It occurred to be me before to try adding digital noise in DSP to possibly mask some of the low-level imperfections of digital audio. Never got around to trying it though.
Thinking about it a little more, maybe better to try in the analog domain with something like a zener noise generator, maybe some noise shaping...
Might improve the sound in more ways than just masking, given a possibility of stochastic resonance benefits. Only way to find out would be to try it.
Last edited:
That's a technical detail, my favorite is "One thing that seems to be missing from the best digital I have heard is the same level of room sound that can be heard on a good record. DSD seems to help with reproduction of room sound." which comes into collision with the First Principles, the digital format adds information to a digital stream.
Oh my God!
I'm still puzzled, when reading a post like this, as it violates the rules of logic and further neglects everything we know about psychoacoustics.
Sound quality by definition relies on the judgement by human listeners and any conclusion about "collision with the first principle" depends on the correctness of the premises.
Hi,
Mark, you could try it in digital. This is called 'noise shaping' and is included in every pro software able to perform a change to quantification ( dithering) characteristic.
For mastering duty i have a collection of plug ins or soft for that as the effectiveness depend from the source material ( Apogee's UV22hr, Ozone mbit+, Airwindows NJAD,...).
Anatech, i'm impressed you are still open to objective evaluation of such things as a by-by (or bee bee or buzz buzz whatever). If you weren't located on the other side of the planet i think i would bother you regularly to learn a thing or two.
Mark, you could try it in digital. This is called 'noise shaping' and is included in every pro software able to perform a change to quantification ( dithering) characteristic.
For mastering duty i have a collection of plug ins or soft for that as the effectiveness depend from the source material ( Apogee's UV22hr, Ozone mbit+, Airwindows NJAD,...).
Anatech, i'm impressed you are still open to objective evaluation of such things as a by-by (or bee bee or buzz buzz whatever). If you weren't located on the other side of the planet i think i would bother you regularly to learn a thing or two.
Last edited:
Through accredited course?
It might or might not. Let us know after you try it.
Yes, they probably used the term expectation bias, but, guess what, they knew what was meant by it, pedants excepted.
Very strange.
I have owned a Naim CD3 for 25 years and still today when I turn it on I find it very pleasant to listen to.
It's not torture, I can enjoy music for hours without getting tired.
Of course, in my subjective opinion.
But even for those who just believe in measurements it's not too bad for a 16-bit DAC.
Not much worse than the Mark Levinson 360S that the Lord of measurements has been using for 21 years.
Except that he eventually measured it and, according to his own measurements, he noticed that the MEIZU dongle (45 USD) measures better and so he has replaced it.
The world of audio is really full of cheaters as the Lord of measurements has even managed to cheat himself for 21 years.
I am interested in Psychology too. And I know about theoretical cognitive biases. I think I have seen in this forum someone mentioned that nobody is free from these biases. So, how do we know that we are free of it to a certain degree? Difficult question, isn't it?
I thought I knew all about it but not. I had a deeper understanding of expectation/confirmation biases in stock market where a bias can cost me money. I had to learn to really free myself from this kind of biases. Not only in stock analysis but also in life (because I then understood the cost of having this bias).
Not saying it sounds bad or anything like that. Kastor L seemed to have an issue with the proponents of the PCM63 and I was pointing out that the followers of the TDA1541 are far more dedicated to their idol.
Oh my God!
I'm still puzzled, when reading a post like this, as it violates the rules of logic and further neglects everything we know about psychoacoustics.
Sound quality by definition relies on the judgement by human listeners and any conclusion about "collision with the first principle" depends on the correctness of the premises.
If you believe a program converted from PCM to DSD can provide a different information to a listener, then you are once again preaching bull chips. A reading in the thermodynamics principles would help.
I could understand if a program would be originally recorded and made available in both PCM and DSD, but this never happens for obvious reasons.
I believe you are not interested in any serious discussions, but only have a weird need to stir the (subjective) pot. The fact that the brain can add or remove information from a program, and the reasons and circumstances in which the brain is doing it, doesn't make it a true, reproducible fact. If you want to study the ways the brain can be tricked, you go ahead, but don't conflate this with the reality.
Hi N101N,
The only valid comparison that is useful to the buying public, or to the experimenter, is how accurate the device is, and how much noise exists in the output signal. Of course, reliability is also a factor, so it would be helpful to know if it is tube based anyway.
Beyond that, the device either sounds good, or not depending on the subject. But at least that person can narrow the field of equipment based on price and preference. They may care about technology, that's fine but not always accurate.
For example, you can have multiple units out that use the same op amp, or DAC (for example), yet they all sound different. Some can sound awful while others are the best thing ever heard - all using the same devices. Board layout, component choices and actual schematic all play a part. But at the end of the day, performance is what matters most to a buyer. For the experimenter, their execution of a design makes all the difference in the world.
As a side note, over the years I have found that the better something is, the more people like it. So high performance clearly matters a great deal. You want the most valid information you can get to make a choice and you can't listen to everything. Measurements help you narrow down the field of things to listen to. You then have half a hope of buying something you will like for the long term.
-Chris
You are absolutely correct! And as such, those issues remove themselves from our concern. That entire argument is simply noise in this discussion.
The only valid comparison that is useful to the buying public, or to the experimenter, is how accurate the device is, and how much noise exists in the output signal. Of course, reliability is also a factor, so it would be helpful to know if it is tube based anyway.
Beyond that, the device either sounds good, or not depending on the subject. But at least that person can narrow the field of equipment based on price and preference. They may care about technology, that's fine but not always accurate.
For example, you can have multiple units out that use the same op amp, or DAC (for example), yet they all sound different. Some can sound awful while others are the best thing ever heard - all using the same devices. Board layout, component choices and actual schematic all play a part. But at the end of the day, performance is what matters most to a buyer. For the experimenter, their execution of a design makes all the difference in the world.
As a side note, over the years I have found that the better something is, the more people like it. So high performance clearly matters a great deal. You want the most valid information you can get to make a choice and you can't listen to everything. Measurements help you narrow down the field of things to listen to. You then have half a hope of buying something you will like for the long term.
-Chris
You have quoted the first part of my post but the most important is the last.
Any comment?
No one is free from all cognitive biases. Some are probably impossible to overcome. Some can be overcome fairly easily, at least if one remembers to think about them. Anchoring bias is an example of one that is fairly easy for most people to overcome if they remember it will influence them unless they consciously overcome it when forming judgements relative to a preexisting anchor point.
Last edited:
Nobody claimed information was being added. That is an assumption on your part.
If anything, some dacs seem to produce less audible linear and nonlinear distortion artifacts when operating in DSD mode. So, using DSD may reduce 'noise,' where the term noise is used a broad sense meaning any audible unwanted signal aberration.
Pants on fire. Before making such statements, please go back and read again what you posted:
Otherwise, you can't use the concept of "noise" to your convenience. "Noise" has a precise definition, no room for subjective BS peddlers, sorry.
"Noise is a signal we don't like," Bart Kosko, Noise.
Dr. Bart Kosko is a Professor of Electrical and Computer Engineering in the University of Southern California's Viterbi School of Engineering and a Professor of Engineering and Law in USC's Gould School of Law.
Dr. Kosko is a fellow of the IEEE, a fellow of the International Neural Network Society (INNS), and a fellow of the International Fuzzy Systems Association (IFSA).
Dr. Bart Kosko is a Professor of Electrical and Computer Engineering in the University of Southern California's Viterbi School of Engineering and a Professor of Engineering and Law in USC's Gould School of Law.
Dr. Kosko is a fellow of the IEEE, a fellow of the International Neural Network Society (INNS), and a fellow of the International Fuzzy Systems Association (IFSA).
- Status
- Not open for further replies.