Just for those who think listening tests are a wrong incentive; I came along a video of the Chief Scientist of ESS (at that time), explaining the concept behind HyperStreaming which is a crucial part of their DAC design. It is about the DAC that is inside the appraised and fantastic measuring Topping DAC.
He explains that people could (and can) hear differences between DAC implementations (in this case multi-bit versus sigma delta). Be assured that these observations did NOT start with controlled and double blinded tests. Those audible differences are below the threshold of what was measurable at that time (e.g. SNR, distortion, jitter, etc).
I worked at the Philips Research Labs (1996-2006), next to the departments working on sigma-delta convertors. The audible observations about those different type convertors was already known to them at that time. The obvious reaction of most of the crew was pointing to their foreheads, claiming those people were biased, and characterised as idiots buying expensive cables and painting the edges of CDs. Did we mention cognitive bias before?
The same observations made ESS curious, and when deep diving they found intrinsic audible noise correlated effects in the sigma-delta modulators. Dedicated research and dedicated measurements were required to reveal those effects technically. When eliminating those effects with a different implementation of the modulator, the differences became unnoticeable.
Again this shows that listening tests are a key start for the (acquisition or) design of audio equipment and their components, to steer design choices and to find/apply very specific measurements to explain audible phenomena. Focussing purely on well-known measurement results only is, well, I dare to call it cognitive bias by itself. The appraised DAC in the Topping is optimised by audible incentives, and somewhere in the movie it is even claimed that measurement results are optimised simply because the market asks for tick-box specs. Whether say the distortion is 0,01% or 0,0001% is obviously way less important than something else not represented in standard measurement results.
This is different from claiming that you should trust any opinion of whatever reviewer and take it as a golden truth. When you state you cannot take Fedde's or my word on some perceptual observations, you are fully right. You don't know us, and have no clue whether you can even trust us. Print it on toilet paper, and do your thing with it if you want.
Putting whatever observations in the bucket "unreliable" because people are subject to cognitive bias, is claiming like the taste of a sandwich with peanut butter cannot be reliable distinguished from the taste of an apple when these are visible to the eater. Such a reasoning deserves a prominent place on that same piece of toilet paper if you on a distance cannot judge the amount of difference that has been evaluated.
He explains that people could (and can) hear differences between DAC implementations (in this case multi-bit versus sigma delta). Be assured that these observations did NOT start with controlled and double blinded tests. Those audible differences are below the threshold of what was measurable at that time (e.g. SNR, distortion, jitter, etc).
I worked at the Philips Research Labs (1996-2006), next to the departments working on sigma-delta convertors. The audible observations about those different type convertors was already known to them at that time. The obvious reaction of most of the crew was pointing to their foreheads, claiming those people were biased, and characterised as idiots buying expensive cables and painting the edges of CDs. Did we mention cognitive bias before?
The same observations made ESS curious, and when deep diving they found intrinsic audible noise correlated effects in the sigma-delta modulators. Dedicated research and dedicated measurements were required to reveal those effects technically. When eliminating those effects with a different implementation of the modulator, the differences became unnoticeable.
Again this shows that listening tests are a key start for the (acquisition or) design of audio equipment and their components, to steer design choices and to find/apply very specific measurements to explain audible phenomena. Focussing purely on well-known measurement results only is, well, I dare to call it cognitive bias by itself. The appraised DAC in the Topping is optimised by audible incentives, and somewhere in the movie it is even claimed that measurement results are optimised simply because the market asks for tick-box specs. Whether say the distortion is 0,01% or 0,0001% is obviously way less important than something else not represented in standard measurement results.
This is different from claiming that you should trust any opinion of whatever reviewer and take it as a golden truth. When you state you cannot take Fedde's or my word on some perceptual observations, you are fully right. You don't know us, and have no clue whether you can even trust us. Print it on toilet paper, and do your thing with it if you want.
Putting whatever observations in the bucket "unreliable" because people are subject to cognitive bias, is claiming like the taste of a sandwich with peanut butter cannot be reliable distinguished from the taste of an apple when these are visible to the eater. Such a reasoning deserves a prominent place on that same piece of toilet paper if you on a distance cannot judge the amount of difference that has been evaluated.
If anything this just shows the shortcomings of listening tests. Despite their extensive listening test efforts ESS dacs have received quite a lot of criticism for their sound 😉
ESS is the one pushing the "numbers game" with their rather useless THD compensations and other optimizations aimed at top notch numbers in AP tests. Some device manufacturers use these gimmicks to get impressive test results. Unfortunately these gimmicks have adverse impact on sound (e.g. high order HD, higher phase noise). IME there are lot of issues with ESS chips: reliability, high chip-to-chip variation, overoptimistic datasheets that leave out important details, foundry errors (mousebite in wrong corner).
Typical sighted listening tests suffer from cognitive bias and the results are only subjective. Others may have different opinions. What is especially misleading is to draw any general conclusions of technical or audible superiority based on these subjective preferences. You may hear a difference or you may prefer the sound but it still only subjectively better.
ESS is the one pushing the "numbers game" with their rather useless THD compensations and other optimizations aimed at top notch numbers in AP tests. Some device manufacturers use these gimmicks to get impressive test results. Unfortunately these gimmicks have adverse impact on sound (e.g. high order HD, higher phase noise). IME there are lot of issues with ESS chips: reliability, high chip-to-chip variation, overoptimistic datasheets that leave out important details, foundry errors (mousebite in wrong corner).
Typical sighted listening tests suffer from cognitive bias and the results are only subjective. Others may have different opinions. What is especially misleading is to draw any general conclusions of technical or audible superiority based on these subjective preferences. You may hear a difference or you may prefer the sound but it still only subjectively better.
Well, I can tell some very gifted and highly educated people have extraordinary capacities: I made DACs with ESS chips that were judged without measurements or listening tests. The fact that they had an ESS DAC chip was enough for a thorough and final judgment.
When someone won't study the literature, won't listen to experts, claims the equivalent of the earth being flat, and rejects any evidence to contrary, that isn't a case of name dropping. It was an effort to see if someone too lazy to study, and holding 'alternative fact' based beliefs, would at least consider listening to the opinion of some well-informed 3rd party.
Also, the comment wasn't directed at you.
Hi, that fact does not seem to be of importance when you give comments.
Overvaluation and mentioning of “names” does not improve anything in a discussion. It is probably a cultural habit as here it means admitting defeat.
Last edited:
The problem is not what certain people believe. The problem is when they keep repeating 'false news' about what every single human on earth can or can't audibly discern that newbies might be fooled into believing. That sort of thing can be corrosive and harmful. If the other guy will stop then I will be happy to stop too.
Can you give us a quote for these "false news" from this thread? Otherwise you are just repeating false news.
You don’t seem to get it. Certain people don’t believe you’re hearing what you say you’re hearing. You will not be able to convince them, who cares! It does no “ real” damage.
It does cause damage because they try to persuade other people to join their hypocritical quasi-religion. It also causes damage when their words are laced with sarcasm that expresses their disgust with people that report hearing differences. Again, the damage is that it tends to result in newbies believing they should feel disgust too. Its contagious. Don't know how long you have been in this forum, but at one point a few years ago there was a lot of anger and disgust aimed at people who were not of the 'measurements only' belief system. To them sighted listening is utterly unreliable and nobody would say otherwise unless they were up to some kind of no good, probably trying to sell some overpriced snake oil or something equivalently evil. Seriously. Those kind of days should not repeat.
A guy who has been here a long time and saw what happened made a graphic showing the range of personality types and beliefs along a couple of dimensions. Bonsai is the name he goes by here. Its actually a pretty accurate pegging of people's personalities. One thing you may notice in the graphic is that some people are of extreme beliefs and tend to be intolerant of other beliefs. It is the intolerance, the disgust associated with it, and the way it is expressed that is bad.
A guy who has been here a long time and saw what happened made a graphic showing the range of personality types and beliefs along a couple of dimensions. Bonsai is the name he goes by here. Its actually a pretty accurate pegging of people's personalities. One thing you may notice in the graphic is that some people are of extreme beliefs and tend to be intolerant of other beliefs. It is the intolerance, the disgust associated with it, and the way it is expressed that is bad.
Last edited:
This ain't politics are war, it does not matter. There is no wrong way to do hifi. Disgust?, so what. I'm an objectivisist, but this is a subjectivist hobby. Trying to convince somebody of something they can't know (what you hear) is fruitless. You're being the classic meme of typing madly at your computer because someone is "wrong" on the internet. I've been here since 2005.
I not trying to convince somebody to believe what I hear. That doesn't matter. Its the intolerance and incivility that needs to change. Sometimes people do reign in the intolerance for a while when its its sufficiently brought to their attention. Its that bad habits tend to recur. When that happens too much, then a corrective response is not inappropriate. Sorry if its not fun to watch. Its not fun to do either, but putting up with bullies abuses is not the right answer either.
It's not intolerance or incivility for me to say I don't believe you without numbers. You can also say I like what I hear I'm not trying to prove anything to anybody. that's it. You seem bothered that people don't believe your ears.
One can easily distuingish those that are lucky to have sisters.
Please keep it nice folks. BTW camp Bohrok here. Just joking, be nice.
Please keep it nice folks. BTW camp Bohrok here. Just joking, be nice.
That only phase was different is true, as viewed in the frequency domain using typical audio spectral analysis.
When phase is different the time-domain waveform must always be different. Always. Its a big nothing burger to say so. The waveforms will also be different in cases where you don't hear a difference due to phase. It therefore means nothing and proves nothing to say the waveforms look different. You may recall that traditional psychoacoustics says the ear is like a phase-insensitive FFT analyzer. Therefore you shouldn't hear any difference since the waveforms are identical to human hearing according to that theory. Therefore you should have done DBT ABX to be sure you weren't imaging something. You know all that already right? Maybe you forgot?
That is not what normally is meant by phase difference. Other than that I leave you to admire that silly marketing stuff.
The message of the video and my posting is not whether the ESS is a good or bad chip in whatever sense, the message is that listening did reveal solid audible properties that were unmeasurable and assumed to be non-existing. Research was done to come up with a hypothesis and correlating measurements (amplitude and state transition dependent noise levels), and when fixing the issue the audible effects were gone. That's all, nothing more.
What are "typical sighted tests"? Which tests do not suffer from cognitive bias?
Totally agree, over-generalisations are dangerous.
In practice recording and reproducing sound is a subjective matter by definition. Nobody knows the intended sound stored on a CD except the producer that mixed and tuned the final result together on specific equipment in some room at some loudness level. Only under the same exact conditions, you will be able to hear what was the intent (and you may not even like it). Deviating from that situation is by definition a subjective choice, including the choice for a "transparent audio chain" that "reliably" reproduces "the content" of a CD.
- Status
- Not open for further replies.