• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

RME ADI-2 DAC FS - AKM Versus ESS Measurements (DAC, Preamp & Headamp)

Scytales

Active Member
Forum Donor
Joined
Jan 17, 2020
Messages
129
Likes
182
Location
France
Bravo VintageFlanker !
Salut à toi ! Toujours sur la brèche, je vois :) !

Bravissimo to VintageFlanker. To make such detailed measurements do take time...

I have one important suggestion to improve this bench test results. Whenever dB are used, the 0 dB reference level must always been explicitly stated. Otherwise, the results are meaningless.

For instance, on the noise spectrographs and tables in the first page, the results means nothing without the reference level. Is it dBu (dB referenced to 0,7746 V RMS=0 dBu), dBV (dB referenced to 1 V RMS= 0 dBV) or another reference level ? When noise is measured, the measurement bandwidth should also clearly be stated (A-weighted implies the bandwith of the stated filter, but linear measurements do not).
 
Last edited:
OP
VintageFlanker

VintageFlanker

Major Contributor
Forum Donor
Joined
Sep 20, 2018
Messages
4,926
Likes
19,611
Location
Paris
if would be more credible to say for example THD is 112.1db +- 0.1db if that really is the case.
- Method : 8 runs for each test, then I choose the closest to the average
And if you wanna know, even 0.1dB was barely a margin of error in this specific test. Probably 6 runs out of 8 gave me the exact same THD+N in REW. RMAA results were unsurprisingly more variable, tho.
So I'm wondering how much your measurement variations are statistical noise due to sample to sample variations, temperature effects, grounding, cabling and connector effects, etc. as opposed to meaningful differences between the implementations
In fact, couldn't one reasonably expect to see similar differences between two randomly selected production units using either chip, simply based on internal component tolerances?
When you look closer at each graphs (REW FFT in particular), you know that we are outside sample-to-sample variations. Look at the bottom of the 1Khz tone, for example. Two same DACs may sure have disparities in harmonics or SNR between samples, but not different shapes like this.

Is English even your native?
Absolutely not, and I still find myself terrible at it.

If you still have the devices at hand
I don't.:(

@VintageFlanker , do you have an idea if the analog part of the DACs circuitry uses the same opamps and if they share the same circuit design?
I have no clue. Maybe @MC_RME does.;)

Am I misinterpreting something or did AKM's 2nd harmonic disappear in the RMAA's THD graph?
Not quite. It is there, but masked out by the 2nd harmonic of ESS. But yes, there is effectively much more difference between 2nd and 3rd harmonic seen by RMAA for AKM. When pushing levels higher, 2nd harmonic becomes more and more apparent. Here is at 0.0dBr/19dBu:
thd.png


I have one important suggestion to improve this bench test results. Whenever dB are used, the 0 dB reference level must always been explicitly stated. Otherwise, the results are meaningless.

For instance, on the noise spectrographs and tables in the first page, the results means nothing without the reference level. Is it dBu (dB referenced to 0,7746 V RMS=0 dBu), dBV (dB referenced to 1 V RMS= 0 dBV) or another reference level ? When noise is measured, the measurement bandwidth should also clearly be stated
I'm sure reading the review from the beginning should give you some answers...;)

This brings me to a simple question : I suppose that make definitely no difference for listening ? Well, we all know the answer I guess, but I would appreciate a confirmation by VF.
I did not conduct a DBT, so impossible to know. Did not catch any audible difference with headphones either. A significant audible variation would have been seen in my data, anyway.

What were the settings and wiring of the ADC ?
Regular pair of XLR cables. 4.5Vrms (for 15dBu testing) input for RMAA and Multitone (Stereo Mode). Y split XLR cable for Mono Mode (3.5Vrms input) in REW. Minimum Phase Filter.
 
Last edited:

DanTheMan

Active Member
Joined
Oct 19, 2020
Messages
290
Likes
438
All I can say for sure is that my next audio interface will be RME. My old one (9 years old) doesn’t cooperate with the M1, so it’s time.
 

Trell

Major Contributor
Joined
May 13, 2021
Messages
2,752
Likes
3,284
All I can say for sure is that my next audio interface will be RME. My old one (9 years old) doesn’t cooperate with the M1, so it’s time.

Which one do you have in mind?
 

RandomEar

Senior Member
Joined
Feb 14, 2022
Messages
331
Likes
765
I happen to have both AKM and ESS units and in an uncontrolled but level matched test I thought that AKM was actually less mellow. I realize this is pretty meaningless given the measurements but that what my brain decided at the time, now I can’t shake it off :)
If you'd like to shake it off, though, I would recommend to listen to more Taylor Swift ;)
 

nagster

Senior Member
Joined
Jan 24, 2021
Messages
361
Likes
592
Salut à toi ! Toujours sur la brèche, je vois :) !

Bravissimo to VintageFlanker. To make such detailed measurements do take time...

I have one important suggestion to improve this bench test results. Whenever dB are used, the 0 dB reference level must always been explicitly stated. Otherwise, the results are meaningless.

For instance, on the noise spectrographs and tables in the first page, the results means nothing without the reference level. Is it dBu (dB referenced to 0,7746 V RMS=0 dBu), dBV (dB referenced to 1 V RMS= 0 dBV) or another reference level ? When noise is measured, the measurement bandwidth should also clearly be stated (A-weighted implies the bandwith of the stated filter, but linear measurements do not).
In the case of normal simple substance measurement, you are of course correct. Bandwidth, the exact absolute value of 0dB, etc. need to be clarified.
Since the main purpose of this time is to know the difference between the two units, the provided information is sufficient if the measurement conditions are the same.
 

Scytales

Active Member
Forum Donor
Joined
Jan 17, 2020
Messages
129
Likes
182
Location
France
I have one important suggestion to improve this bench test results. Whenever dB are used, the 0 dB reference level must always been explicitly stated. Otherwise, the results are meaningless.

For instance, on the noise spectrographs and tables in the first page, the results means nothing without the reference level. Is it dBu (dB referenced to 0,7746 V RMS=0 dBu), dBV (dB referenced to 1 V RMS= 0 dBV) or another reference level ? When noise is measured, the measurement bandwidth should also clearly be stated
I'm sure reading the review from the beginning should give you some answers...;)
I read that you made measurements with signals at 15 dBu at the DACs outputs, but I do not think that answers the issue about the absolute noise level measurements, which are made without signals. What is the 0 dB reference of the scale of the RMAA software apparently used to make the noise measurements ? The terme "RMS power" used to tabulate the results under the spectrographs is confusing. Are the measurements made in a power unit (for instance dBm) or a voltage unit (dBu, dBV, or others) or something else, as a purely digital measurement (dBFS) ? Is there a detailled RMAA spec sheet or user manual that adresses those issues ?
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,318
Likes
3,273
Location
.de
At last, a quick test out of the headphones amp. Note that it is only a run at fixed revel (which basically is @2.8Vrms) with no load.

Headphones out (High Power, -7dB Level)
Test
AKM HF High -7dB
ESS HF High -7dB
Frequency response (from 40 Hz to 15 kHz), dB
+0.04, -0.02​
+0.08, -0.02​
Noise level, dB(A)
-103.6​
-102.5​
Dynamic range, dB(A)
103.6​
102.6​
THD, % (RMAA)
0.00018​
0.00014​
THD + Noise, dB (REW)
-101.6​
-100.7​
IMD + Noise, %
0.00256​
0.00279​
Stereo crosstalk, dB
-92.1​
-91.6​

AKM Headphones High -7dB - THD+N-merged.jpg

Note: @MC_RME had remarks about these plots. They do not match his results at all (same setting, High Power, -7dBr). It should reach 109dB SINAD with a proper rig, and under 32, 300, or 600 Ohms loads. One of the issues in my data is obviously the lack of load board, IMHO, and the still-not-optimal use of the Cosmos when single ended is in question. Anyway, these are only intended to catch "differences" between the two DUTs, certainly not to evaluate pure performance of the (amazing) ADI-2 DAC headamp.
That smells of issues with common-mode rejection, or even a plain ground loop. You'll want to dig up the discussion that I had with @pma on the matter recently. (EDIT: In this thread.) In a nutshell, this measurement requires either an input with reasonably high common-mode input impedance even when there is no output ground to input shield connection, or galvanic isolation on the USB side (plugging the ADC into a separate battery-operated laptop also counts). The "barefoot" Cosmos ADC has low input impedance as we all know, and if no particular constructive measures are taken, differential mode and common mode input impedance are directly related.

You have a pretty good chance of ironing this out with the APU's notch filter section if cabling is adequate.
 

junki

Member
Joined
Mar 24, 2018
Messages
55
Likes
43
Looks like a bunch of charts that show there is no experiential difference between the two.
 

Trell

Major Contributor
Joined
May 13, 2021
Messages
2,752
Likes
3,284
Looks like a bunch of charts that show there is no experiential difference between the two.

There is no audible difference between them, if that is what you imply.
 

ezra_s

Active Member
Forum Donor
Joined
Mar 23, 2020
Messages
293
Likes
327
Location
Spain
wonderful review.
 

Ra1zel

Addicted to Fun and Learning
Joined
Jul 6, 2021
Messages
530
Likes
1,045
Location
Poland
RME scale this to 8 channels at least, pretty please.
 
Top Bottom