[BS5711]

If "he" is AI do they give up when exposed ?

 

[knownothing2]

I wonder if knownothing2 was trying to tell us something with the username and this is all a joke, most audiophiles would have given up the fight long ago.

I took on this name on audio forums years ago because I look at the hobby as a constant learning process. I definitely know more than I did 20 years ago. I also know how much more I don’t know now than I did then. On this forum I am in the uncomfortable position to be questioning people with much greater expertise in specific technical elements related to audio. As a trained scientist used to evaluating complex systems, it’s a curious role for me.

I do think it might be helpful to do some serious work to bridge the gapping canyon between avid audio listeners who depend on their ears and people with technical backgrounds who are driven primarily by theory and practice as professionals. This if only to more productively exchange information and outlooks, even if the flow of information is mostly one way to the great unwashed.

I support the suggestions throughout this thread challenging people who claim they hear differences in sound stage and other audible affects and attributing them to specific drives, servers/streamers, DACs and even digital cables, to prove that it isn’t just observer-expectancy effect by running ABX/DBTs. I do believe in bias, and I know that I am biased. But I suggest not every audio enthusiasts is as smitten and deluded by marketing and their own expectations that they cannot hear differences. And some claim the differences are so clear and repeatable that they should withstand rigorous testing. Let’s see.

You are wondering why I don’t give up? Several reasons.

-I have played with equipment a lot and when I suspect a difference, I have done limited blind testing with digital cables and digital gear, and the subjects support the difference when one in fact exists.

-I am learning through the exchange on this thread, but having been an academic, I am both sure of my opinions and a competitive arguer, and for that I apologize.

-Especially in this last discussion about the Node, some/many of the answers have frankly seemed deeply circuitous, vague, and dismissive, with no one directly addressing the question of switching power supply noise in the Node potentially scaling with demand from the drive and potentially becoming problematic for the digital output in audibly noticeable but non-catastrophic ways. The general answer being “it’s digital, can’t happen”, or something like “the company didn’t specify what you’re saying could happen specifically and it’s not my job to connect your dots for you”. That is supremely unsatisfying - time for a test, right?

-Finally, the level of passion dedicated to attacking folks who claim to hear audible differences in systems level outcomes borders on dogma and is brought with the same religious fervor projected here on “audiophiles” with great distain. But no-one’s perfect.

kn

 

[knownothing2]

If "he" is AI do they give up when exposed ?

No.

 

[BS5711]

But I suggest not every audio enthusiasts is as smitten and deluded by marketing and their own expectations that they cannot hear differences. And some claim the differences are so clear and repeatable that they should withstand rigorous testing. Let’s see.


Finally, the level of passion dedicated to attacking folks who claim to hear audible differences in systems level outcomes borders on dogma and is brought with the same religious fervor projected on “audiophiles” here with great distain.

The first point is the critical one, the claims of hearing things are always only claims. I doubt the claims because my own assessments tell me otherwise. I can be convinced I hear changes with something until I test and retest and discover that the difference was more often than not down to not volume matching. Intentionally change the volume either direction and the results are skewed the other way. How many folks claiming that the more powerful amplifier drives a tiny IEM with more dynamics and authority actually volume match ? I would say none.

The second point is entirely valid but it is a two way street, out in the wider Head Fi threads me venturing an opinion that the differences someone talks of are not in the change in amplifier for example get shot down with fervor and the inevitable "just because your hearing isn't good enough" and the "your gear isn't resolving enough" mantra.

But I take your point, most folks are generally pretty polite here even if they get frustrated with the repetitive back and forth.

 

[BS5711]

No.

Touché

 

[knownothing2]

“AI” could be considered a compliment based on the name, or a cut based on the current performance level of the tool. I am guessing it’s a cut in this context, but I assure you I am generally hallucination-free

 

[Davesrose]

If "he" is AI do they give up when exposed ?

Given *his post* right after this....I would say it's going to stay relentless. With bigshot's theory that it could be using this thread as gathering ammunition for other areas of Head-Fi, I notice almost all *his* activity is this thread. I suppose since AI is just computer algorithms that can run 24/7, it can spend all its cycles scraping this one thread: then during a certain time or when no one is paying attention to it, it will move on to another thread. Edit: also notice *he* simply replied "no" to the question "If "he" is AI do they give up when exposed ?". Maybe that's a response the developer(s) intentionally added as an Easter Egg for a wink wink.

 

[KinGensai]

@sander99
Bias refers to heuristic shortcuts the psyche uses when interpreting data, so perceptual errors fall into the definition, but is just one example. Confirmation bias is the other big one relevant to this topic.

 

[Davesrose]

As a developer myself....I've also been wondering if this "Node" DAC is also an easter egg: Node.js has become a popular web application framework. It's entirely possible there's an AI running on it. I'm thinking this also maybe a wink wink.

 

[BS5711]

As a developer myself....I've also been wondering if this "Node" DAC is also an easter egg: Node.js has become a popular web application framework. It's entirely possible there's an AI running on it. I'm thinking this also maybe a wink wink.

"He" is very persistent and unerringly polite.

 

[Davesrose]

"He" is very persistent and unerringly polite.

I will give credit in how polite and unobtrusive *he* has been. I know I spent too much time getting into the weeds about what specs the Node DAC has, how there's no documentation of a sound difference with USB devices, and how conceptually it's all bit perfect. Now there's been a hard reset with the "arguments" about how noise doesn't create a fault but somehow alters a digital audio file's sound. Let's just see how persistent it is if we don't engage.

 

[VNandor]

It has been explained over and over again that why the "timing" or "jitter" doesn't matter for this case. It is just some digital data sent over the USB. How and when the data is sent doesn't play any role in when the audio sample gets converted (except when it doesn't get sent), they are completely separate. The timing of the conversion is being controlled by the DACs own clock, it's not somehow encoded in the data sent over USB. If the data is somehow not sent by the time the clock triggers the conversion, the output will be made from whatever garbage filled the memory (instead of the audio samples' value) at the time of conversion. The clock not being perfect is what causes jitter. The audio samples arrive in just a fraction of the sampling time, the data just sits there in a queue waiting to be converted 16bits by 16bits every 1/48000 seconds.

 

[bigshot]

I take it back. He’s a troll. An ordinary person would be embarassed.

There’s no way to move forward without accepting ABX testing. That’s step one and all the rest build upon it. I don’t feel bad about dismissing people as soon as they say they trust their subjective impression and don’t believe in blind testing.

 

[knownothing2]

“

The timing of the conversion is being controlled by the DACs own clock, it's not somehow encoded in the data sent over USB. If the data is somehow not sent by the time the clock triggers the conversion, the output will be made from whatever garbage filled the memory (instead of the audio samples' value) at the time of conversion. The clock not being perfect is what causes jitter. The audio samples arrive in just a fraction of the sampling time, the data just sits there in a queue waiting to be converted 16bits by 16bits every 1/48000 seconds.

In the case of my implementation with the Node, the data is not being sent to the external DAC from the Node asynchronously via USB. It is being sent to the DAC via coax cable synchronously, so the timing information is processed at the Node, and sent to my DAC via coax - USB is not used to connect the DAC in this case.

The external data storage devices I have used were all connected to and powered by the Node via USB. I noticed a difference in sound quality between HDDs and solid state storage devices with the Node functioning as a server. Bluesound has notes on there website about problems with some drives being underpowered by the Node 0.5A or 1A output, depending on the Node version. Mine has a 1A output. While I did not experience some of the more dramatic symptoms of low power with the HDD noted by Bluesound, the audio performance was clearly affected, both in terms of clarity, timing and leading edge attack and trailing edge decay in comparison with solid state drives.

My question is based on the information that the internal switching power supply in the Node has been reported to have challenges powering some mechanical hard drives with higher current draws. In theory or in practice is there a possibility that the switching power supply in the Node is generating more noise that affects the Node’s internal processing of the asynchronous data stream from the storage device for export to the external DAC as a synchronous signal? My suggestion is that this could explain the audibly better performance from the Node and external DAC when music files are served via external SSD or flash drive.

kn

 

[knownothing2]

I take it back. He’s a troll. An ordinary person would be embarassed.

There’s no way to move forward without accepting ABX testing. That’s step one and all the rest build upon it. I don’t feel bad about dismissing people as soon as they say they trust their subjective impression and don’t believe in blind testing.

A question for you? Is it possible to hold two potentially opposing concepts in your head at the same time - those being that the results of well designed ABX/DBT tests show no statistical difference between components A and B in group settings are real and valid, and that individuals can reliably hear differences in their own systems that allow them to tune their system to their room and liking, even when evaluating things like cables and digital front ends? Because I can, and if that makes me a “troll” in your eyes, then so be it.

kn