Gustard X16 Balanced MQA DAC Review

[Pdxwayne]

Do you get different readings if you do multiple tests? I know you said your meter isnt the best like a true RMS meter?

Although not accurate, very consistent (like within 0.001V)

 

[ABall]

Although not accurate, very consistent.

Your dealing with very small diferences. I wonder how much can be down to consistency and how much down to voltage fluctuation? Anyone here have some SCIENCE they can throw in?

 

[Frio]

SCIENCE

Is usually a method to take out confirmation bias, but those whole little measurements were made and interpreted sold to support a subjective feeling.

 

[Pdxwayne]

Your dealing with very small diferences. I wonder how much can be down to consistency and how much down to voltage fluctuation? Anyone here have some SCIENCE they can throw in?

Correct, very small, but pretty consistent.

Here is what I have for E30 vs KTB from a while back:

 

[A Surfer]

I hope you are not going to deny measured voltage differences, even with certain filters, are not audible, and keep repeating that everything is imagined, subjective placebo....

Those differences could have been measurement instrument error if I remember correctly, not to mention you need to provide evidence that those voltage differences are audible, and again if I recall (yes I did read the entire thread, albeit a little quickly) they were very small differences and not likely to be audible. I may be mistaken and I freely admit that I did not manage to memorize all of the thread.

 

[A Surfer]

Here's the thing, im coming from 30 years of being one of those people that have been replacing caps, power supplies resistors etc and I have enjoyed doing the work and listening to music on my equipment. However I have stuck around here because I want to stop spending money on this stuff and its working! Just to point out I am on both sides..... But who makes these rules up? Ive read all about ABX testing, how you have to have so many positive results and 7 out of 10 is just the same as guessing. There is no SCIENCE in these parameters that I can find, only opinions, how do you prove these things, maybe I should of posted this on that other thread but its been brought up here.... I used to avoid equipment with opamp outputs like the plague, always thought the sound of discrete circuits sounded better.... I am much happier trusting numbers now, I may have sold myself out but Its going to save me a lot of grief and money and guess what, the music still sounds good. Oh and for us subjectivists, I cant count how many times Ive had life changing sound stages only to find out a couple of days later its not there, mood could play a massive part in this, I used to measure in my mind where every note and every instrument was in my lounge, I was obsessed, changing days could have the same effect as changing a component or device..... Just food for thought.

Not true, it is about doing enough trials to reduce the likelihood of chance being responsible. If you do only say six trials it is possible to guess 5 times correctly, unlikely perhaps but quite possible. The more trials that you do the less likely you are to consistently be lucky. These are very valid scientific sampling/probability principles. You would also have less confidence in a significant result based on one trial participant hence the need to replicate the results with as large a sample of participants as possible.

 

[ABall]

Not true, it is about doing enough trials to reduce the likelihood of chance being responsible. If you do only say six trials it is possible to guess 5 times correctly, unlikely perhaps but quite possible. The more trials that you do the less likely you are to consistently be lucky. These are very valid scientific sampling/probability principles. You would also have less confidence in a significant result based on one trial participant hence the need to replicate the results with as large a sample of participants as possible.

what exactly isn't true? I said I have read "they say 7 out of ten right guesses isnt enough" man those people should stop pissing around with hifi and go to a casino! But Where is the SCIENCE for the number that is ok? And what is it based on? Theory? Just asking..

 

[Atanasi]

But Where is the SCIENCE for the number that is ok? And what is it based on? Theory? Just asking..

It is based on binomial distribution. When people are guessing, they are equally likely to guess correctly or not each round, so p=0.5. The parameter n is the number of rounds. You can use Wolfram Alpha to compute probabilities.
For example, "binomial distribution, p=0.5, n=10, at least 7" gives the result that at least 7 correct answers out of 10 is given by chance with a probability of 17 %.
Usually the result is considered significant if the probability of guessing is less than 5 %.

 

[A Surfer]

what exactly isn't true? I said I have read "they say 7 out of ten right guesses isnt enough" man those people should stop pissing around with hifi and go to a casino! But Where is the SCIENCE for the number that is ok? And what is it based on? Theory? Just asking..

Look into probability theory.

 

[Pdxwayne]

Those differences could have been measurement instrument error if I remember correctly, not to mention you need to provide evidence that those voltage differences are audible, and again if I recall (yes I did read the entire thread, albeit a little quickly) they were very small differences and not likely to be audible. I may be mistaken and I freely admit that I did not manage to memorize all of the thread.

If I could voltages matched from 120hz to 5khz, it is extremely difficult to differentiate DAC, even if one DAC have higher outputs in 6Khz and up.

But, when test E30 with a certain filter, it was like 6 out of 6 times my kids can differentiate which is which in blind tests, when playing songs with high notes. I was going to do more, but people are saying I am wasting my time doing such test for a certain filter with large drop in high frequency. So I stopped.

So yes, indeed when used as it, with certain filter, it is possible to hear a different. My kids don't care which DAC is which.

 

[ABall]

Look into probability theory.

I dont think I would look into anything that has the word theory In it, by definition that is the opposite of fact isnt it?

 

[ABall]

It is based on binomial distribution. When people are guessing, they are equally likely to guess correctly or not each round, so p=0.5. The parameter n is the number of rounds. You can use Wolfram Alpha to compute probabilities.
For example, "binomial distribution, p=0.5, n=10, at least 7" gives the result that at least 7 correct answers out of 10 is given by chance with a probability of 17 %.
Usually the result is considered significant if the probability of guessing is less than 5 %.

Well that's sounds very sciencey, so I googled it....
Seems Theory is popping up a lot here and as I said that's not the same as fact, dont even get me started on statistics, lies damn lies an all that....

In probability theory and statistics, the binomial distribution with parameters n and p is the discrete probability distribution of the number of successes in a sequence of n independent experiments, each asking a yes–no question, and each with its own Boolean-valued outcome: success or failure.

 

[A Surfer]

Well that's sounds very sciencey, so I googled it....
Seems Theory is popping up a lot here and as I said that's not the same as fact, dont even get me started on statistics, lies damn lies an all that....

In probability theory and statistics, the binomial distribution with parameters n and p is the discrete probability distribution of the number of successes in a sequence of n independent experiments, each asking a yes–no question, and each with its own Boolean-valued outcome: success or failure.

So you think all of science is in question then? You do realize that virtually everything that has to do with sensory perception, experience, emotion, behaviour ultimately needs proxy measures to get at and with that there will always be error in measurement and chance. Statistics are not willy-nilly little things, statistics are extremely robust mature mathematical procedures that have been refined and honed by hundreds if not thousands of brilliant mathematicians. They are not perfect, but they are the best thing that we have and I trust them far more than I trust some dude on the Internet who spends a boatload of money on fancy, rarified earth metal cables wrapped in fairy wings and then hears a difference.

I am not sure what level of evidence you will believe in or accept, but thank you, I am good with imperfect science. Far better than the untestable "trust me I can hear it" claims. Beyond that, properly done, blind level matched, multiple trial listening tests when subjected to validated statistical analysis are more than capable of answering these questions whether you or I chose to believe so or not. At one time people believed the world was flat and that the gods determined your fate. Thankfully, in my opinion, science had at those beliefs and sent them scurrying away into the annals of history.

As well, the more we sample the more we diminish the effects of chance in our measurement.

 

[A Surfer]

If I could voltages matched from 120hz to 5khz, it is extremely difficult to differentiate DAC, even if one DAC have higher outputs in 6Khz and up.

But, when test E30 with a certain filter, it was like 6 out of 6 times my kids can differentiate which is which in blind tests, when playing songs with high notes. I was going to do more, but people are saying I am wasting my time doing such test for a certain filter with large drop in high frequency. So I stopped.

So yes, indeed when used as it, with certain filter, it is possible to hear a different. My kids don't care which DAC is which.

I think we have all agreed that it is possible to compare a well done and properly implemented filter to one that is neither and find that the poorly designed filter may indeed introduce audible artifacts. That however, is not the same thing as saying that two DACs sound different. When we talk about a DAC being audibly transparent it is perfectly valid, in fact required, to make some assumptions. In this particular instance the assumption we are making is that a well implemented, audibly transparent DAC will not be using a flawed filter. I do not think it is a stretch to say that using a flawed filter is an error and not at all an example of an acceptable DAC.

If you are trying to extend the reach of your results to generalize beyond that one filter implementation with the evidence you have at hand I would have to say clearly you cannot. All you can say is it appears that the filter produced audible discernable differences in that implementation. I am not sure what else you can use the results for. I could be wrong and if I have misunderstood the situation my apology, but as it stands now, that is my understanding.

 

[Pdxwayne]

I think we have all agreed that it is possible to compare a well done and properly implemented filter to one that is neither and find that the poorly designed filter may indeed introduce audible artifacts. That however, is not the same thing as saying that two DACs sound different. When we talk about a DAC being audibly transparent it is perfectly valid, in fact required, to make some assumptions. In this particular instance the assumption we are making is that a well implemented, audibly transparent DAC will not be using a flawed filter. I do not think it is a stretch to say that using a flawed filter is an error and not at all an example of an acceptable DAC.

If you are trying to extend the reach of your results to generalize beyond that one filter implementation with the evidence you have at hand I would have to say clearly you cannot. All you can say is it appears that the filter produced audible discernable differences in that implementation. I am not sure what else you can use the results for. I could be wrong and if I have misunderstood the situation my apology, but as it stands now, that is my understanding.

What I am saying is that without voltage/volume matching, if a user simply switch DAC to compare them, an audible difference is possible. And no, not just a broken filter. I am claiming using any filters for E30 is possible to decern a difference, if no voltage match to KTB.

Instead of keep throwing out rhetorics, how about you help me out and test the x16 vs E30? Prove me wrong with your own measurements and listening tests.

 

[A Surfer]

You would never compare without matching those parameters, not that I am aware of anyway. Why on earth would you not match them? The moment you allow confounds to flourish the validity of your analysis is gone. Match and compare, otherwise, as far as I have been taught and understand don't bother.

 

[cbayschm]

So you think all of science is in question then? You do realize that virtually everything that has to do with sensory perception, experience, emotion, behaviour ultimately needs proxy measures to get at and with that there will always be error in measurement and chance. Statistics are not willy-nilly little things, statistics are extremely robust mature mathematical procedures that have been refined and honed by hundreds if not thousands of brilliant mathematicians. They are not perfect, but they are the best thing that we have and I trust them far more than I trust some dude on the Internet who spends a boatload of money on fancy, rarified earth metal cables wrapped in fairy wings and then hears a difference.

I am not sure what level of evidence you will believe in or accept, but thank you, I am good with imperfect science. Far better than the untestable "trust me I can hear it" claims. Beyond that, properly done, blind level matched, multiple trial listening tests when subjected to validated statistical analysis are more than capable of answering these questions whether you or I chose to believe so or not. At one time people believed the world was flat and that the gods determined your fate. Thankfully, in my opinion, science had at those beliefs and sent them scurrying away into the annals of history.

As well, the more we sample the more we diminish the effects of chance in our measurement.

There will always be flat earthers as there always will be people willing to depart with thousands of dollars for fairy wing laced cables.

Don't waste your time trying to change that, let them be.

This has gone like 2-3 pages of:
I'm right...
No, I'm right...
No, I'm right....
Repeat....

Neither of you will convince the other, get over it and move on.

 

[MiradoOne]

Hi, I'm new here and new to this Gustard brand as well. For first and foremost, please forgive me if I sound a bit too forwarding.

I have a similar setup (Node2i & Mac as streamers and this Gustard as DAC). I have tried several ways to play Tidal MQA: signal sent from Node to Gustard via coax/optical or Tidal app on Mac to dac via USB. I have never been able to have MQA displayed on Gustard. Honestly I'm not even sure if Gustard will display MQA. As above post, it seemed to do just that for @axeman. Am I missing something?

So, this is my quick and dirty Dr. Frankenstein-esq setup for a MQA test (I don't intend to utilize MQA very often if at all so it's only temporary). It's a Raspberry Pi 4 running Volumio, with a SSD in a USB3 enclosure connected to it that has a MQA test file (Britten: Simple Symphony, Op. 4 by TrondheimSolistene from this page), hooked into a 7 inch touch screen (the Pi is screwed into the back of it), and connected to the X16 via USB. As you can see, the X16 correctly displays when it's doing the full MQA process.

I frankly don't think you are missing much as the MQA sounds the same as every other high res format on that page to my ears, but this proves it will indeed display MQA when the stars are aligned correctly.

 

[Pdxwayne]

You would never compare without matching those parameters, not that I am aware of anyway. Why on earth would you not match them? The moment you allow confounds to flourish the validity of your analysis is gone. Match and compare, otherwise, as far as I have been taught and understand don't bother.

My whole responses were triggered when you and other claim all differences heard are imaginations. I disagree with the “all“ part. I am not arguing about the correct way to do blind tests. I am disagreeing with your blanket statement.

Pretty simple: DAC do not all output same voltages when fed same signals. Typical users didn't realize they are supposed to voltage/volume match. Thus they heard differences. That is all. That is not imagination and not bias, just a lack of knowledge doing proper blind tests.

Now, let's get back to x16. How about helping me do some measurements and compare x16 vs E30 when being fed -16db signals? Isn't this what ASR is about?

 

[JohnYang1997]

I know. But how often we listen to 2V? There is a reason why I output my source at -16 db when trying to voltage match.

I was trying to measure at kind of normal listening volume and matched voltages at 120hz.

A normal user who compares two DAC would use voltages way less than 2V....

Come on. Been watching this for a few days now.
It's probably just that the high frequency modulation noise tricked the multimeter. And Multimeter is almost never accurate in measuring AC higher than 100Hz(or has large components that are in high frequency) .
Get an oscilloscope at least or better a proper sound card to remotely measure something like this accurately.
Or you could ask anybody else who has the tool to measure for you. Are you really believing what you are measuring?