Review and Measurements of Nobsound NS-DAC3 Pro DAC & Amp

[Krunok]

My preference is the pretty girl.

Oh you better keep quiet on that topic - your avatar was the prettiest girl around and now look at what you have!

 

[restorer-john]

...I started to give you a like on this. But I reconsidered...

I started to like your post too, but being as you reconsidered, I thought it only fair to reconsider my like- so I unliked it!

But then I changed my mind and liked it again because I'm just not that petty.... LOL

 

[Blumlein 88]

Oh you better keep quiet on that topic - your avatar was the prettiest girl around and now look at what you have!

Well I'll change it in time. We have to keep some solidarity in our anti_MQA quest.

 

[Blumlein 88]

Was this aimed at me?

No, no another thread about digital vs analog signals. Was really aimed at Wombat. He knows who he is.

 

[Krunok]

Well I'll change it in time. We have to keep some solidarity in our anti_MQA quest.

Aha, so that's what it's all about..

On the MQA topic, this is by far the most amusing title I have seen:

 

[Krunok]

Was really aimed at Wombat. He knows who he is.

I used to think so, but lately I have my doubts..

 

[restorer-john]

Well I think it was a 2nd gen or early 3rd gen Philips unit. Roughly 18 months later. There was a Labor day sale of the Magnavox brand for $199. This had the 14 bit DAC chips (one per channel TDA1540 I think) with 4x oversampling.

The TDA-1540Ds (14 bit) were used for quite a while by Philips, and I'm pretty sure the number they quoted for ENOB was 15.4 or 15.6 at the time.

You must have bought at the time when initial sales around the world had stalled (the deep pocketed audiophiles had paid the big prices and the mums and dads wouldn't cough up) and the heavy discounting to move stock kicked in.

My father bought an Akai CD-D1 about 6-9 months in, and paid AU$699 which was about half price. The Sony had come down to about $1000 IIRC. I've actually got some original brochures with handwritten prices someplace I got in 1983. I'll dig them out for old times' sake.

 

[garbulky]

That is exactly what he means.

BTW, whats with the generic Avatar pic. Come on man, get with the program. Pick a gooffy audio format logo, or a pretty girl. My preference is the pretty girl.

BTW who was that lady on your previous avatar?

 

[sergeauckland]

If we are going to make a statement of inaudibility, it better be true of all people, all content and all conditions. Otherwise it is not inaudible.

I don't agree with this at all. I don't care if somebody somewhere can reliably hear 0.0001% distortion, or frequency response variations of 0.1dB, or noise 80dB below programme levels, as the vast majority of even trained listeners can't. My interest is in what is audible by the vast majority, not the occasional (lucky or unlucky, depending on your viewpoint) freak. That's why I look for adequacy, which is transparency, not SOTA.

I'm far more interested in how reliable, resilient and serviceable something is, than ultimate performance provided that performance is 'good enough'. Others clearly have different criteria.

S

 

[Krunok]

I will allow myself one more car analogy: just try to imagine @amirm as a car tester..

"On my dyno I measured peak power of 230HP, peak torque of 320Nm and max speed of only 240 km/h. I cannot really recommend this car, so I returned it to the owner without test driving it. There are much better engineered engines that are well worth your money."

 

[pwjazz]

take a look at home appliances

This has been a pet peeve of mine for a while. I remember reading an article about a convenience store near San Francisco whose refrigerator from the 1950s was still in operation. Yes, it's not as efficient as modern refrigerators, and no it doesn't have a built-in ice maker, but the thing was built to last.

the batteries have been mostly soldered in

That's part of why my most recent phone purchase was an LG V20 with a removable battery. Sure it's not TOTL anymore, but it's plenty fast and has a great DAC.

 

[pwjazz]

I don't agree with this at all. I don't care if somebody somewhere can reliably hear 0.0001% distortion, or frequency response variations of 0.1dB, or noise 80dB below programme levels, as the vast majority of even trained listeners can't. My interest is in what is audible by the vast majority, not the occasional (lucky or unlucky, depending on your viewpoint) freak. That's why I look for adequacy, which is transparency, not SOTA.

I'm far more interested in how reliable, resilient and serviceable something is, than ultimate performance provided that performance is 'good enough'. Others clearly have different criteria.

S

I'm sympathetic to this view, as well as @amirm's view that, all else being equal, if you can get much better measuring devices for the same or slightly more money, you might as well.

Building a product that performs "well enough" and that wins on features is a respectable proposition. From an engineering perspective, that would require setting some specification targets as "good enough", like 80% SINAD or whatever you deem appropriate, and then making sure the product hits these targets. In this particular case though, that's not how the Nobsound NS-DAC3 is being marketed. It's being marketed on the basis of using a high end DAC chip and hitting an impressive sounding published distortion figure. The fact that the actual device doesn't get anywhere near that figure is a red flag that the manufacturer didn't verify that their design or implementation hits their target, which implies that they don't actually have an idea of what "good enough" performance actually means and doesn't give me as the consumer any confidence in what I'm getting.

Since everyone loves car analogies, I'll use the analogy of the Suzuki Swift, which is probably the crappiest car I ever owned. It was slow, noisy, a little cramped and wandered around at high speeds like a puppy on cocaine. But it was cheap (for a while it was the cheapest new car available in the U.S.), and it was mostly reliable, which is all that Suzuki advertised it as. They never pitched it as a BMW competitor, or for that matter even a Honda Civic competitor. It had one mission - provide transportation from A to B at a low price, and it did the job. From that perspective, I consider it an engineering success, even though it loses out to pretty much any other car on the basis of just measurements.

If the Nobsound where a car, it would be more like the Hyundai XG300. That car had pretensions of being a luxury car on a budget, but having driven one as a rental I can say that it was slow, ugly, and not particularly luxurious.

 

[amirm]

I got into second gen CD players for

Pharmaceutical comparison is crap. Analogies are usually only for the un-educated and usually wrong.

Stick to the relevant(audio} science.

Let's be respectful please.

 

[amirm]

I don't agree with this at all. I don't care if somebody somewhere can reliably hear 0.0001% distortion, or frequency response variations of 0.1dB, or noise 80dB below programme levels, as the vast majority of even trained listeners can't. My interest is in what is audible by the vast majority, not the occasional (lucky or unlucky, depending on your viewpoint) freak. That's why I look for adequacy, which is transparency, not SOTA.

I assume you also want that set to include the music you listen to. If I proved that point using pink noise it would not be sufficient. Or if I used techno music you didn't listened to.

Here is the issue again: the 80 number could be made up of any set of components. Here is the formula for it:

So you are given one number on the left (THD+N) but the components that create that are infinite in variation. You can easily get two devices with identical THD+N that sound different. As an extreme example, if the noise is around 20 to 22 kHz, it would be inaudible whereas if it is at 2 to 3 Khz, it may not be. From: https://www.audiosciencereview.com/forum/index.php?threads/dynamic-range-how-quiet-is-quiet.14/

Notice how our sensitivity at 3 kHz is far, far higher than in low or high frequencies by potentially tens of dBs.

This measurement came about with respect to gross distortions that were rather easy to measure as one number. It was not meant to be and is not an accurate gauge of what is audible. If it were, we would never perform listening tests.

Even full perceptual model of the hearing in every lossy audio codec is insufficient to know what is audible accurately. It is far better than above but we still perform listening tests to approve lossy codec performance.

Back to your hypothesis, let's say there are none of these problems. Where is that study with respect to 80 dB? It doesn't exist, right?

So are you looking to me to demonstrate that for every device I review? I am supposed to recruit larger number of people to perform listening tests across wide range of music, listening levels and headphones to arrive at that conclusion?

If so, then I am not that person. It is not what I signed up for. I have dozens and dozens of gear to review and more is coming every day (no joke). I am taking you from being totally in the dark about a device -- both from instrumentation and audibility -- to tons more data than you had the day before. You can choose to consider -80 dB distortion and noise to be inaudible and go and purchase the gear now. That would be your judgement. Mine would be that based on what I know of engineering there are much better products out there. So I will spend more time characterizing them.

Note that I did this review. I ran a test for the owner a couple of weeks ago. I had a choice of not reviewing this product at all. When the owner gave me this choice, I said I would still do a short review and here we are. It cost me a couple of hours to take pictures, annotate graphs, re-run the tests and write the review. And much more time now writing these posts.

 

[pwjazz]

When the owner gave me this choice, I said I would still do a short review and here we are

For which I'm grateful, btw

 

[amirm]

I will allow myself one more car analogy: just try to imagine @amirm as a car tester..

"On my dyno I measured peak power of 230HP, peak torque of 320Nm and max speed of only 240 km/h. I cannot really recommend this car, so I returned it to the owner without test driving it. There are much better engineered engines that are well worth your money."

That's not the analogy at all. Here is the correct one.

You are not a car tester. But rather, a retired auto engineer/manager who decided in his spare time to set up a dyno in his garage to test cars. People bring in cars from major brands to be tested there. Occasionally, someone imports a car from China that few have even heard of to test. He agrees to also test those. One of those cars advertises to have the same engine, transmission and handling of a Corvette at much cheaper prices. You put it on the dyno and realize that the actual horsepower is half of what Corvette engine produces. You also look around and realize the fit and finish is nowhere near where Corvette is. With a line of brand name cars goring around the block waiting to be tested, you tell the owner the manufacturer claims are false and that there are so many better choices at similar price ranges. Owner agrees and that is that.

You are saying I should have gone above board and test driven the car? Which auto magazine reviewer does what I do and would also entertain that in the same circumstances?

I know your real argument is not any of this anyway. You have no interest in this device. Your interest is to show that all of these devices sound the same. Maybe they do. But I am not in the business of proving or disproving that. If you feel strongly about this, you should put together a project to test such. My job is to put a spotlight on manufacturer claims of measured performance. And where possible, comment on audibility of what I find. That's it. I am not here to perform the project you have in mind. You or someone else should take that on especially since it doesn't take the specialty gear, and knowledge of running it as does measurements.

 

[Thomas savage]

That's not the analogy at all. Here is the correct one.

You are not a car tester. But rather, a retired auto engineer/manager who decided in his spare time to set up a dyno in his garage to test cars. People bring in cars from major brands to be tested there. Occasionally, someone imports a car from China that few have even heard of to test. He agrees to also test those. One of those cars advertises to have the same engine, transmission and handling of a Corvette at much cheaper prices. You put it on the dyno and realize that the actual horsepower is half of what Corvette engine produces. You also look around and realize the fit and finish is nowhere near where Corvette is. With a line of brand name cars goring around the block waiting to be tested, you tell the owner the manufacturer claims are false and that there are so many better choices at similar price ranges. Owner agrees and that is that.

You are saying I should have gone above board and test driven the car? Which auto magazine reviewer does what I do and would also entertain that in the same circumstances?

I know your real argument is not any of this anyway. You have no interest in this device. Your interest is to show that all of these devices sound the same. Maybe they do. But I am not in the business of proving or disproving that. If you feel strongly about this, you should put together a project to test such. My job is to put a spotlight on manufacturer claims of measured performance. And where possible, comment on audibility of what I find. That's it. I am not here to perform the project you have in mind. You or someone else should take that on especially since it doesn't take the specialty gear, and knowledge of running it as does measurements.

Pants on your head, it’s the rules.

 

[Thomas savage]

I don't agree with this at all. I don't care if somebody somewhere can reliably hear 0.0001% distortion, or frequency response variations of 0.1dB, or noise 80dB below programme levels, as the vast majority of even trained listeners can't. My interest is in what is audible by the vast majority, not the occasional (lucky or unlucky, depending on your viewpoint) freak. That's why I look for adequacy, which is transparency, not SOTA.

I'm far more interested in how reliable, resilient and serviceable something is, than ultimate performance provided that performance is 'good enough'. Others clearly have different criteria.

S

I’d agree in so much as your reasoning is impeccably consistent and I like that, can’t fault it.

You can’t make a argument based on a tiny sample of people who might hear something, if you do then where dose that leave the ‘audio science ‘ we praise Harman for when it comes to the spin’o’rama speaker testing. They establish trends and follow those bulk preferences rather than extremes of what some tiny percentage might hear or prefer.

 

[sergeauckland]

For which I'm grateful, btw

And so am I.

To Amir, I say that I fully appreciate the work that goes into these reviews, and want to see the numbers reported. What I dispute is the conclusion that a product is/isn't fit for recommendation based on measurements when these fall short of what is possible. I further don't dispute that should there be two products, identical including price, except for their measurements, then why buy the worse measuring one? The answer for me is that there are seldom two products truly identical, and that one product will offer something different to the other, even if only looks, finish or brand name. One might be better suited than another, and the technical performance of the better suited one may well be adequate, even if short of what's possible.

My buying criteria on all substantial items such as cars, HiFi, cameras etc etc has been for years to create a specification against which I'll judge the possible candidates. I'll buy the cheapest that meets the spec, or if none does, then I'll exercise a judgement (a subjective decision) on what specification points are the more important.

Back in my youth, I worked for Philips, and one thing we were taught, it was almost a mantra, was that quality means fitness for use. There is no benefit to making something that far exceeds customers' requirements if that's going to cost more to make. Define the specification required and design the product to exceed that specification by a small margin to allow for production tolerances, but don't waste time and effort overachieving.
The skill then comes in defining the specification carefully, which is why at Philips, Product Managers were so important.

An amplifier with 0.02% distortion at all power levels, all frequencies and permitted loads is quite sufficient, so I don't look for anything better. Ditto a DAC for home listening. If I was looking for a studio ADC/DAC where the recording might go through several ADC/DAC cycles, then possibly an order of magnitude better would be sensible, but it depends on the application.

I am very much against the subjective view that all that matters is what we (think we) hear, sighted and with no level matching, but equally, I feel that only products that achieve SOTA measurements are worthy of recommendation is also flawed if the performance is otherwise adequate for the job they need to do.

S.

 

[sergeauckland]

My job is to put a spotlight on manufacturer claims of measured performance. And where possible, comment on audibility of what I find. That's it. I am not here to perform the project you have in mind. You or someone else should take that on especially since it doesn't take the specialty gear, and knowledge of running it as does measurements.

That I completely agree with, it needs doing, as there's so much out there with inadequate specifications, indeed some manufacturers won't quote a specification, just tell customers to use their ears!

There there are the 'optimistic' specs and the downright fraudulent. They need exposing.

S.