although digital was initially a poor rival, I quickly came to accept that the finest film grain was never going to compete with digital, ultimately. Subjectively I still love the characteristics of film but in resolution and dynamic range digital is undisputed. (I realise that SD v R2R is not a digi v analogue question but the holy grail seems to be that “anogolue” sound, natural and organic). Similarly there maybe characteristics of R2R that are sought for, but I chuckle when a negative attribute to one listener is spun (corrected?) into a positive one by another. Eg, for the same R2R DAC: “imaging has diffuse borders” (a weakness) whereas from another perspective, this is interpreted as, “softer rounder edges, more natural”. Without live experience I’m not qualified to judge, but I feel disappointed this can’t be expressed more neutrally leaving the reader to decide if they want that kind of characteristic/attribute - in this case I’m having to guess the sound has softer resolution (?).

And so going back to the science, I get the impression that R2R, albeit an older technology, has been making significant advances to rival mainstream DS DACs, but is it a technology, in theory at least, more limited to compete long-term in the arms-race with DS? Eg, something inherent in its physics? (correct me if I’m wrong, but Moore’s Law would apply equally to both right?). Or perhaps just purely on economic and consumer led grounds, eg I rarely use any larger format camera than my iPhone these days as the algorithms have become sophisticated enough to be a respectable replacement - ok, we’re not their yet for purists, and still lots of other variables to consider -but in terms of extrapolating this technology to the future, not hard to imagine I’d say

I’m not qualified to explore this further in a useful way, but would appreciate thoughts/contributions from those of you who are! ..and my humble thanks for your time to have read this far

It's the same stuff taken by different ends. The original DS concept is 1bit with very high sample rate(oversampled inside the DAC), so the linearity is great, but there is an enormous amount of noise to move and filter. It works, but limiting everything to 1 bit doesn't feel like the best option. We could work on 2 axes, and we deliberately remove one, then scramble to "noise shape" the consequence of that choice(like with DSD).

R2R has discrete resistors to output voltages corresponding to each bit( each following bit needing an output of half the voltage of the previous one). So long as you don't fall for extreme choices that are objectively bad on purpose, like non-oversampling or ineffective low pass filters(if there is one at all), it also works. The main issue this time is to match all the resistors as precisely as possible, otherwise linearity sucks and the resulting errors are correlated to the bits which is not something desirable(although I'm sure someone, somewhere argues that it is. Up is down and good is bad).

Now, the very common delta sigma chip nowadays doesn't use any of those pigheaded extreme designs. They use delta sigma, and a few bits(depends on the brand, but let's say between 3 and 6). That way they avoid the most extreme noise issues of one bit everything, and also mitigate the linearity issues by not having to match anything huge like 16 or even 24 bit. To me, that's the common sense, problem-solving middle of the road choice that most engineers would probably pick when facing the same questions. I feel like DACs were solved long ago, but that way they are even more solved.
So obviously the market is getting creative. Many chips now have a bunch of silly filter and oversampling choices, in case you are allergic to good things that simply work great. Some are also back to working in true 1bit like in the old days we should not try to go back to, some can simulate the issues of a filterless NOS DAC, hurray!!!! The future has plenty of room for unnecessary choices.

 

[JamoBroGuy]

If you think owning a R2R is cool and makes you happier, by all means, go for it !!!

Couldn’t have said it better. A lot of the time in places like Head-Fi SS and ASR, I get the vibe of “(insert here) isn’t better than an Apple dongle, so don’t buy it and if you have it you shouldn’t enjoy it, get rid of it and get an Apple dongle instead”. Of course the Apple dongle thing is just an example, but you get the point. Hobbies don’t have to make sense. Why do people collect stamps and watch birds? Those, objectively, don’t really make sense. But it’s a hobby, it doesn’t HAVE to.

Here are 2 more examples with gear I happen to own.

I like owning a dongle with a ROHM DAC. I think it’s cool. Is it objectively better than an Apple dongle? Absolutely not, both are audibly transparent. Would I take it over an Apple dongle every day of the week? Absolutely.

I like owning an IEM with 14 BAs per side. Is it objectively better than a single DD? No, a single DD can reproduce the entire frequency range without issues. Would I take it over a single DD? Heck yeah.

You get what I mean? The issue isn’t people enjoying these “snake oil” (I kind of hate that term lmao) products. The issue is people claiming that they’re objectively better when in reality it’s just some intangible quality due to bias (which CAN and SHOULD improve your listening experience, by the way, we listen with our brains after all, as long as you RECOGNISE that it isn’t an objective measurable quality and don’t try to push that agenda).

 

[gregorio]

Thank you so much, I value being corrected to improve my understanding (eg my so-called “sound resolution”) - very helpful.

You’re welcome, glad it was useful. The “sound resolution” thing is the dirty little secret of audiophile marketing. To explain it better, sound has a physical limitation: Given a hypothetically perfect anechoic chamber (IE. Perfect isolation from exterior sound/noise), no sound occurring within the chamber and therefore effectively perfect silence, the sound pressure level would be about -23dB SPL (sound pressure level), which is effectively the noise created by the random collisions of the air molecules within the chamber. It is impossible to have sound level lower than this (in air). If you look up the measurements of DACs, you’ll typically see a dynamic range of around 120dB, maybe a bit higher on really expensive units and maybe only 110dB or so on cheaper units, you’ll also see measurements of jitter down around -115dB but expensive ones down as far as -160dB, distortion also around those ranges. Last fact in the equation; typically people will listen to headphones at a peak level well below 80dB SPL, if you like it really loud or listen to really dynamic orchestral recordings, maybe as high as 85dB SPL, beyond that is not a “reasonable listening level” because it can cause hearing damage.

Hopefully you can see the issue, taking the highest reasonable listening 85dB SPL and a cheaper DAC with artefacts at say -110dB, 110dB below 85dB SPL would be -25dB SPL but that’s impossible, there can be no sound below -23dB SPL. So these reports of better sounding dynamics, less audible jitter, more resolution, etc., are effectively claiming to hear something that doesn’t even exist as sound, let alone is audible. Some things could theoretically exist as sound, some can even be resolved by transducers into sound and even on occasion enough sound to actually be audible but that’s very rare these days. Only a tiny number of tube or NOS DACs fit into this category, even then, only in some circumstances and cheap DACs don’t. Taking the example of say a $9 Apple Dongle, in typical usage it’s artefacts (distortions/noise) can’t exist as sound, given the highest reasonable listening level they could exist theoretically but it’s doubtful whether even the most resolving headphones could actually resolve them into sound but even if they could, they would be well below the threshold of human hearing.

I did mean trends to broadly capture real technological developments but also what’s popular according to peoples’ current tastes (I think there is some correlation but obviously a subjective one, and I still find this interesting) …

I’m not quite sure I understand what you mean. I’m sure some people’s “current tastes” are for the best performance on paper, regardless of the fact it’s inaudible. In most cases though, according to posts on this site, most people are basing their “tastes” on their belief of actual audible differences. Those differences either literally don’t exist (as sound) or are inaudible, with some very rare exceptions. What they’re actually reporting is the result of their perception biases, which is why they state they “trust their ears” but eschew the test methodology that specifically eliminates those biases and only leaves what their ears “hear”.

G