I wouldn't have believed it!

[gmgraves]

I just borrowed an AudioQuest Dragonfly v.1.2 DAC. Now I own one of the original DragonFlys (DragonFlies?) and have been using it for almost a year. I had read that the new one was much improved, but you know how that is. One man's "much improved" is another's "OK, it might sound a smidgen better..."

 

But I was not prepared for the magnitude of difference between them. Swapping out the old from the new and using both streaming BSO concerts and High-Resolution downloads via J. River showed an unmistakable improvement in clarity and imaging. The new DragonFly just sounds better. It's as if (cliche alert!) somebody had lifted several veils from the music or replaced the soft focus lens with a really sharp one. More detail, cleaner midrange, smoother sounding highs. By comparison, the older DragonFly, while good, added a slightly rough texture to massed strings. This roughness is totally gone in v.1.2 version. I'm impressed and I suggest that all you original DragonFly owners out there find a way to audition the new one against your current model. BTW, anybody wanna buy a slightly used original DragonFly?

[wgscott]

I just borrowed an AudioQuest Dragonfly v.1.2 DAC. Now I own one of the original DragonFlys (DragonFlies?) and have been using it for almost a year. I had read that the new one was much improved, but you know how that is. One man's "much improved" is another's "OK, it might sound a smidgen better..."

 

But I was not prepared for the magnitude of difference between them. Swapping out the old from the new and using both streaming BSO concerts and High-Resolution downloads via J. River showed an unmistakable improvement in clarity and imaging. The new DragonFly just sounds better. It's as if (cliche alert!) somebody had lifted several veils from the music or replaced the soft focus lens with a really sharp one. More detail, cleaner midrange, smoother sounding highs. By comparison, the older DragonFly, while good, added a slightly rough texture to massed strings. This roughness is totally gone in v.1.2 version. I'm impressed and I suggest that all you original DragonFly owners out there find a way to audition the new one against your current model. BTW, anybody wanna buy a slightly used original DragonFly?

 

Did you A/B it, or is this subjective?

 

My loaned my original one out to someone here, so I need to replace it anyway, but I think I will buy the new one (and hang on to it this time).

[Paul R]

I just borrowed an AudioQuest Dragonfly v.1.2 DAC. Now I own one of the original DragonFlys (DragonFlies?) and have been using it for almost a year. I had read that the new one was much improved, but you know how that is. One man's "much improved" is another's "OK, it might sound a smidgen better..."

 

But I was not prepared for the magnitude of difference between them. Swapping out the old from the new and using both streaming BSO concerts and High-Resolution downloads via J. River showed an unmistakable improvement in clarity and imaging. The new DragonFly just sounds better. It's as if (cliche alert!) somebody had lifted several veils from the music or replaced the soft focus lens with a really sharp one. More detail, cleaner midrange, smoother sounding highs. By comparison, the older DragonFly, while good, added a slightly rough texture to massed strings. This roughness is totally gone in v.1.2 version. I'm impressed and I suggest that all you original DragonFly owners out there find a way to audition the new one against your current model. BTW, anybody wanna buy a slightly used original DragonFly?

 

 

LOL! Is it not a great time to be alive? The DragonFly 1.2 sounds astonishingly similar to my litte Proton- a tad bit better in the treble even. Nobody would believe it, looking only at the price.

[sandyk]

I would be willing to bet that a typical DBT would show there was no difference whatsoever in the sound from both!

Having said that, I notice from a recent post, that George has expressed a few doubts about the validity of some DBT findings though.

 

Alex

[Skeptic]

The original Dragonfly would clip audio when the volume was set to its maximum - which would account for the harshness that you describe.

I'm not familiar with the hardware - it sounds like there is a 64-step volume control and that you should avoid the top four steps: AudioQuest DragonFly USB D/A converter Measurements | Stereophile.com

[wgscott]

The problem is if you use it as a conventional DAC only, you are instructed to turn the volume all the way up, so clipping would be the norm.

 

They produced a defective product.

 

Had the one I loaned out ever been returned to me, I would demand a free replacement. Everyone who has version 1.0 should do the same. It is kind of shocking they expect their victims to pick up the tab.

[Paul R]

Maybe not defective, just misunderstood.

 

I do not think it went into clipping with headphones, perhaps because only a fool would turn the volume all the way up with most headphones. And if memory serves me correctly, they replaced the original model upon request, at no cost to the purchaser.

 

Okay, I admit, I am fond of the little thingie, and that may be clouding my judgement just a little...

 

-Paul

 

 

The problem is if you use it as a conventional DAC only, you are instructed to turn the volume all the way up, so clipping would be the norm.

 

They produced a defective product.

 

Had the one I loaned out ever been returned to me, I would demand a free replacement. Everyone who has version 1.0 should do the same. It is kind of shocking they expect their victims to pick up the tab.

[wgscott]

In much the same way that Toyota's sticky accelerators were really just misunderstandings.

[gmgraves]

Did you A/B it, or is this subjective?

 

Both. The nature of the DragonFly's 60-step computer-controlled volume control is that when swapping the original one for the v.1.2 version and vice-versa, you don't have to worry about matching volume exactly. If you don't change the volume setting on the computer, they will both come-up exactly the same. Makes comparison easy.

 

My loaned my original one out to someone here, so I need to replace it anyway, but I think I will buy the new one (and hang on to it this time).

 

Yeah, I don't blame you. I'm going to have to buy a new one as well! I had to give the v.1.2 that I borrowed back to its owner last night. I was trying to figure out how to slip him mine and keep his, but alas the v.1.2 has markings on it that the original lacks. So there was no way I could have fooled him. Darn!*

[gmgraves]

The original Dragonfly would clip audio when the volume was set to its maximum - which would account for the harshness that you describe.

I'm not familiar with the hardware - it sounds like there is a 64-step volume control and that you should avoid the top four steps: AudioQuest DragonFly USB D/A converter Measurements | Stereophile.com

 

 

That is correct. However, I know that and even though AudioQuest assured me that the one they sent me didn't have that problem (only the early ones exhibited that behavior, apparently), I didn't have either DragonFly set to maximum output. Remember, whatever volume level the computer is set at is reflected identically in both the new and the old DragonFly, so when swapping them out, the output level will be identical on both.

 

The original DragonFly doesn't sound bad by any stretch of the imagination, it's just that the v.1.2 version sounds so much better!

[gmgraves]

In much the same way that Toyota's sticky accelerators were really just misunderstandings.

 

No, you're right, the early samples of the DragonFly did have that defect, but they quickly fixed it, according to Joe Harley and Steve Silberman at AudioQuest and, I understand that they replaced the units affected at no cost to the owners.

[gmgraves]

I would be willing to bet that a typical DBT would show there was no difference whatsoever in the sound from both!

 

I don't know about that Alex. I did a single blind test where both Dragonflys were connected to two different USB ports on a hub, and connected to two different analog inputs of my HK990 using identical audio cables,(not that I think it matters, I just happen to have two mini-phone-plug-to-RCA cables from mycablemart.com) and, when the owner of the new one switched between them, I could pick which was the v.1.2 EVERY TIME and when I did the switching and he did the listening, so could he. The differences are pretty gross between the two DACs.

 

Now, I didn't do that as a blind test so much as I did it to facilitate instantaneous switching between the two units using my remote control from my listening chair. What I was doing before, was getting up, pausing the playback from my laptop, muting the amp (to avoid the hum of disconnecting and reconnecting the audio cable), physically swapping the DragonFly, starting the playback again and un-muting the amp and sitting back down. Now, swapping was instantaneous rather tan taking more than 30 seconds to accomplish, making comparisons likewise instantaneous (and also obviating the need for me to pop up and down from my easy chair like a yo-yo. Hey, I'm lazy and audio is supposed to be fun, not work!)

 

When the v.1.2's owner stopped by last evening (to retrieve his DragonFly - sigh!), It occurred to us that if we took turns on the remote switching between the two DACs, that the other wouldn't know to which he was listening. So, while not double blind, and hardly scientific nor statistically significant, we both could easily tell the new one from the old one.

 

Having said that, I notice from a recent post, that George has expressed a few doubts about the validity of some DBT findings though.

 

True enough, I am becoming more than a bit skeptical about the efficacy of DBTs and audio. When a test procedure always seems to return a null result, on a myriad of product types, with amplifiers, CD players, DACs, showing no differentiation regardless of price or build quality, then I think we had best start questioning the test rather than the units being tested.

 

What started me down this road was the ABX test results found by Meyers and Moran in their study: "Audibility of a CD Standard A/D/A Loop Inserted into High-Resolution Audio Playback"

 

http://colors.webatu.com/wp-content/uploads/2011/03/audibility-of-a-cd-standard-ada-loop-inserted.pdf

 

Their results were clear that statistically speaking, using a large listening sample, they found that a 16-bit/44.1 KHz digital audio signal, decoded and re-encoded a number of times using a daisy chain of processors was indistinguishable from a straight SACD playback.

 

 

This simply cannot be irrespective of Mssrs. Meyer and Moran's findings. While their preparation and execution is impeccable, the results make no sense. I don't question that a carefully mastered CD can sound as good as (or even better than) some SACDs, I've a number of examples of that phenomenon in my own collection. But to take an SACD and decode the audio, and then re-encode it to 16-bit/44.1 KHz, decode that, and then re-encode it to 16-bit/44.1 KHz again, and then decode it to audio yet a second time before routing it to the ABX switcher, and finally to the audio playback system, and finding no difference, is just too silly a proposition to even take seriously, much less conclusively!

 

Then I started to notice the results of other DBTs and ABX tests. Except for those done by several audio magazines, (which I consider tainted for commercial reasons), they invariably returned null results for things that I know for a fact don't sound alike in any way other than that they all played the same music!

 

But I'm not about to throw the baby out with the bath water just yet. Obviously more personal study on this subject by me is required before I completely dismiss DBT audio testing.

[Old Listener]

George,

 

People are quite sensitive to small differences in volume level. If the two Dragonfly versions have even a 0.1-0.2 dB difference, your test result may doing nothing more than documenting the difference.

 

Since the fix to the original Dragonfly was to reduce the volume, this is a possibility to be checked out.

[sandyk]

Hi George

Did the later unit further improve when it was plugged in directly, and not via a hub ?

Some hubs are USB powered, and others supplied with a cheap external SMPS, but I doubt that they sound the same through a good system as when used with a product such as iFi USB etc. as reported by many members.

 

Regards

Alex.

[Don Hills]

... But I'm not about to throw the baby out with the bath water just yet. Obviously more personal study on this subject by me is required before I completely dismiss DBT audio testing.

 

"The confirmation bias is strong with this one." - Darth Fader

:)

 

("Darth Fader" is the username of JJ Johnston on another forum.)

[wgscott]

... they invariably returned null results for things that I know for a fact don't sound alike in any way other than that they all played the same music!

 

How do you "know" this?

 

(The point of having an A/B/X test is to disprove the null hypothesis, not to prove it.) There are probably many many examples where differences exist that aren't detected. There are vanishingly few examples where differences don't exist but can be reproducibly detected in a statistically significant manner with A/B/X testing. (I tried to make that point once on H audio. They didn't dig what I had to say.)

[esldude]

George did you ever play a 500 hz tone and check with a meter at the speaker terminals to confirm equal volume? That is the first most important and simplest step.

[gmgraves]

George,

 

People are quite sensitive to small differences in volume level. If the two Dragonfly versions have even a 0.1-0.2 dB difference, your test result may doing nothing more than documenting the difference.

 

Since the fix to the original Dragonfly was to reduce the volume, this is a possibility to be checked out.

 

 

I use an audio Voltmeter to check levels when I do comparisons. The two DragonFlys track each other perfectly.

[gmgraves]

George did you ever play a 500 hz tone and check with a meter at the speaker terminals to confirm equal volume? That is the first most important and simplest step.

 

I use an audio Voltmeter to check levels when I do comparisons. The two DragonFlys track each other perfectly.

[esldude]

Would be interesting to connect the voltmeter across the two positive leads with a different Dragonfly on each channel.

[gmgraves]

How do you "know" this?

 

 

Same way everybody else "knows" things like that. I listened, picked out certain characteristics of one device, then listened for the same characteristics in another device. Repeated this procedure on different music samples until I'm sure that the two units really sound different

 

(The point of having an A/B/X test is to disprove the null hypothesis, not to prove it.) There are probably many many examples where differences exist that aren't detected. There are vanishingly few examples where differences don't exist but can be reproducibly detected in a statistically significant manner with A/B/X testing. (I tried to make that point once on H audio. They didn't dig what I had to say.)

 

It's the first variable that gets these types of tests in trouble wrt audio, not the last one. Too often, DBTs return a null result when there really are differences. I don't why this should be so. I'm not talking about differences so small that they could be imaginary, either. I'm talking about gross differences of the kind one would easily detect between, say, an MSB Diamond DAC IV ($14000+) and a Cambridge Audio DAC Magic (<$300) merely by listening to the two. I.E. Not as great a difference as that between say, a pair of Wilson Audio Sasha and a Pair of Magnapan MG20.7's, but definitely greater than the differences between a Krell Duo 300 and a Pass Labs XS300 amp.

[gmgraves]

Would be interesting to connect the voltmeter across the two positive leads with a different Dragonfly on each channel.

 

 

You'd need two Audio voltmeters to do that. Otherwise, it's just the same as measuring each separately, isn't it?

[gmgraves]

Hi George

Did the later unit further improve when it was plugged in directly, and not via a hub ?

Some hubs are USB powered, and others supplied with a cheap external SMPS, but I doubt that they sound the same through a good system as when used with a product such as iFi USB etc. as reported by many members.

 

Regards

Alex.

 

 

Not that I noticed, and yes, this hub is powered by a wall wart.

 

I have an iUSB power supply by Ifi as well as the iFi iUSB, USB/SPDIF interface. I've never been able to get them to work reliably together, but the iUSB interface works poorly enough on its own. It spits and clicks when used with streaming audio, and I don't understand why it should. Somebody I mentioned this to on some forum (don't remember which one) said that their Benchmark DAC did the same thing on USB streaming audio. ????? However the iFI iUSB sounds superb playing high-res files our of J. River.

[wgscott]

Same way everybody else "knows" things like that. I listened, picked out certain characteristics of one device, then listened for the same characteristics in another device. Repeated this procedure on different music samples until I'm sure that the two units really sound different

 

 

 

It's the first variable that gets these types of tests in trouble wrt audio, not the last one. Too often, DBTs return a null result when there really are differences. I don't why this should be so. I'm not talking about differences so small that they could be imaginary, either. I'm talking about gross differences of the kind one would easily detect between, say, an MSB Diamond DAC IV ($14000+) and a Cambridge Audio DAC Magic (<$300) merely by listening to the two. I.E. Not as great a difference as that between say, a pair of Wilson Audio Sasha and a Pair of Magnapan MG20.7's, but definitely greater than the differences between a Krell Duo 300 and a Pass Labs XS300 amp.

 

There are many who claim all DACs sound more or less the same. Is it possible at least some of the others "expect" or "know" that the two you listed "must" sound different, and therefore expect to hear differences?

 

Aural memory is demonstrably very fallible. This is why having measurements is so important. If there are measurable differences that in principle might be audible, it at least gives some objective validation. If you hear differences that cannot be measured, it is harder to convince a skeptic that is something apart from expectation bias.

 

But the main point is that the double-blind test is crude and simplistic, and a null result really doesn't tell you much one way or the other. A positive result on the other hand is very hard to argue with. If you start out with the null hypothesis that all DACs sound the same, and in a double-blind test you can pick out one from another in a statistically significant and reproducible way, it clearly refutes the null hypothesis.

 

In science we cannot ever hope to confirm a hypothesis (see Popper and Hume). All we can do is test and refute them. That is why it is so important to be able to state under what circumstances we could accept that we are wrong as clearly and as unambiguously as possible.

[gmgraves]

There are many who claim all DACs sound more or less the same. Is it possible at least some of the others "expect" or "know" that the two you listed "must" sound different, and therefore expect to hear differences?

 

I know several people who have that fixed idea. Most of them have so much of their own ego wrapped up in that notion, that they wouldn't admit that they were wrong if you rubbed their noses in it. I'll even mention a name here: Arnie Kruger - the guy who famously did the Great Audio Debate with John Atkinson (and was creamed by him). Kruger thinks everything sounds the same; CD players, DACs, pre-amps, amps, tuners - basically all electronics. He claims that a cheap Panasonic AV receiver purchased from Costco sounds exactly like an Amp from Krell or Audio Research or Bryston, etc. I suspect that he has talked himself into this belief to justify his "frugal" lifestyle. Someone who says they know him very well told me that Kruger is an unbelievable niggard. Never met the man, can't say. Just parried with him on-line in another forum not connected with CA.

 

Now, my expectation was to hear no differences between the new DragonFly and the original one, and if there were any differences, I would have expected them to be of that ephemeral quality that those who hear audio interconnects talk about. But these are so pronounced that they are like the differences between an inexpensive DAC and a very high end one. Not to infer that the v.1.2 is in the same league as an MSB DAC IV Diamond or a dCS Puccini, but the magnitude of difference between the old one and the new one is that dramatic and un-subtle.

 

Aural memory is demonstrably very fallible. This is why having measurements is so important. If there are measurable differences that in principle might be audible, it at least gives some objective validation. If you hear differences that cannot be measured, it is harder to convince a skeptic that is something apart from expectation bias.

 

Agreed, aural memory is very fleeting. That's why I was motivated to find a way to switch between the two DragonFlys instantaneously (well, that and my naturally inherent laziness ). I felt that the time it took to swap the two DACs out physically was too long to make any accurate assessment as to whether the two units were truly different or fundamentally the same. Of course, my observations were further validated by the single-blind test, where I listened and my friend who owned the v.1.2 switched (I didn't tell him which input on my amp's remote control was assigned to which DAC, so he was merely pushing buttons not knowing what he was switching to...) and I was able to pick-out the v.1.2 ten out of ten tries. When he took the listening chair and I stood behind him and did the switching, he was able to pick the v.1.2 ten out of 10 tries as well. That's pretty much enough for me to say what I said in another post. While the differences aren't as great as speaker differences, they are greater than the differences between two very good amplifiers and thus very identifiable.

 

But the main point is that the double-blind test is crude and simplistic, and a null result really doesn't tell you much one way or the other. A positive result on the other hand is very hard to argue with. If you start out with the null hypothesis that all DACs sound the same, and in a double-blind test you can pick out one from another in a statistically significant and reproducible way, it clearly refutes the null hypothesis.

 

That's my opinion as well.

 

In science we cannot ever hope to confirm a hypothesis (see Popper and Hume). All we can do is test and refute them. That is why it is so important to be able to state under what circumstances we could accept that we are wrong as clearly and as unambiguously as possible.

 

Agreed.