Article: Weiss Engineering DAC202 Review

[thotdoc]

Chris-<br />

Do you desire to have a rule about no discussions of 'whether cables have a voice' in any discussion except a designated 'Do cables have a voice?'chat.<br />

<br />

I would like that rule. Can we vote on it? These 'do cables sound different' discussions and their protagonists are the most inane discussions and they keep turning up in my mail in box.<br />

<br />

Gordon<br />

<br />

PS: Do cables sound different? Yes. Cables do sound different because they filter differently. <br />

Have experimental results been confusing? Yes. There are listener biases, and experimental flaws, and information overload that confuse some experimental studies? <br />

<br />

What to do?<br />

<br />

Trust your ears. If you can't hear the difference, don't buy the cable. If you can hear the difference and the difference is important to you, buy the cable. <br />

<br />

Why make it more complicated? It's a hobby and supposed to bring pleasure. It's not right or wrong. It's a hobby and a process. It gives pleasure when you don't obsess about it. <br />

[Charles Hansen]

>> My original post was to ask Chris what he thought of the fact that many very well conducted double-blind and ABX tests have shown, that when listeners didn't know what they were listening to, all differences between DACs disappeared.<br />

<br />

I never even said that *I* believed it. <<<br />

<br />

There's a name for people who do that sort of thing. It's called "troll".<br />

<br />

>> The main difference between the two op-amps were that the Nattys had much lower noise (2 µV/root Hz vs. 8 µV/root Hz) and this we could hear in quiet passages <<<br />

<br />

Nice to see that you're on a "friendly nickname" basis with some op-amps. It's not so amazing that you could hear the difference between 2 uV/root Hz and 8 uV/root Hz. That would be a 12 dB increase in an already horrendous noise level.<br />

<br />

But I believe that you meant *nV*, not *uV*. 2 nV/root Hz is a staggering low figure and just about good enough for use in a phono stage for low output MC cartridges. I'm not sure that you could hear a 12 dB increase in noise on the op-amps if you were using them in line level applications. Probably a case of hearing what you expected to hear, eh?<br />

<br />

>> the LME 49710s were designed to have symmetrical slew rates while the OPA134s were not. This manifested itself in greater sense of clarity <<<br />

<br />

Nice to see you once again hearing what you expected to hear once again.<br />

<br />

>> If an interconnect or a speaker cable can alter any parameter in any way which would have any affect on the sound of an audio system, then those cables are not just cables, but are designed purposely to be filters and that can easily be measured. <<<br />

<br />

And for a third time, we find you hearing exactly what you expect to hear.<br />

<br />

>> I do know that there is a such thing as sighted and expectational bias, and that human beings are VERY susceptible to it <<<br />

<br />

Yes, it sound like you are very familiar with "expectational bias", George.<br />

<br />

>> I am very interested in making my system sound more like real, live music. <<<br />

<br />

That's easy. Just expect it to, and it will.<br />

<br />

>> I am NOT Arny Kruger and I do want to discuss the sound of different components. <<<br />

<br />

That's strange. I could have sworn you made a post referencing a "study" that found that all DAC's sound the same. So which is it? Do you want to dispute any listening that isn't backed by the mythically perfect DBT, or do you want to discuss the sound of different components?

[Charles Hansen]

>> But these cable companies like Nordost and Kimber et al don't even have patents on these "breakthroughs" in wire design. Because there are none. <<<br />

<br />

You sure make a lot of confident sounding statements for someone who is so staggeringly ill-informed:<br />

<br />

7,674,973<br />

6,004,166<br />

5,335,282<br />

4,980,517<br />

4,628,151<br />

7,126,055<br />

6,215,062<br />

5,376,758<br />

<br />

There's more, but this is getting boring. Do some homework before you go spouting off your completely uninformed "pronouncements" in public.

[Audio_ELF]

Watch out for Billy Goats...<br />

<br />

True Charles, but then NTP own a patent on "Wireless email"...<br />

<br />

Troll ends :-)

[artk]

He replaced a part designed in 1997 with current technology, it should sound better. The specs on the two parts show they are measurably better. Few if any of the differences between various high end DACs priced from 1 to 10 thousand dollars appear in specifications; many are based on the same parts; the difference are claimed to be beyond measurement; and well designed double blind testing results are unavailable.<br />

<br />

[4est]

Bob Stern on the Pure Music thread stated:<br />

<br />

"Audiophile-quality computer audio is a fledgling hobby, and we're not going to attract new enthusiasts by creating a hostile environment. Until recently, the congeniality of this forum has been a plus compared to the Asylum, but that no longer seems true. Let's all try to return to our former congeniality."<br />

<br />

Guilty as charged, and I for one would hate this site to become like the asylum. I promise to tone it down.

[The Computer Audiophile]

Art - Consider this your one warning. You've been trolling long enough here without adding anything to the site. In fact your comments are frequently false. <br />

<br />

Many differences in DACs are based on specs. The DAC chip itself has little to do with the whole component. Most of us consider items like filters "specs." Filters are easily measurable.<br />

<br />

Again, add something other than headaches to the site. There are plenty other consumer goods sites that I'm sure you can attack with half-truths. I encourage differing opinions, facts, and constructive criticism here on CA, but you've offered none. <br />

<br />

[gmgraves]

Now there's a question. It used to be thought that large amounts of global feedback linearized analog amplifiers and reduced distortion. And it does - for steady-state distortion, anyway. This can be proven with sine waves. Increase feedback (to a point) and watch THD drop. But we don't listen to sine waves, we listen to complex waveforms called music. When tested with complex waveforms (a methodology that didn't exist before the late 1970's) it was found that large amounts of global feedback actually slowed the amplifier down as it tried to correct for non-linearities by feeding back a signal that had already happened upon a new waveform that was totally non-correlated to the feedback signal. IOW, the feedback was too slow to linearize a non-repeating signal. This had been postulated by Harry F. Olson of RCA Laboratories back in the 1940's but since there was no way to measure the effect, it was mostly dismissed by the electronics establishment. It wasn't until the 1970's when engineers from Bruel and Kjaer of Denmark came up with a device that could actually measure these kinds of distortion, that engineers started to pay attention to things like slew-induced distortion, dielectric absorption distortion, etc. Now it's pretty well known that high-amounts of global feedback in solid-state amplifiers is not conducive to good sound. I might also point out that this is not as much of a problem with tube equipment than it is with solid-state because the tube circuits are usually simpler than SS, so the path isn't as long. Tubes tend to have wider gain bandwidth than transistors and thus faster slew rates (this isn't the problem it was in the 70's. Solid-state devices have come a long way) and the capacitors in the tube feedback loop were tiny compared to those used in solid-state designs so the delay isn't as great.<br />

<br />

Today, no SS amp designer uses anything other small local feedback loops around each stage. Global feedback is mostly a thing of the past (except in powered subwoofers using servo loops which include the speaker driver itself in the loop. But here, speed isn't that important as the harmonics that would be affected by the sluggish feedback are generally out of the passband). <br />

<br />

WRT Nordost and Kimber (Ray is an acquaintance of mine, I really admire his work with the IsoMike microphone system and his commitment to helping the music department of Weber State University in Utah, but I avoid talking cables with him) et al, They certainly do make cables for professional and aerospace applications, but that doesn't mean that their audio cables are anything special (other than price) or that they have found any material or process technologies that would make their expensive cables sound any different from any other wire suitable for audio applications.<br />

<br />

Wire is so simple that DBTs easily tell the whole story. One either can hear the difference between cables or one can't. There's not much more to say about the subject and DBTs say that a $4000 pair of 1 -meter Nordost Valhalla interconnects and a pair of radio shack interconnects sound identical (just as electrical theory says they should).

[The Computer Audiophile]

<i>"DBTs say that a $4000 pair of 1 -meter Nordost Valhalla interconnects and a pair of radio shack interconnects sound identical"</i><br />

<br />

Hi George - Please direct me to the testing methodology and the results of the aforementioned double blind test.

[barrows]

"Wire is so simple that DBTs easily tell the whole story. One either can hear the difference between cables or one can't. There's not much more to say about the subject and DBTs say that a $4000 pair of 1 -meter Nordost Valhalla interconnects and a pair of radio shack interconnects sound identical (just as electrical theory says they should)."<br />

<br />

Another case of the blind leading the blind (and apparently deaf).<br />

<br />

I have heard the above comparison (Valhalla vs Generic ICs), and I did not need to be "blinded" to tell the difference. In fact, the difference was enormous, and no one with any ability to hear anything would doubt it.<br />

<br />

I mentioned Nordost's wire being used in aerospace because George claimed that audio cable companies wires were not sought out for use in other industries, just another of George's statements proven false (as with the patents).

[artk]

What were the specific details of that Valhalla Double Blind Test? Was it their analog or digital version? What were the specifications of the interconnect it was tested against? How was the switching performed? How were you shielded from the knowing the interface in use? Any other details, length of test, number of switching between the two interconnects? <br />

<br />

[Charles Hansen]

>> Increase feedback (to a point) and watch THD drop. <<<br />

<br />

Wrong. At first as you apply feedback, the lower order distortion drops and the higher order distortion increases. Eventually when you apply massive amounts of feedback (>40 dB) *all* of the harmonics drop. Then as you increase feedback, the distortion continues to drop (without limit).<br />

<br />

>> engineers from Bruel and Kjaer of Denmark came up with a device that could actually measure these kinds of distortion, that engineers started to pay attention to things like slew-induced distortion, dielectric absorption distortion <<<br />

<br />

Wrong again. Slewing induced distortion was first investigated by Matti Otala and later John Curl. Dielectric absorption was first measured for non-audio applications by Tektronix. Shortly after it was investigated for audio by John Curl and Richard Marsh.<br />

<br />

>> Now it's pretty well known that high-amounts of global feedback in solid-state amplifiers is not conducive to good sound. <<<br />

<br />

Wrong. When I started Ayre 17 years ago, we built only zero-feedback solid-state designs. The only predecessor was the Wingate amp, a short-lived amp available by mail order only. (His patent is online and it's a nice circuit.) Now it has become fashionable to make claims of "low feedback" or "no global feedback" or some other phrase that includes a weasel word. But I would say that even today less than 5% of all commercial solid-state amps are zero-feedback designs.<br />

<br />

>> this is not as much of a problem with tube equipment than it is with solid-state because the tube circuits are usually simpler than SS, so the path isn't as long. <<<br />

<br />

George, you're only half-wrong with this one! Keep it going, slugger! It's true that using feedback with tubes is less of a sonic problem (how do we know? -- listening tests!!!). But it has nothing to do with the signal path being shorter. The fact is that nobody knows why this is. I have some theories, but nothing concrete and there are probably only a handful of people in the world with as much experience as I have in this area.<br />

<br />

>> Tubes tend to have wider gain bandwidth than transistors and thus faster slew rates <<<br />

<br />

Wrong.<br />

<br />

>> Today, no SS amp designer uses anything other small local feedback loops around each stage. <<<br />

<br />

Wrong. This is so wrong it's laughable. Name me three amplifiers (besides Ayre) that are designed this way.<br />

<br />

>> Global feedback is mostly a thing of the past <<<br />

<br />

George, you're killing me! Have you been smoking crack cocaine?<br />

<br />

>> Ray is an acquaintance of mine but I avoid talking cables with him <<<br />

<br />

What? Are you afraid you might learn something?<br />

<br />

Ray is one of the best cable designers on the planet. I'd take advantage of the opportunity instead of wasting my time posting a bunch of falsehoods on an internet forum.

[Charles Hansen]

>> What were the specific details of that Valhalla Double Blind Test? <<<br />

<br />

Art, I would give you a big yawn but it might be wasted. I think your privileges are about to be revoked...

[artk]

So, don't waste the response on me, educate the other readers.

[Charles Hansen]

>> educate the other readers <<<br />

<br />

Gee, Art, look down a few posts and I did just that.

[gmgraves]

Sure Chris. Glad to.<br />

<br />

The test was performed at a friend's home listening room (built into an old photography studio that was originally a 3-car garage for a large old mansion). Chairs were set up for 10 participants. The equipment consisted of a high-end Marantz SACD player (Don't remember the model #, but I think it cost about five-grand. It seemed at least as good as my own Sony XA777ES). I supplied the amplifier, a Krell KAV-300iL integrated. We used that because it had remote controllable input selection and the equipment owned by our host did not. Speakers were Magneplanar MG 3.6s. We used a variety of source discs, both SACD and regular CD; mostly classical and jazz. <br />

<br />

The Marantz was connected to the the two cables via a pair of "Y" adapters. I believe these were from Audioquest, but I might be wrong. One of the cables went to one set of Krell inputs and the other went to another set of Krell inputs. Since we're talking about two 1-meter (or thereabouts, I think the Radio Shack cable was slightly shorter at three feet) lengths of interconnects<br />

the question of level matching was moot. An HP 400GL audio voltmeter showed that playing a test CD's 400 Hz tone yielded identical readings for both cables. I had previously done a frequency response sweep of both cables from 10 to 30 KHz and found both cables absolutely flat in frequency response with respect to each other as noted by the HP 400GL. <br />

<br />

Because the Krell has LEDs to indicate which input is selected, we covered the row of LEDs on the amp with black electrical tape so they could not be seen.<br />

<br />

A clock with a second hand hung on the wall between and behind the two speakers. My friend's bored girlfriend, Beth, was tagged to stand at the absolute back of the room behind all of the listeners and operate the remote which had a label pasted on two of the buttons marked "A" and "B". All of us were given clipboards and pencils to mark 10 tries for 5 musical selections. the form was to be marked A or B by each listener for each try. It was agreed that Beth would switch inputs at 12 O'clock and 6 O'clock of the sweep second hand on the clock. Beth's form was identical to everyone else's except that she was to mark whether she switched or didn't switch and if she did switch at the appointed times, she would mark whether she selected A or B. She didn't know (or care) which button on the remote denoted which cable. Only our host knew which was which, and he didn't participate in the test. He merely changed discs and started and stopped the player. Each selection was played for a full five minutes. Before the test started he wrote down which input was the Valhalla and which was the Radio Shack "Gold" interconnect. and folded the piece of paper and stuck it under the amplifier. Since he couldn't see the LED's on the amp he didn't know which input had been selected at any time. We believed that this set of circumstances satisfies the requirements that the test be double blind. At each thirty second interval. we marked down whether we detected a change between A and B. We all knew that at the sweep of 12 and the 6 we were either listening to the "other" cable or the same one again.<br />

<br />

When the results were tallied, it was found That all 10 participants over 50 tries were correct in guessing which cable they were listening to less than 60% of the time. This falls in the statistical area of "blind chance", which means that merely guessing would yield a similar result.<br />

<br />

Afterwards we discussed the results and all agreed that it was very difficult to impossible to hear any difference at all and that most of the guesses were, just that, guesses. Most had expected there to be some change in the sound, no matter how subtle, at each 30 second interval when the inputs were switched, but they heard nothing change. One listener said that at first, he suspected that Beth wasn't switching at all, but when he saw the sheet she had marked, it was apparent to him that she had (unless she was lying which is unlikely).<br />

<br />

Certainly the Valhalla cable was very pretty. Beautifully made audio jewelry. Needless to say when the owner of the cable retrieved it to take it home, he was a very unhappy camper. "I wasted $4000," was the comment he made. After he left, our host assured us that he was extremely rich and that $4000 was pocket change to this guy. <br />

<br />

I have heard of many such tests being performed (but probably not with Nordost Valhalla cables). The Boston Audio Society (I understand) has done a number of DBTs involving high-priced interconnects and speaker cables as has the AES. All of them show that when sighted and expectational bias is removed from an evaluation, the differences between interconnects and speaker cables disappear. It certainly did in this evaluation<br />

[The Computer Audiophile]

Hi George - Thanks for taking the time to provide all the details. As with anything in life I'm sure some people will take your results as fact and others will find many holes in the methodology and results. Either way you've provided a data point for people to consider. <br />

<br />

Thanks George.

[gmgraves]

OK, so you can produce a bunch of patent number, but what do they denote? I never meant that there weren't any cable patents, Of course their are. That should be obvious. I was referring to breakthroughs in the science of conductors, not in coaxial cable designs. I meant that there aren't any patents on the claims for the improvements in sound emanating from the use of these cables (since audio is what we are talking about). Take, for instance, the first patent that you list, #7,674,973. It only claims to equalize the "velocity of propagation of an electromagnetic wave in the conductive element along the length of the central element". Nowhere in the patent does it mention that this equalization results in any improvement of conduction at AUDIO FREQUENCIES. It also doesn't even pretend to be a breakthrough in cable technology. Nor does it prove that the cable design being patented actually ACHIEVES the claim that it makes.<br />

<br />

Remember a patent doesn't mean that the device patented actually works. The patent files are full of patents given to quacks for "electrical belts" and other infernal machines that claim to cure everything from rupture to "women's problems". In fact, devices in this category are technological snake oil, nothing more. Patents on these types of cable designs are hardly breakthroughs (remember I did specify breakthroughs) unless they ACTUALLY solve some real electrical or electronic problem (the guy who invented waveguides for microwaves, for instance, solved a big problem in the conduction and transmission of those frequencies. now, THAT was a breakthrough). <br />

<br />

I agree that this is getting boring. We shouldn't argue religion 8^)

[gmgraves]

"But of course, all DACs and amplifiers sound the same, don't they?"<br />

<br />

No. And there are no rules in electronics that say that they should. Aside from DAC chips which merely turn PCM into a series of voltage levels which proscribe a waveform, every decision about component quality made after the DAC chip can affect it's sound. Amps will sound different using different components. For instance, an amp built with low-noise resistors is going to be quieter that one built with carbon composite resistors in the signal path, now, isn't it (not that anyone does that anymore, it's just an example of how component choices can affect sound). <br />

<br />

As for cables, Cardas or otherwise, Let's just agree to disagree on that score.

[barrows]

this when reading the description of the (scientific...really?) blind test:<br />

<br />

"The Marantz was connected to the the two cables via a pair of "Y" adapters"<br />

<br />

Are you kidding? I did not need to read anymore. Really, y adapters, and you expected this to be a good way to test a highly developed audio cable????????

[gmgraves]

I just have one comment. <br />

<br />

You said: "Do cables sound different? Yes. Cables do sound different because they filter differently."<br />

<br />

Cables are supposed to be conductors, not filters. If a cable is a filter, then it's been improperly designed either on purpose or accidently. The X-sub-c and X-sub-l of either an interconnect or a speaker cable simply cannot, as far as any math I know is concerned, be significant enough at audio frequencies and at domestic lengths to have any filtering effect on a signal between DC and 20 Khz (a 25 ft piece of coax, might be down a half-dB or so at 20 Khz, but that shouldn't be audible unless your name is Rover. 8^) That doesn't mean that a cable company can't hang outside capacitors and inductors on their cables of sufficient value to peak at some frequency within the passband. But those components would have to be pretty large, physically, and I doubt that they would fit inside the conductor itself or the connectors terminating those conductors. That's why I view cables and interconnects with big wooden blocks or boxes in them somewhere with suspicion. I don't want a non-adjustable tone control. I want a CONDUCTOR.<br />

<br />

<br />

[gmgraves]

I knew that somebody was going to bring this up! These Y adapters are THREE INCHES LONG, fer crissake! Do you actually think that Y adapters would be any different than two parallel pairs of RCA outputs on the back of the unit? Don't look now, but your religion is showing.

[barrows]

"THREE INCHES LONG, fer crissake! Do you actually think that Y adapters would be any different than two parallel pairs of RCA outputs on the back of the unit? Don't look now, but your religion is showing."<br />

<br />

I would not suggest testing cables by using different outputs from a source either (or inputs on a preamp), different outputs (inputs) may sound different!<br />

<br />

Here is how I would evaluate a Valhalla vs Generic cable:<br />

<br />

Wire full system with Valhalla, ICs and Speaker cables. Remove Valhalla cable from source to preamp/integrated, plug generic cable into the same output from the source, and the same input on the preamp. Listen.<br />

Any other method would introduce too many variables to be valid.<br />

<br />

This is not religion, this would be scientific, eliminating other variables. I would not choose to be blinded, as doing so introduces stressors not conducive listening, but if you would prefer to be blinded, feel free to do so.<br />

<br />

I am now less surprised that you claim not to hear differences between certain things.

[SteveS1]

George,<br />

<br />

Is it religion to know that the limiting part of that connection isn't either cable?<br />

<br />

Equalising the cables by shoving an adapter in will certainly give you the results you want. But that's no different to your accusations of people producing results they want. It really was school boy error in testing anything - not to equalise the conditions of use completely .<br />

<br />

As for the general debate, I'm not convinced by DBT methodology mainly because I, like some of the other contributors here, feel it usually introduces stresses and doubts of it's own. But I'm always quite happy, if invited to participate.<br />

<br />

Do I have a dog in the fight, well no. I have found cables to sound 'different', not necessarily 'better' and there are quite possibly measurable reasons - I wouldn't know without a lab. It's an easy thing to compare. I have interconnect cables that I chose from a selection of three sets all around $80-$120. There were slight differences in tonal balance, bass weight and top end. Again, quite possibly measurable if I had the gear - I have used the ones I prefer, ever since. No names - no expectation bias for anyone. :-)<br />

<br />

I have since compared the odd set of more exotically priced cables and again found slight differences in presentation, none that I have felt justified replacing the ones I have. All sighted testing and although often given a nice look and the appearance of being obviously well made, it really cuts no ice with me and frankly it's very arrogant of folk to think that anyone making choices from sighted auditions is somehow so gullible that the mere view of a shiny plug does it for them.<br />

<br />

If I ever hear a set that I prefer they will stay (just like a new box), but I'm in no rush. I also think quick A-B tests are misleading and prefer to live with a change for a few days going through familiar music. Unless of course there is an obvious negative factor as there has occasionally been, like a slightly over-blown bass or bright effect.<br />

<br />

I think the number of people leaping to buy exotic cables because they 'expect' to hear differences has been greatly exaggerated. I didn't blind test my first (ugly and cheaper) DAC against the lovely looking, rather expensive CD player it replaced either.<br />

<br />

But anyway, however you test anything blind or sighted, shoving an adapter into the mix will help equalise results for sure.<br />

<br />

Steve

[gmgraves]

The problems with your alternate procedure are:<br />

<br />

1) There is only one pair of interconnects in use at a time: those from the CD player to the amp. It is the ONLY variable in the test, there ARE no other variables. <br />

<br />

2) Since both sets of interconnects are playing through the same amp and the same speakers, with the same speaker cable. What speaker cable is actually used is irrelevant to the test. Remember, there is only ONE variable, the interconnects.<br />

<br />

3) All the listeners are asked to do is discern whether or not either cable is distinguishable from the other in a statistically meaningful way. Which is "best" is not the job of a DBT<br />

<br />

You said: "I am now less surprised that you claim not to hear differences between certain things."<br />

<br />

Perhaps if you had read the entire procedure, you would understand why there were no statistically significant differences to be heard.<br />

<br />

And one other thing. If a DBT between these wildly disparate sets of interconnects can be fouled-up by merely using a pair of Y adapters (hint: It can't) to allow the two cable sets under test to share the same source component output, then I put it to you that the difference between a $10 pair of interconnects and a $4000 pair of interconnects is so slight that to buy the Valhallas would be a monumental waste of money for anybody, no matter how rich. <br />

<br />

All of these cable manufacturers are selling snake oil. Conductors don't and can't have any sound. If you think they do, it's expectational bias at work. If you understand this and want to buy the cables anyway because having them makes you feel better about your system or makes you enjoy the music more, then it's probably a good use of your money. And make no mistake, it is your money to do with as you please, and far be it from me to burst anyone's balloon. Me? I'd rather put my eight to ten thousand dollars toward a new pair of Martin-Logan Spires or Summit X's than spend it on audio jewelry, but then I don't want a diamond-studded gold Rolex either, and for the same reason. <br />

<br />

I think this is enough between us on this subject, don't you? We're just going to have to agree to disagree. <br />

<br />