|Hmmm... What does "LESS DIGITAL NOISE" actually mean? After all, noise that's bad enough to affect digital data will result in errors. Obviously, that would be very bad!|
Normally, I don't review stuff like this because there's simply little reason to believe these kind of things do anything of value.
I know, this comes across as highly biased right at the start of a review post! But I want to be honest and I wasn't born yesterday so I know there are all kinds of claims in the audiophile world that simply do not pan out. Furthermore, I've had experience with the AudioQuest company (see the Dragonfly Cobalt / Red / V1.2 DAC review a few years back) and know that they have made claims which turn out to be untrue when tested.
Another admission going into this is that I was introduced to the AQ JitterBug back in 2015, and over the years I have seen other tests like the Audio Science Review thread from 2019. Having said this, I am happy to be swayed if I truly hear a difference or measurements show me that I should change my mind.
So, when a local audiophile friend bought the new AudioQuest Jitterbug FMJ ["Full Metal Jacket" - cute] (~US$60, released mid-May 2021) and suggested I borrow it for a week as he went on vacation, I figured it would be fair to give it a try and see if I can run some measurements to provide data to audiophiles at large on this update to the product. I promised him that I would give him my honest opinion.
To start let's look at the box and device:
It's a small rectangular USB "stick" that measures 1.6cm wide x 1.2cm tall x 5.5cm long including USB Type A male connector (4cm long without the connector). It feels reasonably sturdy in the hand with metal casing, rounded edges, and the rubber USB cover out back is a nice design change from the old JitterBug which is just a plastic thing with exposed rear connector. The rubber cover comes off if I pull it too hard; I could use my fingernail to shove the hinge part back into the slit with a little bit of effort.
Something I did not like was that the unit rattles a little bit which reminded me of the construction of the Dragonfly Cobalt I previously described. It weighs <20g with thin metal casing. I have not seen any pictures of this device open so have no idea if they significantly changed the circuitry beyond the metal casing. Word on the street is that there has been no change internally.
A look behind the rubber cover - standard USB Type A female connector, good to see gold plated pins inside the connector:
Here it is hanging off the side of my laptop. It is relatively long and could be ridiculously cumbersome if you have a USB DAC sticking even further out:
My old AudioEngine D3 DAC from back in 2014 itself is ~5.5cm (not including male connector) so in total that's sticking out around 9.5cm (~3.75"). Have to be careful if I move the laptop around not to hit anything! Probably best to use a USB extension cable to reduce strain on the port if I want to use this combination.
Something I noticed about the construction was that the angle of the USB connector had a slight downward tilt from the metal body looking from the side. This won't affect function and who knows, maybe it was meant to be like this but the construction looks a little "off". Don't know if other units are the same.
Ostensibly, the purpose of this device is for "data & power noise filtering" as per the product description. Also, within the name is the implication that jitter can be improved. Note that this is a USB 2.0 product so don't stick a USB 3 hard drive into it which presumably will just degrade data transfer speed (did not try).
For more of what the company claims this device does, here's the rear of the box:
|Cool, that rubber cap at the back is "carbon-loaded" and apparently dissipates RF noise. ;-)|
Knowing this then, from a testing perspective, we should see if noise floor is lower, check for improvements in distortion, and whether this thing affects jitter with DACs. They also talk about funky stuff like "internal noise-intermodulation" and giddily suggests you can buy another one for parallel noise reduction purposes - "a 2nd JitterBug in parallel is always a delightful improvement as it pulls more RF Noise off the USB power bus"... Uh huh... Sure...
First, let's discuss the "materials" for this little experiment. Here are the DACs I'll use with the device:
As you can see, I have my RME ADI-2 Pro FS warming up which is representative of a more expensive "high fidelity" DAC (with ADC) >US$1000 capable of very low noise thanks to balanced outputs. Note that I will use the ADC function for measurements of course. Less expensive, but still around $200 to the right is the Topping DX3 Pro (V2) which is powered by a wallwart and I know from previous measurements that there is some playback jitter. Finally, let's try the old silver AudioEngine D3 (from 2014) sitting on top of the RME. This is a USB-powered USB Audio Class 1 "stick" DAC which is capable of 24/96 hi-res and measurements years ago showed reasonably well controlled noise level (but higher than the other two devices) and low jitter.
In combination with the DACs, let's also measure using a few USB outputs as organized below. Let's get started!
A. COMPUTER USB PORT
Let's put the JitterBug right off the front USB port of my Intel NUC 6i5SYH computer with WiFi turned on, using the NUC's stock switching power supply, close to the HDMI monitor. Notice that there's a wireless keyboard/mouse dongle on the adjacent USB port that could add extra "noise":
|Technical detail: because at 100% volume the D3 clips, I'm measuring at 90% output level.|
|You can see the average noise level of the DACs with RME ADI-2 Pto FS < Topping DX3 Pro < AudioEngine D3.|
C. RASPBERRY PI 3 B+ "TOUCH" STREAMER USB PORT
Finally, let's run measurements using the Raspberry Pi 3B+ "Touch" device playing from the USB storage you see plugged into one of the ports in the picture below. This is commonly the type of set-up I use for DAC measurements. With a separate playback system disconnected from the same computer/USB hub, we will not see ground loop issues like above.
I'll just use a standard 3A switching power supply for the Pi with touchscreen and WiFi turned on:
As you can see in the picture above, I've already plugged in the AudioEngine D3 DAC to the JitterBug. Test signals were copied on the USB stick, playback software is the recent Volumio v2.882.
For this set-up, I'll test the Topping DX3 Pro V2 and AudioEngine D3 DACs since there's no such thing as running a loopback measurement with RightMark on the ADI-2 Pro off the Pi.
Alright, there we go with "proper" measurements of the Topping DX3 Pro and AudioEngine D3 free from general computer and specifically the 8kHz USB PHY noise. As you can see, there's really not much difference numerically at all between a direct USB connection and using the JitterBug.
Here are the noise FFT and IMD+N graphs:
Wow. Look at how clean/jitterless the old AudioEngine D3 is off the Raspberry Pi 3 B+ streamer! Again, some low-level "skirting" around the 12kHz signal. No evidence that the JitterBug has the ability to clean any of this up.
I noticed that some of AudioQuest's announcements for this product suggests that using the device will result in "more spacious and engaging sound". Okay... Let's have a listen. I used 2 systems:
A. Headphone: Intel NUC --> +/- JitterBug FMJ --> Topping DX3 Pro --> Sennheiser HD800
B. Sound room: Raspberry Pi 4 "Touch" streaming from Roon server --> +/- JitterBug FMJ --> RME ADI-2 Pro FS R Black Edition --> Emotiva XSP-1 analogue preamp --> Hypex nCore NC252MP amp --> Paradigm Signature Reference S8 speakers + subwoofers
|The JitterBug sticking out behind my Pi 4 "Touch" - I have a 90° adaptor for the USB port. (A 1993 Hacken Lee 李克勤 Cantopop album cover on the screen - called Album in fact.)|
I spent a couple evenings listening, switching in and out the JitterBug three or four times with each set-up. Instantaneous A/B switching isn't possible so I did the switch as quickly as I could.
Seriously folks, after taking the JitterBug in and out of the USB path a couple of times, it's clear there's no audible difference worth the effort. The novelty wears off very quickly and definitely I am not hearing any "more spacious" nor "engaging" sound compared to what I already enjoy directly through the USB port without the JitterBug FMJ. No differences in background noise level or any impression of "blacker blacks" or other such clichés.
Sorry, no creative subjective fictions of what I heard to describe for you guys. ;-)
Here are the albums I had a listen to those evenings: Olivia Rodrigo's Sour (modern pop), Brian Bromberg's Jaco (jazz fusion), Kaori Muraji's Cinema: Movie Themes for Classical Guitar, and Melody Gardot's Sunset in the Blue.
At least I did not experience any issues with playback or dropouts using the JitterBug. It basically worked and sounded "transparent".
III. Discussions & Summary...
Well, across multiple DACs and various USB ports - the Intel NUC, Raspberry Pi 3B+, 10-port USB 3 hub - as you can see, there is nothing here to suggest that the AudioQuest JitterBug FMJ made any worthwhile difference to audio playback.
Despite me not taking any "audiophile precautions" at all - WiFi turned on with devices, Bluetooth on if present, stock switch mode power supplies, USB hub with other stuff plugged in - there is no evidence that noise has in any way been significantly improved by the JitterBug FMJ. Do not buy this product thinking that it is capable of suppressing ground loops and stuff like the 8kHz USB noise if you run into such issues - there's no galvanic isolation here. Likewise, despite the JitterBug moniker, there is no evidence that this device improves the timing precision of the DACs plugged into it to affect jitter performance in a positive way.
In fact, depending on which DAC was used and with which USB port, sometimes the JitterBug FMJ might even result in slightly more noise or worse jitter. Nothing major, but just the same, why spend money on something that doesn't actually make a difference and just as likely could result in slightly poorer audio output?
Subjectively, I hear no difference when doing A/B listening and I do not feel compelled at all to own one of these. Admittedly, I do not claim to have any super powers but I did receive a "Golden Ears" certificate a number of years back, I have avoided loud concerts through life, and am not quite 50 years old yet. ;-) (Obviously, I believe personal characteristics like age and listening acuity are important for subjective evaluation.)
As you can see in the picture of the box above, this product (presumably based on reviews of the original plastic JitterBug) has been endorsed by The Absolute Sound, Stereophile, the understandably defunct AudioStream, Hi-Fi News, and Positive Feedback. Not listed there, we also can find Darko Audio, What Hi-Fi?, Hans Beekhuyzen ("if you don't hear any difference, you know you have to improve how you set up your stereo", LOL), Paul McGowan of PS Audio and TONEAudio with their accounts of improvements. Likewise, recently this Part-Time Audiophile article with thoughts by Graig and Grover Neville was rather hard to believe including the parts on the JitterBug. Then there's IMO YouTube BS like this.
Stereophile's measurements of the original JitterBug (2015) showed no difference yet claiming that it "improved the sound - which it most certainly did" with no further sense of curiosity to investigate why the reviewer believes such a thing is simply ridiculous. I believe John Atkinson was wrong - he most certainly was.
I would strongly suggest that audiophiles take endorsements and awards from audiophile "news" outlets with caution whether it be very expensive stuff or cheaper trinkets like this.
IMO, it's a shame that writings / videos like these have permeated the mainstream audiophile "culture"/ "lore" so thoroughly. Claims of audibility hidden behind subjective "evidence" are not hard to obtain (or to influence, or to buy). After all, one could always say to the unbeliever that "your ears and gear are just not resolving enough", right? Thus absolving the subjective commentator of need to provide any evidence of what they supposedly heard, or that he could just plainly be wrong.
The fact that there has been a chorus of positive reviews for something as questionable as the JitterBug I believe is an example of how this system of subjective-mainly reviewing is incapable of separating truth from advertising, and in the process treats the interests of the consumer as secondary to the claims of manufacturers. Until this changes, much of the audiophile Industry and its mouthpieces simply cannot be taken seriously by most music lovers. Thankfully, my sense over the years is that audiophiles have become more rational as a group and hopefully can maintain a healthy level of skepticism going forward.
I was thinking about audiophile "snake oil" and how I would classify the JitterBug FMJ. It's probably at best a Class B "Unnecessary Accessory" that provides some high frequency filtering in the circuit path similar to a USB jack with integrated EMC filtering. There's a range of chokes and ferrite beads one can find for USB data and power line filtering (especially for noisy industrial applications) but it would be hard to justify these as helping achieve better sound quality on the other end of a consumer audio DAC! Presumably AQ incorporated a few of these inexpensive parts in the Dragonfly Cobalt which is why they claim the JitterBug is not "needed" with that.
From a marketing perspective, I can see AudioQuest wanting to capture the "cheapo audiophile tweak", "trinket" segment with this product suggesting to audiophiles that it's inexpensive enough to take a chance on, or as some kind of "stocking stuffer" at Christmastime. Smart marketing move I guess. AudioQuest also suggests that you might want to buy 2 of these and plug up an unused USB port. Again, smart marketing move but IMO misleading and definitely "snake oilish". By the way, for giggles, I did a couple of measurements putting the JitterBug on an adjacent USB port and as expected, found no difference at all. I am certainly not curious enough to get 2 of these to try in parallel given the lack of difference even when used in a serial path!
Bottom line: I think it's best to spend your US$60 elsewhere. For example, go treat family/friends to a good bottle of wine - you'll get more enjoyment and they'll love you for it. Giving the money to a good charity will also likely improve the world more than whatever change to the sound of one's system.
I would love to see AudioQuest come up with evidence that this product makes a difference when used in conjunction with a DAC - what equipment they choose, and with what tests would actually be highly informative!
Note that even though the JitterBug failed to show value, there are relatively inexpensive USB isolators out there that could be useful in some circumstances. For example, ADuM4160-based devices are inexpensive and easily available but they are limited to 12Mbps "full speed" - fine for up to 24/96 stereo. Maybe at some point I'll give the iFi iDefender+ a try to see if ground loop elimination can deal with stuff like the 8kHz USB noise while maintaining high speed. There are other high-speed USB isolators (eg. Intona) I might look into later. Historically, I've also tried the Corning USB 3 optical cable, and USB-over-ethernet cable extender which might be useful in some situations.
Big thanks to my friend (who wanted to remain nameless) for letting me borrow this for the week. I suspect he will be returning this after I talked to him about my findings and showed a few of the graphs above.
With all that's done and said, if you still want to give AudioQuest your money, please feel free to use my Amazon Affiliate link here: AudioQuest JitterBug FMJ. While AudioQuest makes a few bucks, I'll also make a few bucks which I will subsequently recycle into buying some CDs to support the artists. Thanks for the charity. ;-)
Cheers (with a glass of that $60 bottle of wine of course)!
As you can see in the measurements above, there are improvements in noise level to be had when using balanced (XLR) output if we look at the RME compared to Topping DX3 Pro (RCA) and AudioEngine D3 (3.5mm headphone) single-ended outputs.
This reminds me of a recent paper by Dr. Milind Kunchur - "Cable Pathways Between Audio Components Can Affect Perceived Sound Quality" (here's a PDF copy from the author) - that shows statistically in listening tests that a short XLR cable output vs. longer RCA cable out made a difference using the same Cambridge DAC. (We've come across Kunchur before with his research on auditory temporal acuity.)
In 2021, apparently we still need research from the University of South Carolina into balanced vs. unbalanced cables to show that the common mode rejection provided by XLR output is beneficial and can be audible!
Notice that Kunchur doesn't even show a standard log audio frequency FFT between the RCA and XLR outputs to help us appreciate the difference in noise floor and the relative amounts of stuff like 60Hz hum. Instead, a linear graph out to 2.5MHz seemed important to the author for analogue audio and human listening?! I don't understand what kind of review process the Audio Engineering Society has to allow an article like this to pass! Hasn't the science behind balanced audio transmission been known since before the 1950's? Maybe it's a slow time for journals given that so much research effort has gone into COVID-19 lately.
Anyhow, here's a summary I posted on an Audiophile Style thread recently:
Posted June 13, 2021:
Yes, the author sets out to find a difference in audibility between 2 cable paths. One with a 2m RCA cable and the other with a 0.5m XLR cable. Both plugged in simultaneously as far as I can tell to the same Berkeley Audio Alpha 2 DAC. Both cables connected to the same Spectral DMA-250S Amplifier.
Potential causes of differences between the RCA and XLR outputs:
1. The Berkeley DAC may perform differently between the single-ended and balanced outputs - distortion levels, frequency response can be different. Even as output level controlled.
2. I didn't see a measurement of the output impedance of the Berkeley DAC and the input impedance of the Spectral amp; I assume they're OK for both balanced and unbalanced operation.
3. Odd that RCA cable longer than XLR. Unbalanced cables obviously more susceptible to noise which is what they found. No surprise. (And no surprise that this could be audible.)
4. Even more strange, why did he choose an expensive $500 XLR short cable vs. $50 RCA long cable?
If we give the benefit to the Straight Wire Virtuoso cable as better built with better connectors since it is 10x the price for 1/4 the length, then he's totally stacking the listening test against the RCA output! And guess what, the RCA output path sounds worse with more noise which he concluded as "However, the electrical measurements conducted here indicate that noise levels may be one determining factor of sonic performance".
"May be"!? Why doesn't he just show us an FFT of the noise floor from the RCA output vs. the XLR from 20Hz-20kHz? Let's see the level difference between the 60Hz hum for a start...
What exactly is the new "scientific" finding here or the uniqueness of this experiment?
Hope the Canadians had a great July 1st, and happy Independence Day to the Americans ahead.
Have an awesome summer enjoying the music, dear audiophiles...