Measurements of RME ADI-2 DAC and Headphone Amp

[amirm]

I reviewed and measured the RME ADI-2 DAC a year ago. That was just prior to getting my new Audio Precision APx555 so the measurements don't directly compare to what I publish now. So a local member kindly loaned me his ADI-2 DAC to get fresh measurements.

If you are not familiar with RME ADI-2 DAC, it is a DAC-only version of the RME ADI-2 Pro. The former costs US $1,099 and the latter, US $1,999. It is an all-in-one DAC, headphone amplifier plus remote control. It has better connectivity with all the jacks in the back rather than dongle in RME ADI-2 Pro. The menu system is world's easier to navigate than the RME ADI-2 Pro. It has built-in parametric EQ and pretty spectrum display.

For the measurements, I took an abbreviated set but I think it paints the picture quite well.

DAC Measurements
Dashboard view is with USB input using RME ASIO driver and XLR balanced output set to nearly 4 volt output:

SINAD (signal over noise and distortion) is 112 which is excellent. Even more excellent is that the highest harmonic is at less than 122 dB! So what brings it down to 112 dB is noise which is much more benign than distortion. The SINAD firmly plants the RME ADI-2 DAC in excellent category of well over 100 DACs tested:

Here is the RCA/unbalanced output performance:

Jitter is picture perfect:

As is linearity:

And here is the IMD graph:

At max level, it beats the Topping DX3 Pro. It has higher noise though that may be an artifact of where I have set the volume.

Per later request in the thread, here are the multitone and THD+N vs frequency:

Headphone Amplifier Measurements
Here is the power versus distortion at 300 ohm:

These are both using the 1/4 inch jack. In low mode the RME ADI-2 DAC matches the Topping DX3 Pro in noise but stops just a bit short in power. It makes up for that in high power mode where it produces twice as much power and at lower distortion level.

Switching to 33 ohm and measuring both the 1/4 inch and IEM jacks we get:

The IEM jack does what it is supposed to do: producing even lower noise than the low mode (by about 5 dB). This is reflected in signal to noise ratio measurements:

Which gets us this for sensitive IEMs as far as noise:

An excellent showing all around.

Conclusions
What we had guessed from my old measurements is confirmed using new instrumentation. The DAC portion is near perfect as is the headphone amplifier. Yes, a dedicated headphone amplifier like JDS Labs Atom or Massdrop THX AAA789 is a hair better but they lack the integration that the RME ADI-2 brings.

In some sense, I think of the RME ADI-2 DAC as a better Topping DX3 Pro. It has more power and features. And of course it is German made for those of you who prefer to get western gear for whatever preference (import costs, warranty, etc.). You will pay four to five times more for these differences so you would have to decide the trade offs.

Of course I continue to recommend the RME ADI-2 DAC.

-----
We are going to have a huge party for those of you who have donated to the forum. It will be open bar and all you can eat. Since I am not too fond of spending my own money on such things, I hope the rest of you can donate so that we can eat and drink well. Use either:

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).

 

[mi-fu]

Thanks Amir for the re-measurement. I bought my ADI-2 DAC last year after frequenting this site.

Once again it shows ASR saves me lots of money

 

[Typevii]

Great to see the RME ADI-2 DAC now formally in the ASR SINAD master list.

 

[Music1969]

Hi @amirm , when you show the jitter measurements in review can you indicate which DAC input is being measured ?

And maybe compare TOSlinks vs USB the way JA does at Stereophile?

And also show 16bit vs 24bit the way JA does?

Ideally TOSlink and USB would measure the same but as you know, sometimes they don’t.

Also, for headphone amp power and distortion measurements can you do 16 ohm measurements?

 

[amirm]

Hi @amirm , when you show the jitter measurements in review can you indicate which DAC input is being measured ?

You should always assume USB. I usually say that in the graphs but this is a quick set of measurements so I just post the raw graph.

 

[amirm]

Also, for headphone amp power and distortion measurements can you do 16 ohm measurements?

I don't have that resistance on my dummy load. Will have to modify it to add it which is a bit more work than my lazy self wants to get into right now.

 

[Blumlein 88]

Hi @amirm , when you show the jitter measurements in review can you indicate which DAC input is being measured ?

And maybe compare TOSlinks vs USB the way JA does at Stereophile?

And also show 16bit vs 24bit the way JA does?

Ideally TOSlink and USB would measure the same but as you know, sometimes they don’t.

Also, for headphone amp power and distortion measurements can you do 16 ohm measurements?

Of course Amir will decide, but really what is the point of the 16 bit vs 24 bit the way JA shows it? In 16 bit undithered you'll get a basic squarewave pattern with ringing tops if the device is low noise. If the device has a higher noise level it will look more messed up. So you can glean the same information looking at the noise floor. And actually about the same with 24 bit. Amir has shown those graphs in some past reviews. And how good they look (except for some gear that couldn't really do more than 16 bit) is just down to noise levels.

 

[amirm]

And maybe compare TOSlinks vs USB the way JA does at Stereophile?

Testing other inputs doubles the work-load. JA does far fewer reviews than me, and measures less things. Seeing how most people use USB, am focused on that unless asked.

 

[LTig]

The arrow in the dash board says 2nd but points to 3rd harmonic.

 

[amirm]

The arrow in the dash board says 2nd but points to 3rd harmonic.

Yeh, I make that mistake often.

 

[Music1969]

but really what is the point of the 16 bit vs 24 bit the way JA shows it?

I agree there'd be no point, unless something unexpected results, as has been the case in many of JA's measurements over the years... otherwise I doubt even he would bother with 16bit vs 24bit either, if everything was always consistent and as expected...

And of course, even Amir has seen differences between SPDIF and USB before...

A couple examples below. Note I am not claiming any of these anomalies are audible - just that they may be interesting to see/know, versus not seeing/knowing...

Just one example of 16 bit vs 24 bit (both SPDIF in this case):

One example of USB vs TOSlink:

 

[VintageFlanker]

Thanks you so much @amirm !

It has been several months a lot of people were waiting for it... This even became a running gag on ADI-2 DAC's thread: "No ADI-2 DAC in SINAD Chart?!"

Still, I'm a bit sad to don't see RCA measurements (Which is the way I use mine).

In some sense, I think of the RME ADI-2 Pro

*ADI-2 DAC!

 

[Blumlein 88]

I agree there'd be no point, unless something unexpected results, as has been the case in many of JA's measurements over the years... otherwise I doubt even he would bother with 16bit vs 24bit either, if everything was always consistent and as expected...

And of course, even Amir has seen differences between SPDIF and USB before...

A couple examples below. Note I am not claiming any of these anomalies are audible - just that they may be interesting to see/know, versus not seeing/knowing...

Just one example of 16 bit vs 24 bit (both SPDIF in this case):

View attachment 26236

One example of USB vs TOSlink:

View attachment 26237

You examples on the jitter test show what I've been talking about. In 16 bit those spaced spikes are supposed to be there touching the green line. In 24 bit those spaced spikes are lower and below the noise floor of the device which is why they don't show up. You have a few unwanted spikes around the central tone in both cases.

In the USB vs Toslink example they are essentially identical other than those stray tones in the USB at -133 db and lower. It is mildly interesting, but otherwise not that revealing.

I like showing spectrograms of sweeps with different depths of the spectrogram. There are many things you can see in that one graphic. But it isn't common, and many would look and wonder what are they seeing. I'd like to see a spectrogram sweep with a depth of -100db and -120 db for each device (and -140 db if -120 were unusually clean). But Amir makes the most of what he is doing. So I can't expect him to show all these other views just in case something interesting is there.

 

[Tks]

Gotta contact RME and give them the good news. So they can get on working releasing a clear-top Anniversery Edition for the DAC like they did with the Pro

 

[Veri]

I agree there'd be no point, unless something unexpected results, as has been the case in many of JA's measurements over the years... otherwise I doubt even he would bother with 16bit vs 24bit either, if everything was always consistent and as expected...

And of course, even Amir has seen differences between SPDIF and USB before...

A couple examples below. Note I am not claiming any of these anomalies are audible - just that they may be interesting to see/know, versus not seeing/knowing...

Just one example of 16 bit vs 24 bit (both SPDIF in this case):

View attachment 26236

One example of USB vs TOSlink:

View attachment 26237

If you want the typical stereophile measurements I suggest you simple check out stereophile. They don't measure as many devices but they're still active. Wouldn't make sense for amir to simply copy the same measurement suite.. let amir do his own thing.

Besides that, 16-bit measurement as Blumlein is saying is not very valuable information. And toslink/spdif jitter is often measured by amir, when it is requested.

 

[Music1969]

In 24 bit those spaced spikes are lower and below the noise floor of the device which is why they don't show up. You have a few unwanted spikes around the central tone in both cases.

In the USB vs Toslink example they are essentially identical other than those stray tones in the USB at -133 db and lower. It is mildly interesting, but otherwise not that revealing.

Yes these two comments are exactly why I showed these examples... I guess I find these anomalies interesting (even if they are below audible levels as per my earlier post) because there are DACs that don't exhibit these anomalies at all...

Anyway it's up to @amirm to decide if he finds them interesting or not.

Just a friendly request.

 

[Tks]

Oh a request if it hasn't gone back to the owner. If possible, the classic THD+N Ratio vs FR? And just maybe the 32 input signal multi-tone test?

 

[KSTR]

For the THD and Noise measurements we should note that the DAC's volume control is digital, complemented by two only selectable (or auto-set) output gains for phones output (as compared to four settings for the XLR and RCA outs). The 4Vrms (+14dBu) level Amir has used can only be established in hi-gain mode (max V_out = +22dBu) and the DAC chip fed with about -8dBr.
This operating point sligthly favors distortion vs noise-floor because internal signal levels are lower. Running the DAC at -1dBFS would likely produce better S/N ratio but more distortion (see datasheet plots/tables). Combined SINAD probably stays where it is, or maybe even a tad better.

The RCA and XLR outputs are fed via four steps of available gains (or auto-set) and there is lesser impact of switching the gains on THD and noise. From the spec sheet I see only the lowest setting for the RCA outs (-5dBu) suffers a bit in S/N. Looks like the specs can be trusted.

 

[daftcombo]

Hi @amirm ,

1) I would be interested in you doing listening tests with the different filters (sharp delay sharp, sharp...) to decide which is better.
You can see here in the manual page 57:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=2ahUKEwjOrPe2kZ3iAhWOzYUKHYoGCCYQFjAAegQIAxAC&url=https://www.rme-audio.de/download/adi2dac_e.pdf&usg=AOvVaw3ypdhqKkVOsE6hXVNSrMI1

2) I read that the RCA out were superior to the XLR out. Could you confirm this?

 

[VintageFlanker]

Hi @amirm ,

1) I would be interested in you doing listening tests with the different filters (sharp delay sharp, sharp...) to decide which is better.
You can see here in the manual page 57:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=2ahUKEwjOrPe2kZ3iAhWOzYUKHYoGCCYQFjAAegQIAxAC&url=https://www.rme-audio.de/download/adi2dac_e.pdf&usg=AOvVaw3ypdhqKkVOsE6hXVNSrMI1

2) I read that the RCA out were superior to the XLR out. Could you confirm this?

1) "Better" when we talk about frequency response is the flattest possible. So the answer is already in the manual: Sharp and SD Sharp!

2) About RCA Vs XLR. It is the opposite. Amir first measured RCA out in the first review and then XLR (still, with the old AP) with slightly better results for linearity. But nothing audible, IMHO.