RME ADI-2/4 Pro SE - Auto-measured with Multitone

[pkane]

Thank you for these measurements! I’ve got two questions:

What is DIM? Some kind of intermodulation I gather, but I haven’t seen that acronym before.

The “THD vs Level” plot has me scratching my head. Wouldn’t the noise graph have to be a perfect complement of the distortion graph to make the THD+N such a straight line?

DIM 30 and 100 are dynamic intermodulation tests from IEC 60268:3 spec. I think AP website has a note on the specifics, since the IEC document itself is not cheap.

When THD is very low, which it is for this converter, THD+N is dominated by noise. Noise being fairly constant, THD+N is inversely proportional to the level of the fundamental.

 

[morillon]

What does this have to do with measuring how a DAC handles a specific test signal? Please post in the threads about intersample overs if you want to discuss that subject.

Sorry..
a classic situation of the net because not the subject of course...
- Of course, I have nothing against this test.. I realize the interest...
- nothing against the fact that your software offers it....

I just imagine knowing this environment, to what extent starting from a particular, rather specific thing... we can quickly transform it into a "" necessary condition ""... even if largely out of context...
etc

but end...!!!
and thank you for your software so rich of so many tests!!!
(even if even if... I know a few missing ones..na! and you know! ;-)) joke )

 

[Rja4000]

Odd.
I tried your test plan and the first test is getting saturated at 0dBFS (+24dBu ranges)

I've run the same test with MI and levels are OK.
For +13dBu ranges, it's all good.
I don't understand why.

Other than that, the plan works.

I'll publish results when it will have run through all tests.
(I modified the test to run in full mono. I just changed output channels to L+R, actually.)

Question:
You don't have THD+N in Distortion test ?
I also see Noise > -SNR, which is strange.
(Same for the plot you published here)

 

[pkane]

Odd.
I tried your test plan and the first test is getting saturated at 0dBFS (+24dBu ranges)

I've run the same test with MI and levels are OK.
For +13dBu ranges, it's all good.
I don't understand why.

Other than that, the plan works.

I'll publish results when it will have run through all tests.
(I modified the test to run in full mono. I just changed output channels to L+R, actually.)

Question:
You don't have THD+N in Distortion test ?
I also see Noise > -SNR, which is strange.
(Same for the plot you published here)

Which test shows saturation @0dBFS? And what does that look like?

THD+N is the same as TD+N when used with a single tone.

 

[Rja4000]

Which test shows saturation @0dBFS? And what does that look like?

THD+N is the same as TD+N when used with a single tone.

First test: 1kHz distortion
I tried several times and double checked everything

The tests are still running.
Ill publish the results when they're finished.

 

[Rja4000]

RME ADI-2/4 Pro SE
All tests L+R output, Mono input (M/S mode)
Output and input ranges +13dBu

The test plan stopped during THD vs Level

Comparing to your tests, Noise is lower (as expected).

EDIT:
Added some comments and more plots (and added the correct Crosstalk measurement)

SINAD 1kHz

Linearity
NB: This is NOT comparable to Amir's results
Amir is using a lower frequency and a bandpass filter.
On top of it, the reference is inversed:
Amir has a positive value if level (Input level - Output level) is above the same delta at -20dBFS level.
it should raise above 0dBr, due to the effect of noise, when the value is very low.

Here is my own attempt to the same @200Hz, with very narrow bandpass filter (This is a bit extreme filtering here) :
NB: Vertical scale is in milli dB, not in dB as above

Jitter

Multitone 32
Again, this is not stricly comparable to Amir's Multitone (although there should be less difference here)

Here is an attempt with the same Multitone Amir is using (generated by Audio Precision software)
Note the lowest tone at 14Hz
I also use "No window" FFT, as the AP test signal is meant for it.

IMD DIN vs Level
Amir is using IMD SMPTE, which is similar but uses different frequencies
Noise level is VERY low. So what may look like a distortion raise is only visible because the noise is low.
For lesser devices, it would be hidden by the noise.

THD vs Frequency
Again, this is not to be compared to Amir's measurements, which use 90kHz Bandwidth to show distortion at high frequencies.

Noise floor

DIM30 was not correct, due to a bug in Multitone

Here is the correct one, after bug correction

Crosstalk
The following plot is completely wrong (my fault)

Here is the real one:
Slightly better

IMD SMPTE vs level
Same comment than for IMD DIN

Intersample distortion
At high level, distortion is expected.
This test should be re-run at loer level

Frequency response
I get "DC" alert on frequency Log sweep.
When I disable all filters, I get this:

I can't zoom on the impulse response either

Here is the SINAD Dashboard, with another tool

@pkane
Fantastic tool !
Awesome work !

 

[Sokel]

Tested it too (not with RME sadly) mainly to steal the settings
Works really good!

(couple of screens):

 

[Rja4000]

Tested it too (not with RME sadly) mainly to steal the settings
Works really good!

(couple of screens):

With the title saying "RME ADI-2/4 Pro SE", it's a bit confusing. ;-)
What device did you measure ?

 

[Sokel]

With the title saying "RME ADI-2/4 Pro SE", it's a bit confusing. ;-)
What device did you measure ?

Oh,I only wanted to see how the plan goes and the settings for each measurement.
E-MU of course!

(it's nice to look how 13 years of progression goes)

 

[VintageFlanker]

AFAIC, I got 118+dB SINAD (19dBu) with my ADI-2/4 PRO SE measured by itself in REW RTA. Same goes for both XLR and 4.4mm headphones out.

DAC performance aside, max I could measure from its ADC was this (44.1kHz):

 

[Rja4000]

First test: 1kHz distortion
I tried several times and double checked everything

I retried and I don't get the saturation at 24dBu anymore.

Here is what I was getting.

And here is what I get now (which is more in line with expected)

I can't explian it, sorry.

 

[Rja4000]

AFAIC, I got 118+dB SINAD (19dBu) with my ADI-2/4 PRO SE measured by itself in REW RTA. Same goes for both XLR and 4.4mm headphones out.

DAC performance aside, max I could measure from its ADC was this (44.1kHz):
View attachment 275798

Here is what I get in 19dBu (Loopback Mono, @997Hz 20Hz-20kHz, 44.1kHz sampling rate)

And with the Topping D90LE (5V Mono) as a DAC @-3.3dB (approx 3.5V) - RME ADI-2/4 Pro SE ADC +13dBu input range Mono

 

[Rja4000]

Same with Multitone

and REW

@pkane
As above: I don't understand why your "Noise" figure is so high.
Should be around -121.4dB here.

 

[Sokel]

RME ADI-2/4 Pro SE
All tests L+R output, Mono input (M/S mode)
Output and input ranges +13dBu

The test plan stopped during THD vs Level

Comparing to your tests, Noise is lower (as expected).
View attachment 275758
View attachment 275760

View attachment 275761


View attachment 275759


View attachment 275763

View attachment 275764
Then I restarted
View attachment 275788
View attachment 275789
View attachment 275790
View attachment 275792


View attachment 275791

NB: Can't change the zoom for Waveform (maybe because my regional settings uses , for decimal separator ?)

I get "DC" alert on frequency Log sweep.

When I disable all filters, I get this:
View attachment 275793
View attachment 275794


I can't zoom on the impulse response either
View attachment 275795


I'll re-run those tests with other tools tomorrow, to compare.


Here is the SINAD Dashboard, as an appetizer.

View attachment 275797



@pkane
Fantastic tool !
Awesome work !

I think your DIM30 is wrong or mislabeled.
Requires 30kHz bandwidth and can't see the rest of the tones.

 

[Rja4000]

I think your DIM30 is wrong or mislabeled.
Requires 30kHz bandwidth and can't see the rest of the tones.

It's not my test, but I agree.

Here is what I get with Multitone for DIM30

And for DIM100

(Unfortunately, I can't see the wave form)

And with my own test at the same -15dB
We see clearly the fundamental Square wave at 3.15kHz and its split in odd harmonics.
(The signal was generated by the AP Wave generator.)

EDIT: There was a bug in Multitone Analyzer
@pkane corrected it.

Here is DIM30 now

 

[Sokel]

It's not my test, y I agree.

Here is what I get with Multitone for DIM30
View attachment 275814

And for DIM100
View attachment 275815
(Unfortunately, I can't see the wave form)

And with my own test at the same -15dB
We see clearly the Square wave split in odd harmonics.
(The signal was generated by the AP Wave generator.)

View attachment 275816

Strange...
That's what I get either from the plan or not:

From plan

 

[Rja4000]

It looks like the signal is wrong for me.

I ran it by removing the other tests from the test plan at the beginning of this thread.

 

[Sokel]

Test plan's BW is wrong by the way,@pkane has told me it should be 30Khz for DIM30 and 100Khz for DIM100.

 

[Rja4000]

Also, the test labeled "IMD SMPTE vs Frequency" is actually IMD SMPTE vs Level.

 

[pkane]

RME ADI-2/4 Pro SE
All tests L+R output, Mono input (M/S mode)
Output and input ranges +13dBu

The test plan stopped during THD vs Level

Comparing to your tests, Noise is lower (as expected).
View attachment 275758
View attachment 275760

View attachment 275761


View attachment 275759


View attachment 275763

View attachment 275764
Then I restarted
View attachment 275788
View attachment 275789
View attachment 275790
View attachment 275792


View attachment 275791

NB: Can't change the zoom for Waveform (maybe because my regional settings uses , for decimal separator ?)

I get "DC" alert on frequency Log sweep.

When I disable all filters, I get this:
View attachment 275793
View attachment 275794


I can't zoom on the impulse response either
View attachment 275795


I'll re-run those tests with other tools tomorrow, to compare.


Here is the SINAD Dashboard, as an appetizer.

View attachment 275797



@pkane
Fantastic tool !
Awesome work !

Looks really good, @Rja4000 ! I'll see about the zoom window with different regional settings. You can always use the mouse scroll wheel to zoom in on any point in the graph (or on individual axis).

[EDIT: no issues with regional settings, and I can zoom in and out with no problems. Maybe you can clarify what you're trying to do when zoom isn't working?]