This is a review, measurement, and detailed comparison of two "thumb drive" sized combo DAC and headphone amplifiers: the Audirect Beam and Audioquest Dragonfly Black. I had reviewed the latter last year but it was with my old measurement gear so I thought I perform a fresh set of data on it given the popularity of Dragonfly. I purchased the Audirect Beam from Amazon a few months ago for USD $104 (about $114 with tax). I purchased the AQ Dragonfly for USD $99 from Amazon just the same. The Audirect Beam is on sale for $80 on massdrop and hence the reason I thought I review it now just in case you want to take advantage of its sale.
The Audirect Beam is the smallest DAC+amp I have seen, sans the phone USB-C dongles:
It has a spring-loaded, nice feeling momentary rocker switch for turning the volume up or down. Sadly in practice it would not change level correctly, requiring multiple toggles.
The volume control changed the Windows native device volume control. As a result, it is just a convenience and not an actual analog volume control. By default the level in Windows was very low. It threw me off for a while as did odd performance in the Dashboard measurement. Then I realized it is the issue I saw before with ASIO4ALL interface to Windows causing truncation of 24 bit samples to 16.
This was a "good thing" as after further investigation, I realized this has nothing to do with ASIO4ALL. But rather, these few devices do not support exclusive WASAPI interface! You can try to use WASAPI but it will still go through the Windows audio stack with volume, sample rate conversion, etc.
Anyway, the appeal of this device is what looks too be very nice specifications:
At 0.0004% THD+N, it would have a SINAD of 107 which would put it solidly in tier-2 desktop DACs.
Notice the odd spec "IR: < 1 Ohm." The reads like input impedance would which would not interest us, nor would it be correct way to talk about USB input. If it is output impedance, then that is what we like to see.
Format support is much richer than Audioquest Dragonfly which is limited to 24-bit/96 kHz PCM only.
Let's run through our measurements and see how the two compare to each other.
Measurements
As I noted, my initial dashboard test using ASIO4ALL interface to Audirect Beam generated results that indicated truncation. So I played my test tone in Roon player which nicely avoided the truncation to 16 bits. Even though that took some distortion components away, they were low level ones so overall picture did not change:
We see that the second harmonic at -95 dB dominates so as a result, SINAD can't be better than that which we see in 93 dB results. We have quite a departure here from the specification. Both output voltage and distortion specs are not remotely close to what is advertised, indicating that those must be chip specs, not complete products. Too bad.
Before getting too depressed though, here is AQ Dragonfly black:
Audioquest provides no specifications for the performance of the device. What we get here is worse than the Beam across the board. Lowering the digital input value increased performance up to a SINAD of 83 but then you have a lot less output.
Measuring the output impedances, the tables reverse with Dragonfly showing an excellent 0.6 ohm output impedance which allows driving of any headphone without changing its frequency response. In contrast, I measured the output impedance of Beam at 10.6 ohm. This is heck of a lot more than stated "< 1.0 volt." I remeasured this value twice more and it constantly produced the same results.
High impedance headphones should not be a problem for Beam. But lower impedance ones can be a problem if they have frequency response variations.
For these types of adapters, power availability is everything when it comes to fidelity so let's see how they do at 300 ohm:
For reference, I have included the power from the Topping DX3 Pro small desktop DAC+AMP (red) and portable (but large) Apogee Groove (green). As you see, there is no competition with these two. The Audirect Beam clips at just 3.3 milliwatts and the AQ Dragonfly Black doesn't even get to one milliwattt! Fortunately it has more usable power since there is no hard clipping so you could get to the same power as Beam but with higher distortion.
Changing the load to 33 ohm we see:
Here again there is some overlap in power depending on how you look at the distortion from Dragonfly Black.
To better understand these power ratings, I have started to create a bar graph of portable audio products, sorted by 300 ohm power rating:
Yes, the tiny version 1 Google Pixel dongle has far more power than either the Beam or Dragonfly black!
Listening Tests
A lot of times my listening tests are perfunctory but not in this case. I first started with Audirect Beam and my AKG K92 headphones. Here, there was decent power and definitely listenable. Switching to Sennheiser HD-650 was a huge let down. Yes, you can hear "music" but there is little bass and music has no impact. Switching to Hifiman HE-400i was the same, providing an anemic experience.
Compared to the Beam, the Audioquest Black was a revelation. With AKG K92 performance was superb with thundering bass that would rattle (nicely) the cups! Quite a satisfying experience
. Experience using either HD-650 or HE-400i was respectable but of course, that thunder was no longer there. All in all, the Audioquest Dragonfly is far superior to Audirect Beam subjectively. That soft clipping and low output impedance clearly allows more performance to translate into sonic energy from the headphone.
Conclusions
I have lost track of how many times we have been misled by "paper" specifications. Such is the case with Audirect Beam with numbers that have little relationship to reality. Sadly, it seems that 90% of these products are the ones I buy with my own money Fortunately that means you don't have to. Unless you have a very efficient headphone like the AKG K92, I cannot in any form recommend the Audirect Beam.
The Audioquestion Dragonfly Black shows why it is popular: it provides a satisfying experience with a range of headphones. I can recommend it now based on this new set of data and listening tests. Then again, if you can get your hands on the Google Pixel V1, it may very well be a better choice at a fraction of the cost.
-------------
As always, questions, comments, recommendations, etc. are welcome.
If you like this review, please consider donating funds to support these reviews using:
Patreon: https://www.patreon.com/audiosciencereview)
or upgrading your membership here though Paypal (https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
Without your support, I wouldn't be able to purchase products like this and test them.
The Audirect Beam is the smallest DAC+amp I have seen, sans the phone USB-C dongles:
It has a spring-loaded, nice feeling momentary rocker switch for turning the volume up or down. Sadly in practice it would not change level correctly, requiring multiple toggles.
The volume control changed the Windows native device volume control. As a result, it is just a convenience and not an actual analog volume control. By default the level in Windows was very low. It threw me off for a while as did odd performance in the Dashboard measurement. Then I realized it is the issue I saw before with ASIO4ALL interface to Windows causing truncation of 24 bit samples to 16.
This was a "good thing" as after further investigation, I realized this has nothing to do with ASIO4ALL. But rather, these few devices do not support exclusive WASAPI interface! You can try to use WASAPI but it will still go through the Windows audio stack with volume, sample rate conversion, etc.
Anyway, the appeal of this device is what looks too be very nice specifications:
At 0.0004% THD+N, it would have a SINAD of 107 which would put it solidly in tier-2 desktop DACs.
Notice the odd spec "IR: < 1 Ohm." The reads like input impedance would which would not interest us, nor would it be correct way to talk about USB input. If it is output impedance, then that is what we like to see.
Format support is much richer than Audioquest Dragonfly which is limited to 24-bit/96 kHz PCM only.
Let's run through our measurements and see how the two compare to each other.
Measurements
As I noted, my initial dashboard test using ASIO4ALL interface to Audirect Beam generated results that indicated truncation. So I played my test tone in Roon player which nicely avoided the truncation to 16 bits. Even though that took some distortion components away, they were low level ones so overall picture did not change:
We see that the second harmonic at -95 dB dominates so as a result, SINAD can't be better than that which we see in 93 dB results. We have quite a departure here from the specification. Both output voltage and distortion specs are not remotely close to what is advertised, indicating that those must be chip specs, not complete products. Too bad.
Before getting too depressed though, here is AQ Dragonfly black:
Audioquest provides no specifications for the performance of the device. What we get here is worse than the Beam across the board. Lowering the digital input value increased performance up to a SINAD of 83 but then you have a lot less output.
Measuring the output impedances, the tables reverse with Dragonfly showing an excellent 0.6 ohm output impedance which allows driving of any headphone without changing its frequency response. In contrast, I measured the output impedance of Beam at 10.6 ohm. This is heck of a lot more than stated "< 1.0 volt." I remeasured this value twice more and it constantly produced the same results.
High impedance headphones should not be a problem for Beam. But lower impedance ones can be a problem if they have frequency response variations.
For these types of adapters, power availability is everything when it comes to fidelity so let's see how they do at 300 ohm:
For reference, I have included the power from the Topping DX3 Pro small desktop DAC+AMP (red) and portable (but large) Apogee Groove (green). As you see, there is no competition with these two. The Audirect Beam clips at just 3.3 milliwatts and the AQ Dragonfly Black doesn't even get to one milliwattt! Fortunately it has more usable power since there is no hard clipping so you could get to the same power as Beam but with higher distortion.
Changing the load to 33 ohm we see:
Here again there is some overlap in power depending on how you look at the distortion from Dragonfly Black.
To better understand these power ratings, I have started to create a bar graph of portable audio products, sorted by 300 ohm power rating:
Yes, the tiny version 1 Google Pixel dongle has far more power than either the Beam or Dragonfly black!
Listening Tests
A lot of times my listening tests are perfunctory but not in this case. I first started with Audirect Beam and my AKG K92 headphones. Here, there was decent power and definitely listenable. Switching to Sennheiser HD-650 was a huge let down. Yes, you can hear "music" but there is little bass and music has no impact. Switching to Hifiman HE-400i was the same, providing an anemic experience.
Compared to the Beam, the Audioquest Black was a revelation. With AKG K92 performance was superb with thundering bass that would rattle (nicely) the cups! Quite a satisfying experience
. Experience using either HD-650 or HE-400i was respectable but of course, that thunder was no longer there. All in all, the Audioquest Dragonfly is far superior to Audirect Beam subjectively. That soft clipping and low output impedance clearly allows more performance to translate into sonic energy from the headphone.Conclusions
I have lost track of how many times we have been misled by "paper" specifications. Such is the case with Audirect Beam with numbers that have little relationship to reality. Sadly, it seems that 90% of these products are the ones I buy with my own money
Fortunately that means you don't have to. Unless you have a very efficient headphone like the AKG K92, I cannot in any form recommend the Audirect Beam.The Audioquestion Dragonfly Black shows why it is popular: it provides a satisfying experience with a range of headphones. I can recommend it now based on this new set of data and listening tests. Then again, if you can get your hands on the Google Pixel V1, it may very well be a better choice at a fraction of the cost.
-------------
As always, questions, comments, recommendations, etc. are welcome.
If you like this review, please consider donating funds to support these reviews using:
Patreon: https://www.patreon.com/audiosciencereview)
or upgrading your membership here though Paypal (https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
Without your support, I wouldn't be able to purchase products like this and test them.
