Can someone explain to me what the difference is between multibit dacs like Ifi Zen Dac or Schiit Bifrost and R2R dacs?

[gregorio]

That’s probably the least sophisticated deliberate mis-quote I’ve ever seen in this subforum and that’s saying something!

G

 

[DarginMahkum]

The biggest difference (technical implementation aside) is the marketing. R2R also gives the audio device developers option to make in house development and not depend on a 3rd party, especially during a chip crises. And then they use it for advertising purposes, as if it is a groundbreaking new and superior technology. It gives them the possiblity of selling one more DAC to those caught in the marketing hype.

DAC technology as an off the shelf chip is already capable of perfectly reproducing the signal from digital data, any imperfections far beyond the limits of human hearing.

 

[bigshot]

That’s a bingo.

 

[Sheep3000]

That’s the whole point of marketing! It’s not that you necessarily believe everything claimed but that it creates a subconscious bias that affects one’s experience. Science has known this for a century or more, marketers themselves have obviously known this for at least that long and even much of the public knows this but for some reason many audiophiles do not know and actively fight against knowing. So, not only do they completely believe they’re manipulated/biased experiences but they have no intention of being “cured” of their ignorance. They’ve even invented their own term for it; “subjectivist”, and naturally discredit science, objective measurements, controlled (blind) testing and anything which demonstrates their experiences are biased.

Nope, while components and implementation does vary, this typically does NOT result in audible differences, particularly with solid state units. However, audiophile marketing of course has to suggest otherwise, so audiophiles often have resultant biased experiences and “hear” differences in almost everything, fuses, capacitors, DAC chips, power cables to Ethernet routers and pretty much anything else. The reality is, that it’s typically only with mechanical components (transducers) that necessarily have low efficiency and accuracy, where differences are significant enough to be audible.

Exactly, and DAC chips are of course “on a technical level all doing the same thing” and “are basically interchangeable” and “indistinguishable”, except in one sense, the marketing is not indistinguishable!

Absolutely, audiophile fuses, cables or whatever are presented very differently to standard equivalents. For starters, they typically look significantly different (which is enough on it’s own) but obviously the price and marketing is also very different.

G

I also know a lot of manufacturers genuinely perceive differences in their own equipment as well. Also I'm pretty sure true multibit r2r dacs don't use as much math to replicate the signal as delta-sigma dacs, right?

 

[gregorio]

I also know a lot of manufacturers genuinely perceive differences in their own equipment as well.

If that’s really true, it’s rather shocking. Personally, I would expect a manufacturer to have a very good understanding of what it is they are designing/engineering, are there really “a lot of manufacturers” who don’t? My guess is there’s very few who don’t but a lot in the audiophile market who make marketing assertions that are not in line with their actual understanding.

Also I'm pretty sure true multibit r2r dacs don't use as much math to replicate the signal as delta-sigma dacs, right?

That kind of depends on what you mean by “math”. If you’re using electronic components to implement mathematical functions on a signal, then that’s still math. However, that’s not quite the same as just using a chip to perform math on numbers (binary data). In the former case there are component tolerances to contend with and therefore a certain level of inaccuracy, which is not an issue with a chip just calculating numbers. This is why chip based DACs are generally more accurate/higher fidelity than R2R DACs and why certain R2R DACs that are very high fidelity actually also employ chips.

G

 

[DarginMahkum]

I also know a lot of manufacturers genuinely perceive differences in their own equipment as well. Also I'm pretty sure true multibit r2r dacs don't use as much math to replicate the signal as delta-sigma dacs, right?

We keep coming back to this "hearing differences" topic. Assuming there's an audible difference: You can always inject differences into your design. Some intentional tweaks here and there in the circuit to justify "the difference", or unintended differences as the designer of the circuit lacks the skills. In any case: A DACs purpose is to reproduce the original signal with the least amount added imperfections. Period. If a DAC is not able to do that, it does not deserve its value. We have every kind of proof that current DS DAC chips are able to achive this. If a signal deviates from the original signal when it reaches the transducer, it was altered after it left the DAC chip, either intentionally or unintentionally. In case of R2R, a specific R2R design might produce imperfections, due to the component tolerances etc., but that doesn't make them a good product if the produced signal deviates from the original signal in an audible way.

If a company is reinventing the wheel to replicate what a simple DS DAC based design can do and charging above its price, the extra money is wasted - assuming all other features of the unit are the same.

That brings me to this: As a music lover, I value the pure replication of the original signal. How it was achieved (DS, R2R etc.) is of secondary value. I do not value them based of how it achieves the main goal. As this goal can easily be achieved today with very simple off-the-shelf chips, there's absolutely zero reason to pay a different technology more just because it includes some buzz words.

Math: Electronics is math. Whether you are designing a circuit or doing digital signal processing. I remember filling up 7-8 pages of mathematical analysis in my second year circuit theory exams for 2-3 questions / simple circuits only. That didn't change later on for DSP. What happens is, you learn along the way, use circuit simulations etc., and might use less math. But math is always there, though DSP is more math intensive. I am unfortunately not doing any analog circuit development and don't have audio experience, but the basics don't change.

 

[caracara08]

great posts here with great info. Thank you.

 

[Sheep3000]

We keep coming back to this "hearing differences" topic. Assuming there's an audible difference: You can always inject differences into your design. Some intentional tweaks here and there in the circuit to justify "the difference", or unintended differences as the designer of the circuit lacks the skills. In any case: A DACs purpose is to reproduce the original signal with the least amount added imperfections. Period. If a DAC is not able to do that, it does not deserve its value. We have every kind of proof that current DS DAC chips are able to achive this. If a signal deviates from the original signal when it reaches the transducer, it was altered after it left the DAC chip, either intentionally or unintentionally. In case of R2R, a specific R2R design might produce imperfections, due to the component tolerances etc., but that doesn't make them a good product if the produced signal deviates from the original signal in an audible way.

If a company is reinventing the wheel to replicate what a simple DS DAC based design can do and charging above its price, the extra money is wasted - assuming all other features of the unit are the same.

That brings me to this: As a music lover, I value the pure replication of the original signal. How it was achieved (DS, R2R etc.) is of secondary value. I do not value them based of how it achieves the main goal. As this goal can easily be achieved today with very simple off-the-shelf chips, there's absolutely zero reason to pay a different technology more just because it includes some buzz words.

Math: Electronics is math. Whether you are designing a circuit or doing digital signal processing. I remember filling up 7-8 pages of mathematical analysis in my second year circuit theory exams for 2-3 questions / simple circuits only. That didn't change later on for DSP. What happens is, you learn along the way, use circuit simulations etc., and might use less math. But math is always there, though DSP is more math intensive. I am unfortunately not doing any analog circuit development and don't have audio experience, but the basics don't change.

Thank you for the explanation! I recently bought a schiit modi multibit 2 to see if i like these stated "imperfections". If I want an accurate DAC, I'll just use my motu m2! I feel like a lot of the reason people like r2r/multibit stuff is for the nostalgia/old-school sound (said imperfections you mentioned, I presume). I'll attempt to do a true blind volume matched a/b/x test with a family member's assistance to see if I can actually tell the difference between the two. Any suggestions on how I'd go about matching the volume levels?

 

[gregorio]

I recently bought a schiit modi multibit 2 to see if i like these stated "imperfections".

You’re going to need some quite sophisticated/expensive analysis equipment in order to see “these stated imperfections” and then you’ll have to invent some criteria for what you “like”, because these imperfections are typically below or way below audibility. However, the Modi 2 is an interesting choice because AFAIK it’s one of the worst performing DACs out there, with more than one problem area. What makes it interesting is that even with such relatively poor performance, it’s artefacts will still be largely inaudible in most cases (though potentially not all).

I feel like a lot of the reason people like r2r/multibit stuff is for the nostalgia/old-school sound (said imperfections you mentioned, I presume).

What “old school sound”? Maybe those few that have dodgy tubes inserted after the DAC but otherwise the audible sound will generally be the same.

I'll attempt to do a true blind volume matched a/b/x test with a family member's assistance to see if I can actually tell the difference between the two. Any suggestions on how I'd go about matching the volume levels?

As a simple, very rough method: First play the loudest parts of some of your favourite recordings and set your amp so the volume is moderately loud. Use a test signal such as a 1kHz sine, and measure the amp’s voltage output, this is your baseline. Use the same test signal with your other DAC and lower or raise the amp until its output voltage measures the same as the previous DAC. Switch between your baseline and other DAC setting when listening to the appropriate DAC. You can use an ADC to measure the output level of the amp (in dB) rather than voltage if you don’t have an accurate multi-meter. You don’t want the difference in level to be much higher than 0.1dB.

Note, depending on your amp/volume control, it might not be possible to switch between baseline and the other setting accurately enough and the switching obviously needs to be done by your assistant otherwise it won’t of course be a blind test. Ideally, one would use an actual ABX switch box, to cut out the potential errors when using an assistant, but that’s not cost effective for the average interested consumer.

G