Review: Apple vs Google USB-C Headphone Adapters

[majingotan]

Saw this report confirming that this little $9 dongle is all you need if you only care about getting the best sound for the money (no extra features of course)

 

[Angstrom]

Honestly, it really depends on the headphone you want to drive, there is a current limit on this dongle that does not match the needs of a few planners for example. Low to medium cost IEM should not have and issue

 

[infinitesymphony]

And even then, you can get around the limitations of the dongle by using an external headphone amplifier (battery-powered or not) if it makes sense for your setup.

 

[Angstrom]

Yep, then your back to do you get integrated DAC/AMP combo

 

[minimalist]

Hallo, first post.
Learned a lot about audio here already, and hope for more
Many thanks for that.

Out of curiosity I've been measuring my "Apple USB-C to Headphone Jack Adapter" (Model A2049) lately.

Measurement hardware is an old Thinkpad T400 running Linux and an even older E-MU 0202 USB,
both powered by laptop battery to avoid ground loops.

The Apple Adapter was first connected to a RaspberryPi 2 (for the measurements shown), and later
for cross checking the results to a Thinkpad T430s (measurements not shown, but nearly identical
to the measurements with Pi).
Both devices were running Linux with pulseaudio disabled. Pure ALSA audio, DAC volume at 100%.

First test was 1kHz, 24 bit, 48kHz, reproducing Amir's result. Good.

Next test was multitone, because I missed that one in the review.
I picked up the multitone test file Amir provided here: Multitone test file

Downsampled to 48k with sox:
sox 'APx555 Multitone 32 192 khz 24 bit.wav' -b 24 -c 2 'Multitone_48000_24_0dB.wav' rate -v -L 48000 dither

Measurement result looked very disappointing: not even 70dB of distortion free range!

Next I 'measured' the used test file via ALSA loopback device with REW to ensure the test file does not contain any unwanted artifacts from downsampling. Test file was OK.

Maybe clipping in the DAC?

Ok, downsampled multitone file again, this time with -3dB gain, providing some headroom to the DAC:
sox 'APx555 Multitone 32 192 khz 24 bit.wav' -b 24 -c 2 'Multitone_48000_24_-3dB.wav' rate -v -L 48000 gain -3 dither

Much better, but still a bit disappointing: slightly more than 90dB of distortion free range. Even my Behringer UCA202 can do that (in 16bit!).

This was the best measurement for multitone test I could come up with.

I'm not absolutely sure that my measurements are correct,
so I would appreciate it very much if someone else would share his findings.

Edit 03/28/2020: Seems my dongle is a defective unit. Please ignore my previous recommendations.

Until then, assuming my measurements are correct, I see two things to take away:

1) Ensure to feed the DAC with about 3dB headroom to avoid clipping.
2) Should we still call the DAC transparent, as I've seen here occasionally in some threads?
I don't think so.

Don't get me wrong. The DAC is cheap and small and good enough for many use cases. Just don't expect wonders.

 

[bobbooo]

Hallo, first post.
Learned a lot about audio here already, and hope for more
Many thanks for that.

Out of curiosity I've been measuring my "Apple USB-C to Headphone Jack Adapter" (Model A2049) lately.

Measurement hardware is an old Thinkpad T400 running Linux and an even older E-MU 0202 USB,
both powered by laptop battery to avoid ground loops.

The Apple Adapter was first connected to a RaspberryPi 2 (for the measurements shown), and later
for cross checking the results to a Thinkpad T430s (measurements not shown, but nearly identical
to the measurements with Pi).
Both devices were running Linux with pulseaudio disabled. Pure ALSA audio, DAC volume at 100%.

First test was 1kHz, 24 bit, 48kHz, reproducing Amir's result. Good.

View attachment 50402

Next test was multitone, because I missed that one in the review.
I picked up the multitone test file Amir provided here: Multitone test file

Downsampled to 48k with sox:
sox 'APx555 Multitone 32 192 khz 24 bit.wav' -b 24 -c 2 'Multitone_48000_24_0dB.wav' rate -v -L 48000 dither

Measurement result looked very disappointing: not even 70dB of distortion free range!

View attachment 50403

Next I 'measured' the used test file via ALSA loopback device with REW to ensure the test file does not contain any unwanted artifacts from downsampling. Test file was OK.

Maybe clipping in the DAC?

Ok, downsampled multitone file again, this time with -3dB gain, providing some headroom to the DAC:
sox 'APx555 Multitone 32 192 khz 24 bit.wav' -b 24 -c 2 'Multitone_48000_24_-3dB.wav' rate -v -L 48000 gain -3 dither

Much better, but still a bit disappointing: slightly more than 90dB of distortion free range. Even my Behringer UCA202 can do that (in 16bit!).

View attachment 50404

This was the best measurement for multitone test I could come up with.

I'm not absolutely sure that my measurements are correct,
so I would appreciate it very much if someone else would share his findings.

Until then, assuming my measurements are correct, I see two things to take away:

1) Ensure to feed the DAC with about 3dB headroom to avoid clipping.
2) Should we still call the DAC transparent, as I've seen here occasionally in some threads?
I don't think so.

Don't get me wrong. The DAC is cheap and small and good enough for many use cases. Just don't expect wonders.

Which model do you have? A2049 (US version), or A2155 (EU version)? Be wary, the model number on the box it came in may be wrong - I ordered what was described as model A2049 on eBay (with that written on the box), but it was actually model A2155 inside. The correct model number will be written on the cable of the dongle. Your measurements maybe look more in keeping with the A2155, which was measured by Reference Audio Analyzer here.

 

[minimalist]

Which model do you have? A2049 (US version), or A2155 (EU version)? Be wary, the model number on the box it came in may be wrong - I ordered what was described as model A2049 on eBay (with that written on the box), but it was actually model A2155 inside. The correct model number will be written on the cable of the dongle. Your measurements maybe look more in keeping with the A2155, which was measured by Reference Audio Analyzer here.

The dongle was purchased in Melbourne, but not in an Apple Store. Both box and cable have printed "A2049" on them. Manufacturing date is 09/2019. I have no reason to doubt that it is an original Apple dongle. Now that you mention it, yes, the multitone measurement of model A2155 looks similar, thanks for hinting. Unfortunately I could not find any multitone measurements of model A2049 for comparison.

 

[bobbooo]

The dongle was purchased in Melbourne, but not in an Apple Store. Both box and cable have printed "A2049" on them. Manufacturing date is 09/2019. I have no reason to doubt that it is an original Apple dongle. Now that you mention it, yes, the measurement of model A2155 looks similar, thanks for hinting. Unfortunately I could not find any multitone measurements of model A2049 for comparison.

If it is the A2049 then I think the poor results might be (at least partially) due to your measuring equipment - even the baseline noise on your 1 kHz test is around 10 dB higher than Amir's result. @amirm could you do a quick multitone test on your Apple dongle (model A2049) at some point?

 

[minimalist]

If it is the A2049 then I think the poor results might be (at least partially) due to your measuring equipment - even the baseline noise on your 1 kHz test is around 10 dB higher than Amir's result. @amirm could you do a quick multitone test on your Apple dongle (model A2049) at some point?

Of course I don't expect my measurements to be comparable with AP measurements. Noise floor of the E-MU 0202 is around -140dB. It's ADC is comparable to the E-MU 0404 USB. I've already measured other DACs below -110db in multitone test with the E-MU 0202, so it's not likely that the measuring gear is at it's limits at -90dB in this case. Time to very politely ask @amirm if he has unpublished multitone measurements of the Apple dongle?

 

[bobbooo]

Of course I don't expect my measurements to be comparable with AP measurements. Noise floor of the E-MU 0202 is around -140dB. It's ADC is comparable to the E-MU 0404 USB. I've already measured other DACs below -110db in multitone test with the E-MU 0202, so it's not likely that the measuring gear is at it's limits at -90dB in this case. Time to very politely ask @amirm if he has unpublished multitone measurements of the Apple dongle?

It might be interesting for you to repeat your measurements at 16-bit output, as at least for the Lightning version, there's an increase in distortion at 24-bit compared with at 16-bit. You might also want to see how changing the sample frequency to 44.1 kHz affects the measurements (it looks like this is what Amir used).

 

[minimalist]

It might be interesting for you to repeat your measurements at 16-bit output, as at least for the Lightning version, there's an increase in distortion at 24-bit compared with at 16-bit. You might also want to see how changing the sample frequency to 44.1 kHz affects the measurements (it looks like this is what Amir used).

Did that already before my first post on the topic. 48k/24 was best. Other measurements were nearly the same or just a few dB (2dB-3dB IIRC) worse. Thanks for your hints anyway. Double checked everything. The only thing left that was not replaced in the measuring setup (except for the DAC) is the USB-A to USB-C adapter, of which I own only exactly one. But I would be very surprised if that was the culprit.

 

[snake3276120]

As a Google Pixel 3 user, I bought this dongle after reading this review and ditched my Pixel dongle...

 

[Mrcrunch08]

As a Google Pixel 3 user, I bought this dongle after reading this review and ditched my Pixel dongle...

I'm surprised you switched. The apple dongle may measure good but to my ears it sounds the worst of all the ones I tried minus the one made by Onn. Plus being a pain to get the full power out of it using an Android device ruins it for me.

 

[snake3276120]

I'm surprised you switched. The apple dongle may measure good but to my ears it sounds the worst of all the ones I tried minus the one made by Onn. Plus being a pain to get the full power out of it using an Android device ruins it for me.

I don't use this for serious listening... mostly it's for work with a relatively crappy headphone which is very easy to drive (19 Ohm). I have dedicate Dac/Amp at home with OK headphones.

When I test it I didn't do volume matching, which I should test and compare again with volume matching. Good point. Another driving factor for me to buy it is the low cost lol. But I doubt there will be auditable differences between Apple/Google dongles.

 

[ninetylol]

I'm surprised you switched. The apple dongle may measure good but to my ears it sounds the worst of all the ones I tried minus the one made by Onn. Plus being a pain to get the full power out of it using an Android device ruins it for me.

You don't have to justify your subjective bias on an objective based science forum.

 

[Mrcrunch08]

You don't have to justify your subjective bias on an objective based science forum.

I'm surprised to hear that sort of response from you considering you clearly experienced the problems I mentioned. Outside of me not liking it everything else I stated seems pretty objective. What other dongle requires an additional $8 purchase to potentially get close to the measured performance? Even then it is a noticeable difference from using it on a PC and on a phone regardless of uapp. From what I can see none of us have any actual proof those numbers hold up when it is used as intended. From what I have read here no one knows if other dongles measure better when used with a phone. You already stated you experienced the lack of power so is it not possible other things go out of wack when used as intended? There's subjective, objective, and just plain ignorance. Bias is nothing more than your assumption.

 

[SIY]

This was the best measurement for multitone test I could come up with.

I'm not absolutely sure that my measurements are correct,
so I would appreciate it very much if someone else would share his findings.

I'd run a loopback at minimum to see what your actual test limits are. I haven't run a USB-C dongle, but I've run tests on the Lightning dongle and it's quite good.

 

[ninetylol]

I'm surprised to hear that sort of response from you considering you clearly experienced the problems I mentioned. Outside of me not liking it everything else I stated seems pretty objective. What other dongle requires an additional $8 purchase to potentially get close to the measured performance? Even then it is a noticeable difference from using it on a PC and on a phone regardless of uapp. From what I can see none of us have any actual proof those numbers hold up when it is used as intended. From what I have read here no one knows if other dongles measure better when used with a phone. You already stated you experienced the lack of power so is it not possible other things go out of wack when used as intended? There's subjective, objective, and just plain ignorance. Bias is nothing more than your assumption.

Actually I sold my 600 dollar desktop setup and replaced it with the apple dongle on my Windows PC. Sounds wonderful to me and no volume problems there either. On android you may need a app for hardware volume - so what? Even the DFR for 200 dollars got this "problem". If you need the volume headroom you sure can spend another 4 bucks. But that's not the real problem. Its you saying it sounds bad to you without any reliable proof of whatsoever. On the opposite the numbers say you are wrong.

 

[Veri]

Actually I sold my 600 dollar desktop setup and replaced it with the apple dongle on my Windows PC.

Respect

 

[minimalist]

On the opposite the numbers say you are wrong.

Which numbers do you mean? Do you have numbers of a reliable measurement of IMD (inter modulation distortion)?