nwavguy vs NuForce

[limp]

It is higly dependent on what you are using.

Conventional headphones and IEMs with a high and/or even impedance across the frequency spectrum won't be much affected.

IEMs with a low and/or un-even impedance across the frequency spectrum, as his UEs and all multi driver multi cross-over IEMs, will be affected.

[faileas]

He also has some unrealistic expectations - he's assuming all headphones are ideally designed, and SHOULD sound the same. They don't (he should try measuring it ;p). Occationally you can even throw together two pieces of utterly junky kit, and end up sounding slightly better than you'd expect (while the fiio e3 is junk, and i'm not a fan of the sennheiser HD202, for some bizzare reason, they sound half decent together).

[NekoAudio]

And I'm still scratching my head about the zero line nonsense. The zero line is where a digital signal clips, by definition. I don't see how a DAC can make a clipped signal an more clipped. Or if I would be offended if it happened. "Hey, who distorted the distortion in my Lady gaga song!".

FWIW, I've heard and measured this phenomenon, if I'm understanding your question correctly. (Admittedly, this was while measuring and listening to a super-cheapo DVD player.) Obviously music that is highly compressed close to 0dBFS is not a 100% block on a waveform graph. But a poor DAC can introduce additional distortion as a result of that music sitting around 0dBFS. Rather than (or more accurately in addition to) the distortion it may already be introducing when music is not sitting near 0dBFS.

This is why I include Groove Armada's Lovebox in my test track list. I found some of those tracks useful for audibly detecting poor handling of highly compressed music audible.

(Side note: I've noticed some publications have started performing their measurements with test signals that are -3dBFS or even lower. You'll get better measurements, usually, than if you take measurements at 0dBFS. I've been wondering why, since that might be fine for highly dynamic music but don't you want to know how things will perform in the worst-case, which also happens to be a somewhat common-case with a lot of music?)

[CarlSeibert]

Wes -

My impression, from some of my own recordings that have the occasional peak over the zero line, is that each time the signal hits the zero line, there will be a burst of distortion. If I'm understanding what's happening here correctly, peaks that are tangent to the zero line or maybe ever so slightly below it will also clip, with the attendant release of hash. Thus, this device will increase the number of occurrences by adding the tangent or almost-so peaks to the total instances. Is that sound?

And how is it that clipping could occur in the digital domain at anything less than zero?

Or is the phenomenon we're talking about really cascading down to the analog section?

[grawk]

There is NO > 0 in the digital domain. You get a hard, nasty splat. If you don't, it's because your recorder is doing some kind of limiting or compression. But no matter what, there's no data > 0. Because digital is recorded as a reference to 0.

[limp]

And how is it that clipping could occur in the digital domain at anything less than zero?

Or is the phenomenon we're talking about really cascading down to the analog section?

The clipping occurs in the DAC SoC, but at the analog output section of it, so not in the digital domain.

The output gain of the ES9023 chip used in the uDAC can be decided by the implementer by changing a feed-back resistor.

What appears to have happened in the case of the uDAC is that the gain have been set so high that it makes the DAC output section clip at input levels above -1dBFS.

[NekoAudio]

As grawk said, you can't have something higher than 0dBFS in the digital signal and if that portion is perfectly reproduced you'll get that splat.

What I was talking about, I suspect is a function of the DAC chip's reproduction capabilities and the following analog circuit adding distortion because it is being pushed into its worst-case scenario. It's possible the analog circuit could end up pushing a signal above the analog reference 0dBFS in such a case, due to constructive interference, but it could also be that the additional distortion is clipping and adding more nasty splats?

[Tyll Hertsens]

I'm sure even Tyll's great initiative is powered by self-glorification on some level,

?

I'll think about that ... and being a reviewer kinda force a degree of (fucking) "celebrity", but I don't really give a shit about that.

Back OT, I think measurements and listening are both VERY important. The trick is you have to weight both apropriately somehow. Its kinda apples and oranges ... both fruit, but different beasts.

[limp]

?

I'll think about that ... and being a reviewer kinda force a degree of (fucking) "celebrity", but I don't really give a shit about that.

Please do not worry too much about my remark, I was just trying to convey the view that genuinely altruistic actions do not exist. We always act in our own self-interest some way or an-other.

I do admire and greatly apprecieate you work.

[Tyll Hertsens]

Please do not worry too much about my remark, I was just trying to convey the view that genuinely altruistic actions do not exist. We always act in our own self-interest some way or an-other.

I do admire and greatly apprecieate you work.

Thanks. And I do it for money, so theres the self-intrest bit ... but I love it because I love the hobby and great listening.

[Dusty Chalk]

I do it for the nookie.

[NwAvGuy]

It's funny, because the whole "this shit measures bad but sounds good" thing is exactly what happened in the Stereophile CDP-8 review. I kinda don't know what to think. NuForce fixed the jitter issue immediately afterward, but I get a vibe that they wouldn't have done anything if they never got called out. The implications there really bother me, personally.

Let me start by saying I agree with the above and many other posts in this thread. This is my first post here, and after reading this thread, I can see I've already had a rather mixed reception.

As a general rule, anyone who's come out against the snake-oil side of the audiophile world hasn't fared very well or has otherwise been marginalized. The most obvious example is probably Peter Aczel and the Audio Critic. He had more people hating him than liking him. But, to my knowledge, the poor guy never said anything that was actually wrong. He was just trying to expose the truth which can be a rather thankless job in this business.

And while I hope I'm not coming across like Jesus, if someone accused me of evangelizing objective measurements I'd probably have to plead guilty. These forums are filled with 98% subjective posts so, to me at least, there's plenty of room for a bit more of the objective side.

I've been told Head-Case has lots of smart people and I don't see ads and links plastered everywhere for the products being discussed. So that's why I'm here. The admins at HF have become increasingly hostile over my saying bad things about one of their sponsors. I'm hoping to find more open minds here. Is that wishful thinking?

Yeah, my NuForce review is by far the longest. But much of that is due to how NuForce responded before I even wrote it. Had they said just about anything other than (paraphrasing) "yup, the crappy measurements are correct, we designed it that way on purpose" I would have just reported the measurements, made some brief comments, and left it at that. The 8000 words are not so much about a $130 DAC, but about the company that designed it, how they've marketed it, and how they respond to having their product measured. The NuForce CDP-8 sells for $1500.

CarlSiebert here asked the question "I wonder how it would sound if he had never measured it". I'm glad you asked! That's why I conducted a blind listening comparison of 3 DACs to find out. It's also why I listen to anything I'm going to review before I measure it.

CarlSiebert went on to question the 0 dBFS clipping issue. As others here have pointed out, it's a valid concern. And a few even identified the NuForce distortion in the blind listening test. And even if you argue it's inaudible, it's not something I've ever seen in all the other devices I've measured. 0 dBFS is used as the reference signal to which many other tests are calibrated against. So it's something you have to run on just about any piece of digital gear you're testing. It's easy for a DAC to reproduce 0 dBFS without breaking a sweat let alone clipping.

CarlSiebert said "The Beringer unit, according to Larry, sounds like ass. So it loses the war after winning all the bench tests. " Actually it faired pretty well in the blind listening test. And it didn't exactly win all the bench tests, only about half of them. The uDAC-2, for example, was quieter--both subjectively and objectively. Larry's sounded bad as he admits he used IEMs--exactly what the measurements would predict.

Grawk said, "If the objective quantitative and repeatable measurements don't result in a better listening experience, what's the point?". If you look at how the listening test turned out, there is some good correlation between how things measure and how they sound. The high output impedance of the NuForce, when combined with IEMs, created flaws that were easily heard and strongly disliked. Larry's observation of the UCA202 could also be predicted from the measurements.

Pars said, " ignoring things on a technical level shouldn't be tolerated either" and I agree. A "high-end" power amp fried my tweeters because of a really half baked current limiting scheme. The manufacture, much like NuForce, could only say "we designed it to sound great". That amp also didn't come close to meeting their published specs. I don't think anyone should have to put up with such sloppy designs that can cause real problems--especially when they're avoidable at zero cost to the sound quality or the manufacture.

Grahame said, "would it sound better if they fixed those glaring technical problems?" and the answer is generally yes. He went on to say, "Why put up with fixable / avoidable glaring technical problems? If Beringer can avoid them for $29 it seems to imply that cost is not the issue in avoiding the problems? If you have a defined test regime, then you can apply the same tests ( + measurements) to different devices irrespective of the cost of the device, and compare them. In the absence of measurements, how do you get a (process/product) improvement feedback loop to make better stuff?" As Lord Kelvin said "If you can not measure it, you can not improve it." I think the above is all very true. But then I'm an engineer and we tend to think in these terms. Even companies like NuForce should either have a competent engineer or hire the services of one. So some of their mistakes are hard to dismiss.

Atothex said, "I get a vibe that they wouldn't have done anything if they never got called out" And that's another reason for my 8000 words. Does anyone here remember the early days of "Secrets of Home Theater" when it was one guy testing DVD players? He'd write these really long detailed reviews of any DVD player he could get his hands on. He'd find $200 players that did most everything right and $2000 players that got far too much wrong--geeky stuff like how they handled various flags, 3:2 pulldown,etc. Lots of people were blindly buying those $2000 DVD players and "accepting" all the visible problems in their movies until he came along and exposed all the problems. Over the following years, even the huge manufactures had to start paying attention because little companies like Oppo were among the first to pass all his tests. It really can make a difference to call out the manufactures who are getting it wrong. Shouldn't someone keep the low volume guys honest?

Headphone addict said, "I agree that a higher output impedance can sometimes make a difference in the sound but not always" and this is true. With balanced armature IEMs even the NuForce's 6 ohms is enough to cause audible problems. It scored an epic FAIL in the blind listening test when used with IEMs for just this reason. And I show the graph of the response which makes it pretty obvious why.

To Failes, I listen with all sorts of headphones but some of the best in-ear types are balanced armature designs--Shure, Etymotic, UE, and many others. So what's wrong with explaining what happens if someone wants to use them? Yeah I do have ideas on how gear should be designed--mostly basic engineering 101 stuff. I'm fine with gear that's tweaked in a way to deliver a predictable "signature" or "sound" (say a soft high-end or whatever). But the problem with any output impedance over 2 ohms is you get an unpredictable sound that varies widely with different headphones. What benefit is that to the Average Joe reading reviews online? Yeah, if you can even find a retail brick-and-mortar audio store with all the gear you want available for demo, and they let you try enough different combinations, you might find some combo that's "synergistic", but how realistic is that?

To Elnero, please see above. I'd say anything under about 8 ohms is (marginally) acceptable with conventional dynamic headphones. Under 2 ohms is best for balanced armature, multiple driver, or other more esoteric cans. There are more examples on my blog such as this one which shows the Behringer with it's original 50 ohms and after being modified at around 2 ohms (note the 14 dB of total deviation for the original):

All of the above is my attempt to "catch up" here. I'm happy to answer questions, or whatever. Here's the link to the listening test results I mentioned above:

DAC Listening Challenge Results

[Tyll Hertsens]

I was going to say, "To long didn't read" but I did.

I think your spot on about keeping manufacturers honest by taking the measurements. its amazing the stuff Ive found on headphones.

I appreciate your efforts.

Thanks for posting.

[adktitan]

nwavguy,please learn to multi-quote...for many reasons..

[mypasswordis]

90% (okay, number pulled out of my ass but I'm pretty sure I'm close) of audio companies put out specs that are better than they actually measure, and sometimes by a lot. It's kind of chicken-and-egg-y, in that once one starts posting really great specs that are false and get away with it, it makes the others look worse so they inflate their numbers as well and pretty soon there's an unwritten agreement to post false specs, sometimes without even measuring their designs. Well, some people also just kind of make the specs really vague so they can't really be proven wrong, and even Zanden puts up "Specifications subject to change without notice." on their site.

[The Monkey]

I'm all for measuring gear and keeping companies (and their fucking marketers) honest (or at least less dishonest). I just happen to think such things can be done more concisely. I'm just a data point. And Nuforce obviously has no idea how to handle this sort of thing.

EDIT: And welcome to headcase.

[NwAvGuy]

nwavguy,please learn to multi-quote...for many reasons..

Good point! That's a feature I'm not used to forums having so thanks for the suggestion. And thanks also to Tyll for the encouragement.

[digger945]

I'm first in line for the "brown bag" wine tasting.

[NwAvGuy]

90% (okay, number pulled out of my ass but I'm pretty sure I'm close) of audio companies put out specs that are better than they actually measure, and sometimes by a lot. It's kind of chicken-and-egg-y, in that once one starts posting really great specs that are false and get away with it, it makes the others look worse so they inflate their numbers as well and pretty soon there's an unwritten agreement to post false specs, sometimes without even measuring their designs. Well, some people also just kind of make the specs really vague so they can't really be proven wrong, and even Zanden puts up "Specifications subject to change without notice." on their site.

I wouldn't say 90% for electronics, but I would agree for speakers, headphones and a few other things. And sometimes, like you say, it's a matter of clever wording. Amplifier power specs got way out of hand a long time ago so the Federal Trade Commission stepped in and, for a long time, strictly regulated it. But that's no longer the case. But still, most manufactures make an attempt to have some element of truth to their specs--especially ones that are easily verified. Even cheap $200 A/V receivers usually have multiple power specs if you dig out the fine print with some of them actually being reproducible (usually with 2 channels operating rather than all 5 or 7). And some companies, like even Apple, provide very few useful audio specs.

But NuForce was literally off by orders of magnitude with the uDAC-2. They claimed 0.0018% THD+N when the real number is 0.05%--about 27 times worse. And they made a similar huge cut to the dynamic range after my review was published. Both of those measurements are pretty black and white and easy to verify as standards exist for both. In my experience, most gear at least gets close to achieving the claimed specs in these areas. The usual exceptions are the little "boutique" audiophile companies like NuForce but even some of those do it right.

It's also fun when a company makes say 3 versions of a similar product. And as you move up in price the specs get slightly better. But, in reality, all three measure the same (typically as good as the most expensive one). That's just pure marketing. And I don't mind so much if they make the cheaper versions look worse than reality.

[faileas]

To Failes' date=' I listen with all sorts of headphones but some of the best in-ear types are balanced armature designs--Shure, Etymotic, UE, and many others.[/quote']

That introduces a bias in the testing in this case, skewing the results towards gear that works well with in-ear types. While i don't suggest a extremely large range of testing, it may be a good idea to include a 30something ohm headphone as part of the test criteria.

That would assume two pairs of different headphones should sound the same - they don't. with my current, admittedly small setup, i have a pair of cheap IEMs (my 'testing for dangerous DIY design pair') which sound like they're at the bottom of the well, my M50s (which are my daily listening pair, and show the least variation with source) and a sennheiser hd202 - which sounds significantly different based on its source/amp. They all have distinct sound.

No matter what, as such, your headphones *will* affect the sound - so i feel it is a wrong assumption to believe the source is the *main* thing that affects the sound.

Unrelatedly to either - i subscribed to the RSS feed for your site, and didn't get any updates... while i disagree with you on some points, i still find the methods interesting ;p.

[NwAvGuy]

That introduces a bias in the testing in this case, skewing the results towards gear that works well with in-ear types. While i don't suggest a extremely large range of testing, it may be a good idea to include a 30something ohm headphone as part of the test criteria.

I'm not sure what you mean? I listened to the NuForce, before I ever tested it, using HD590's, Denon D2000's, the SuperFi's and Mee M11+'s. All represent significantly different sounds and loads.

That would assume two pairs of different headphones should sound the same - they don't. with my current, admittedly small setup, i have a pair of cheap IEMs (my 'testing for dangerous DIY design pair') which sound like they're at the bottom of the well, my M50s (which are my daily listening pair, and show the least variation with source) and a sennheiser hd202 - which sounds significantly different based on its source/amp. They all have distinct sound.

When someone is deciding which amp or DAC to buy, if they can't try it with their personal headphones first, what they care most about is knowing how it's likely to perform with their headphones (or the headphones they plan to buy with it). If the amp or DAC has a relatively high output impedance, the answer is there's essentially no way to know. But if it has a low output impedance, it's a safe bet the amp will perform very similarly in terms of sound quality with any headphone they use it with. You can just about arc weld with the output of the Benchmark DAC1. It literally doesn't care what headphone you plug into it. It cheerfully delivers the same performance regardless. To me, that's the ideal.

A higher output impedance, however, will change the sound with many headphones. And for some it might even change in ways that person likes. But that's entirely dependent on the particular headphone used. So if you rave about the uDAC-2 with say the Mee M11+ headphones (which are almost a constant 16 ohms). And I like your review so much I run out and buy one but I have UE SuperFi's, Shures, Etymotics, etc (which have widely varying impedances--very different than the M11's). I'm not going to be so thrilled with the uDAC-2. My lowly $29 Sansa Clip portable is likely to sound better. This is supported by all the people who didn't like the uDAC-2 in the listening test with the SuperFi's.

No matter what, as such, your headphones *will* affect the sound - so i feel it is a wrong assumption to believe the source is the *main* thing that affects the sound.

It depends on what you mean by "affect the sound". If you mean different headphones sound different, yeah. But that's due to the headphones. I'm talking about reviewing sources, not headphones. I can assure you the Benchmark DAC1 Pre's "sound", for example, will NOT change in any audible way with different headphones. But yeah, my Denon's certainly sound different than my Sennheisers plugged into my DAC1. That's entirely because of the headphone and has nothing to do with the DAC1.

It's, obviously, impossible to review every possible headphone with every possible source. So you have to do what you can to isolate the two from each other if you want to provide anything resembling meaningful reviews unless someone else is interested in the exact same combo you happen to review.

Unrelatedly to either - i subscribed to the RSS feed for your site, and didn't get any updates... while i disagree with you on some points, i still find the methods interesting ;p.

Hmmm... thanks for subscribing. But if there have been updates, and you're not getting them, I'm at the mercy of Google there. Hopefully if there's a problem they'll fix it.

[faileas]

In this case, i suppose i'd specifically mean the 'listening' tests, or as loads for measurements.

Personally when i look for reviews, i look for people who use similar gear - my own decision to get the udac2 was based off this review and another one i can't remember the location of, which had someone using the udac2 with the same headphones i use.

I'm not talking about covering all the bases. I'm talking about measurements with another typical range of headphone impedances, and suggested one that's fairly common - in short, adding another datapoint to better reflect diversity.

At the moment, you've proven

1) 1/2 the testers can't tell the difference - with no idea if the bias is their source

2) the udac2 sounds horrible with low impedance phones - with no data on a typical medium or high impedance phone

3) it sounds worse than something that costs 10 times as much, and is a lot bigger.

[NwAvGuy]

In this case, i suppose i'd specifically mean the 'listening' tests, or as loads for measurements.

Ah. In theory I agree a higher impedance load would have been good to add to the listening test mix. But as it was, it took a lot of time to put it all together, and many of the listeners didn't even bother with all the trials I offered. So adding more sets of files with different headphones wasn't really practical this time around.

At the moment, you've proven

1) 1/2 the testers can't tell the difference - with no idea if the bias is their source

2) the udac2 sounds horrible with low impedance phones - with no data on a typical medium or high impedance phone

3) it sounds worse than something that costs 10 times as much, and is a lot bigger.

Based on similar assumptions, the tests also show the line outputs of the uDAC-2 sound pretty good. And the $29 Behringer gives the uDAC-2 a good run scoring similar or better in most areas. It also doesn't suffer from the well known channel balance problems, misleading specs, etc.

I've readily admitted they were only informal listening tests not a rigorous full blown trial. My bench measurements, however, I fully stand behind. And there's just no real excuse for some of what I discovered.

That said, I've also said many times on HF there are obviously plenty of happy uDAC-2 owners. And if they're happy, that's all that matters. It obviously works well enough for many people. I also think there are significant differences sample-to-sample (especially for channel balance). So some may have a better one than I tested.

Like I said, my 8000 words wasn't all about the uDAC-2 itself. Much of it was about the rest of the story, NuForce, similar companies, etc.

[elnero]

It seems like there's more to this whole output impedance thing than what's been stated or that it's not as black and white as it's been made out to be here. If an output impedance higher than two is universally thought of as being bad then why would a company publish a higher than optimal spec especially when it looks as if the majority of others companies don't publish it at all?

[justin]

It seems like there's more to this whole output impedance thing than what's been stated or that it's not as black and white as it's been made out to be here. If an output impedance higher than two is universally thought of as being bad then why would a company publish a higher than optimal spec especially when it looks as if the majority of others companies don't publish it at all?

this is one of the least understood specs and is a perfect example of knowing just enough to be dangerous. your Best Buy receiver buying consumer would just skip over that spec, but those on HF seek out amps that have the same output impedance as their headphone's impedance. just one possibility for why this spec would be misstated.

another example "The Apache" specs from the web page

Output impedance: headphone amp, 8 – 2K

Frequency Response: 5-100K

Channel separation: Dual mono. What?

Output signal strength: up to 18 Volts

THD: less than .0002