R2R or single-IC for DAC?

Jump to Latest
B

batteryman

Member
Joined 2011
wafflesomd said:
I personally have too much knowledge on how our hearing works and how to do proper testing to hear for differences and biases involved to ever confidently make such a decision just based on what I hear.

Some days my system sounds great, other days it sounds bad. I know nothing has changed and have learned to identity when it's my mental state that is accounting for the difference. Differences in the quality of sleep I get can make a pretty big impact here. I think people in general greatly underestimate how big of an impact mental states and bias can affect things.

I'm not sold on r2r, it just feels to much like the burn in debate, claims are always the same, more musical, less or no "digital glare", etc... Never heard anyone say they thought their r2r sounded bad and all that makes me highly skeptical.
Click to expand...
I think knowing about the mechanics of how our hearing works, means nothing because it is the brain's interpretation that affects our perception and no one knows how that works.

However, I do agree that how we feel affects this perception. Anyone who suffers migraines will know that and smoking certain herbs enhances perception. (many many years ago I did try some and found myself standing on the conductors podium and being able to focus on each instrument - it was so involving)

If you haven't listened to a r2r dac, you aren't qualified to comment and you can't claim they all sound the same.
There is no glare on chiurutu's and no aliasing artefacts that I can hear.
 
Z

Zoran

Member
Joined 2004
zorandimovski.com
MarcelvdG said:
I have no practical experience with using R-2R DACs for audio, only for other applications, but whether you will hear the effect of resistor tolerances and glitches in a pure R-2R DAC could depend on a lot of things, like

-Whether you use digital volume control or listen to music with very soft passages.
-At what volume you listen.
-How accurate the resistors in the most significant sections really are; a 0.1 % tolerance resistor is guaranteed to be within 0.1 %, but it can be much more accurate if you are lucky. (A former colleague once gave me a bunch of 5 % tolerance metal film resistors. I measured some, those were all within 0.01 %.)
-How good your ears are.
Click to expand...
.
About the resistor I had the almost the same experiance. But that was some "pro" series resistors, sort of special purposes, and from the same big cardboard package, same series. And of course some older production.
I dont think that it could be ex[ected to happend with some contemporary pieces? No mateer of cost...
.
I think that any digital decreasing of volume will make the sound outcome worse with R2R dacs. Also I am close to think that any CPLD or FPGA processing causing additional HF noise that affect mid-high spectrum.
Somehow after some experiments I doubt that digital ICs, with mosfets at the output, is good enough for directly driving R2R ladder? Maybe that it should be driven by some transistor circuit. Even apply some isolation if it is possible?
 
M

Markw4

Member
Joined 2016
rfbrw said:
...always thought of psychology as one the more dubious sciences not unlike economics.
Click to expand...
Both started out as primarily theoretical. The former has become much more experimentally derived (although not without some growing pains - i.e. replication crisis), while the latter field has become much more computationally oriented (using fairly advanced mathematics). Kahneman and Tversky in particular were known for their scientific approach and for experiments with solid statistical analysis. For one example, Kahneman measured pupil dealation as in indicator of effortful (slow) thinking. He could thus tell when a test subject finished solving a problem or else gave up. OTOH, Kahneman had no formal training in economics (thus might be considered amateur in that sense). Despite the lack of formal training he still managed to be awarded the Nobel Prize in the field of Economics. Computational models in economics now need to allow for Kahneman's discovery and experimental proof of risk aversion bias.
 
Last edited:
  • Like
Reactions: 1 user
M

Markw4

Member
Joined 2016
I decided to inquire with ChatGPT for an AI opinion (just for fun). Here is what it said:
"The claim that the primary function of computer modeling in the banking sector is to perform financial "onanism" is not accurate. Computer modeling in the banking sector is used for a variety of purposes, including risk management, financial forecasting, and portfolio optimization. These models help banks make informed decisions about lending, investment, and other financial activities. While some models may be used for speculative or non-essential purposes, the primary function of computer modeling in banking is to support responsible and profitable business operations."

With that response in hand I was emboldened to further inquire about what computational economics is actually used for (again just for fun). Here is the response I received:
"Computational economics is a field that uses computational methods and tools to study economic problems. It is mainly used for:

  1. Building models: Computational economics is used to build mathematical models of economic systems. These models help economists understand how different factors, such as changes in government policy or shifts in consumer demand, affect the economy.
  2. Simulation: Computational economics is also used to simulate the behavior of economic systems, which allows economists to predict the impact of different policy choices or market conditions.
  3. Optimization: The computational methods also allow for optimization of certain economic policies, such as to find the optimal taxes, subsidies or other economic policies that would lead to a desired outcome.
  4. Empirical analysis: In recent years, computational economics also has been used to analyze large datasets to uncover patterns and relationships in the economy that are not visible through traditional statistical methods.
Overall, computational economics is a powerful tool that helps economists to gain a deeper understanding of economic systems and make more informed policy decisions."

Now I must ask, do you guys think it is time to get back to the subject of audio dacs? :)
 
M

Markw4

Member
Joined 2016
With some more information I might decide whether its worth going on a test drive. If I have to buy a dac and pay for shipping, then return it for a refund and pay return shipping just to take a listen, I would like to know if I think its likely worth the time, effort, and cost.

Ideally maybe it could be more like this: I know if I want to buy some wine, I know which reviewer has tastes similar to mine and the way he describes the ones I like. Often they are only available by mail order and not returnable. Works well, once figured out.
 
You must log in or register to reply here.
Top Bottom