Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   Gear Shoot-Outs / Sound File Comparisons / Audio Tests (https://www.gearslutz.com/board/gear-shoot-outs-sound-file-comparisons-audio-tests/)
-   -   Evaluating AD/DA loops by means of Audio Diffmaker (https://www.gearslutz.com/board/gear-shoot-outs-sound-file-comparisons-audio-tests/607481-evaluating-ad-da-loops-means-audio-diffmaker.html)

didier.brest 13th November 2011 06:17 PM

Very interesting results: external clocking does not have any impact.

Here below the updated list for single ADDA converter. I will add another list for results from two converters sets later on.


Mytek 8 x 192 (MainTime)
-1,002sec, 0,450dB (L), 0,553dB (R)..Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
-750usec, -0,085dB (L), -0,080dB (R)..Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
-2,286msec, 0,524dB (L), 0,817dB (R)..Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
-1.564usec, 0.463dB (L), 0.461dB (R)..Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
-1sec, 0,075dB (L), 0,064dB (R)..Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
-36,94usec, 0,298dB (L), 0,311dB (R)..Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

Echo Audiofire 8 with ADAT clocked to an RME HDSP 9632 (schtim)
-1.164sec, 0.128dB (L), 0.126dB (R)..Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire 2626 (cylens)
-13,21usec, -0,505dB (L), -0,486dB (R)..Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference; -44,0 dBFS -44,9 dBFS

Echo Audiofire 8 non adat CS4272 (guze)
-2,952msec, 0,840dB (L), 0,885dB (R)..Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
-5,645msec, 0,976dB (L), 0,962dB (R)..Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
-2,101msec, 1,104dB (L), 0,605dB (R)..Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
-8,95msec, 1,015dB (L), 0,605dB (R)..Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
1,555usec, 0,352dB (L), 0,414dB (R)..Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)

Quint 13th November 2011 08:11 PM

This is an interesting test indeed. I'd be interested to see where the SSL Alphalink fits into all of this and maybe some older stuff like Radar?

Also, everyone's ears are different so one person might be able to hear a difference down at -50 dbfs while another may not. That being said and skeptical of it as I might be, for this type of scenario, has anyone ever established any kind of a null threshold in which anything below that threshold is effectively inconsequential?

didier.brest 15th November 2011 10:41 PM

Evaluating DA-AD loops from this wav file by means of Audio DiffMaker.


ADDA converters


Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Please give me the link for downloading an AD-DA loop achieved by any AD-DA converter, even if it is already in the list: it is is interesting for checking the consistency of the performance. Lavry, UA, RME UFX especially welcome.

guze 16th November 2011 03:08 AM

audiofire12.wav

Audiofire12

keep going

didier.brest 16th November 2011 08:41 AM

Thank you guze!

Evaluating DA-AD loops from this wav file by means of Audio DiffMaker.


ADDA converters


Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Please give me the link for downloading an AD-DA loop achieved by any AD-DA converter, even if it is already in the list: it is is interesting for checking the consistency of the performance. Lavry, UA, RME UFX especially welcome.

Timur Born 18th November 2011 04:05 PM

Let's put some more interpretable facts on the table! ;)

This is the measured filter slope of the Babyface's inputs:

http://i.imgur.com/jhJD5.png

As you can see the -0.5 dB point is at around 45 Hz, the -1 dB point is at around 30 Hz.

Furthermore the Babyface's Input is on average 2 dB hotter compared to its output when set to 0 dB gain. With pure sinusoidal signals the peak difference is about 2.0 - 2.3 dB. So any loop-back measurements must be made with the *output* set to at least -2.3 dB, better use some more for possible variance.

http://i.imgur.com/LlYg6.png

With pure square signals you need to pull down the output level even more. At 1 KHz it's the same -2.3 dB as with sinusoidal, at 50 Hz you need to go as low as -6 dB, and all the way down to 1 Hz you need to use -9 dB. That is because the lower the frequency the more the square will overshoot at its ascending ramp.

http://i.imgur.com/bPFxO.png

So before we go on discussing you need to do new measurements with modified output levels, preferably while running your test file through the loop to even out the levels. Else any loop-back measurements, be they made with Audio DiffMaker or other software, that does not apply this output level reduction will measure input *distortion*!

didier.brest 18th November 2011 05:51 PM

Quote:

So before we go on discussing you need to do new measurements with modified output levels
I applied this output level reduction in the Software Playback gain setting. AFAIR it was a bit larger than 2 dB. In my last list of results, I remove the values of the delay and gain corrections applied by Audio DiffMaker to the tested loop before generating the difference with respect to the original. Theses values mentioned in a previous post (November 13th) were :

RME Babyface (didier.brest)
1,555usec, 0,352dB (L), 0,414dB (R)..Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)

Hence the Babyface loop was about 0.4 dB quieter than the original.

I got in touch by PM with another Babyface owner who got similar (a few tenth of dBs worse...) results. However if you have got the Babyface, would you please record a loop from the original file used for this test ?

_ben 18th November 2011 06:33 PM

Quote:

Originally Posted by nms (Post 7224971)
Because white noise isn't generated identically every time and because RMAA uses an assortment of tones and analyzes them itself in a way that is altogether a lot more accurate and relevant than anything you are going to pull off. How's that? When people test gear to get stats they don't do it the way you are trying to. There's professionally designed software for that. I'm floored that you actually wasted time trying to null 2 independently generated white noise clips! Anyhow.. this isn't the place to further discuss it though so go explore it and you'll understand pretty quickly why you want to use that.

Your "because" is just wrong. I explained what i do and why. The fact that you can't understand it speaks a lot.
By the way as i said in this other thread (i think you read this thread too) the white noise was recorded.
https://www.gearslutz.com/board/gear-...ml#post7244975

The fact is that you push a button and read parameters but you don't really know what happends. I'm sorry but i don't call this science.

The corr depth is calculated with a collection of datas, it is by definition a statistic result.
I will give you a small example with a statistic result which show you how wrong it can be interpreted.

Timur Born 18th November 2011 08:02 PM

Quote:

Originally Posted by didier.brest (Post 7244870)
I applied this output level reduction in the Software Playback gain setting. AFAIR it was a bit larger than 2 dB.

Yes, that's why I wrote you need to run your own testfile to find the correct reduction. There is no *one* value, because it depends largely on how transient the signal is (see square vs. sinusoid, triangular only needs -1.7 dB btw).

Quote:

However if you have got the Babyface, would you please record a loop from the original file used for this test ?
I will download and run the test later. The inherent flaw (albeit correct result) of the DiffMaker with the Babyface is that it will always report the input filter slope as *vast* difference. Again, it *is* correct in that, but once we know about the input filter what other results can the software provide?

Here is an example of how much a difference this filter alone can make in a null/phase-inversion test:

Material with low-pass filtered content (aka low signal dB in sloop area of BF's input filter)
http://i.imgur.com/oT4AA.png

Same material without the low-pass filter
http://i.imgur.com/rsh4i.png

nms 18th November 2011 08:42 PM

Quote:

Originally Posted by _ben (Post 7245000)
Your "because" is just wrong. I explained what i do and why. The corr depth is calculated with a collection of datas, it is by definition a statistic result..

What are you on about? Correlation depth? You're quoting a feature from the wrong program (Diffmaker) and saying I am the one who doesn't know what I'm talking about? Do you really not understand that RMAA is software professionally designed to do what you are trying to do? lol.. What are you afraid of? If your results are real then they will be verified in the tests that RMAA does.. Which uses several tones to test properly, not just some white noise, and provides accurate specs across the board including freq response, noise floor, crosstalk, distortion and lots more.

Btw.. are the cables you're testing both terminated the same (both XLR or both TRS)?

_ben 20th November 2011 02:36 PM

I just get a TC Twin Impact so i made the test with AudioDiff to compare it to an EMU1616M.
I used the original file provided by Didier in mono 44khz/16bits.
Here are the settings for AudioDiff :

http://img259.imageshack.us/img259/4981/settingsq.jpg


Result with TC Twin Impact after one pass :
>> ********************************
>> (not configured for EQ response correction)
>> ---Beginning difference extraction---
>> Files are monophonic, at 44,1kHz sample rate, about 121,1 seconds long.
>> The Reference track is 16 bits; the Compared Track is 16 bits
>> Using Large memory algorithms
>> Finding the spectra of the "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 8388608]
>> Correlating the "Compared" with the "Reference"
>> [inverse fourier transform: size 8388608]
>> Coarse Time Offset = 0sec.
>> Move track subsections to freqency domain
>> Transforming overlapped portions of "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 131072]
>> [dual fourier transform: size 2x 131072]
>> [fourier transform: size 131072]
>> [fourier transform: size 131072]
>> Calculating energy in Reference track
>> Beginning fine determination of time offset between tracks
>> Finished fine determination of time offset between tracks
>> Beginning fine determination of gain difference between tracks
>> Finished fine determination of gain difference between tracks
>> Aligning the "Compared" track to the "Reference" track
>> Sample Rate error is -0,0052 ppm, adjustment unnecessary
>> [inverse fourier transform: size 8388608]
>> Saving the aligned COMPARE file and calculating the Difference track
>> Calculating Correlated null depth.
>> [fourier transform: size 8388608]
>> [fourier transform: size 8388608]
>> Complete: saving the "Difference" WAV file
>> Parameters: -118nsec, -0,001dB.Corr Depth: 30,0 dB


Result with EMU1616M after one pass :
>> ********************************
>> (not configured for EQ response correction)
>> ---Beginning difference extraction---
>> Files are monophonic, at 44,1kHz sample rate, about 121,1 seconds long.
>> The Reference track is 16 bits; the Compared Track is 16 bits
>> Using Large memory algorithms
>> Finding the spectra of the "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 8388608]
>> Correlating the "Compared" with the "Reference"
>> [inverse fourier transform: size 8388608]
>> Coarse Time Offset = -113,4usec.
>> Move track subsections to freqency domain
>> Transforming overlapped portions of "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 131072]
>> [dual fourier transform: size 2x 131072]
>> [fourier transform: size 131072]
>> [fourier transform: size 131072]
>> Calculating energy in Reference track
>> Beginning fine determination of time offset between tracks
>> Finished fine determination of time offset between tracks
>> Beginning fine determination of gain difference between tracks
>> Finished fine determination of gain difference between tracks
>> Aligning the "Compared" track to the "Reference" track
>> Correcting "Compared" Sample Rate, adjust by -0,0830 ppm, avg.
>> [multiple inverse fourier transform: size 4x 8388608]
>> Saving the aligned COMPARE file and calculating the Difference track
>> Calculating Correlated null depth.
>> [fourier transform: size 8388608]
>> [fourier transform: size 8388608]
>> Complete: saving the "Difference" WAV file
>> Parameters: -108,8usec, -2,886dB rate adj=-0,0830 ppm.Corr Depth: 29,8 dB

TC Twin Impact :
Parameters: -118nsec, -0,001dB.Corr Depth: 30,0 dB

EMU1616M :
Parameters: -108,8usec, -2,886dB rate adj=-0,0830 ppm.Corr Depth: 29,8 dB

I wonder how you can get 32db with the TC Twin Impact after 10 pass when i get only 30db after one pass.

After this test the EMU1616M = the TC Twin Impact which is in the top of converters (according to your evaluations)....big lol.

By the way, the headphone output (on the dock) sounds slightly clearer and detailled with the EMU than with the TC Impact.
Due to the fact that the headphone output on the pcmcia card sounds better than the dock version, the monitoring of the EMU is far better than the one available for the TC Impact.
The patchbay of the EMU is more powerful.
I have not compared the preamps for the moment but i already know that i don't like the ones of the EMU.

didier.brest 20th November 2011 04:19 PM

Quote:

Originally Posted by _ben (Post 7250141)
I wonder how you can get 32db with the TC Twin Impact after 10 pass when i get only 30db after one pass.

1) There is only one pass in the test reported in this thread.
2) Could you upload at some file sharing site (mediafire for instance) the wav files of the loops from the TC and the EMU so that I can add their results here ? Thanks in advance.jkthtyrt
3) My correlated depth figures are got with the default setting of Audio diffMaker that limits the measurement to 100 Hz-12 kHz. This may explain the 2 dB discrepancy between both our figures for the TC.

didier.brest 20th November 2011 04:55 PM

Quote:

Originally Posted by Timur (Post 7245280)
Yes, that's why I wrote you need to run your own testfile to find the correct reduction. There is no *one* value, because it depends largely on how transient the signal is (see square vs. sinusoid, triangular only needs -1.7 dB btw).

I made another measurement with -10 dB software playback gain. The result is

8,087dB (L), 8,151dB (R)..Corr Depth: 9,1 dB (L), 10,8 dB (R) Difference: -28,7 dBFS (L) -29,7 dBFS (R)

which is within 0.1 dB the same like my first result got with about -2 dB gain.

I might make another measurement with the Fireface 400 using the preamp imputs (1-2) so that to check what is the preamp impact. But the Babyface DA ---> Fireface AD loop performs bad also, which seems to show that the Babyface preamp is not the only issue.

Current results:

Evaluating DA-AD loops from this wav file by means of Audio DiffMaker.


ADDA converters


Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Please give me the link for downloading an AD-DA loop achieved by any AD-DA converter, even if it is already in the list: it is is interesting for checking the consistency of the performance. Lavry, UA, RME UFX especially welcome.

Timur Born 20th November 2011 05:22 PM

Once I understood how to use this software properly I will post my own results. But from what I saw yet the "higher" results you are getting for the BF are simply a result of the BF input's high-pass filter.

See my realtime/manual phase-inversion test above (orange graphs) to get an idea how the input filter alters the readings. Filter your testfile and repeat the test and you'll likely get a lot better results.

farnouts 20th November 2011 05:30 PM

Quote:

But the Babyface DA ---> Fireface AD loop performs bad also, which seems to show that the Babyface preamp is not the only issue.
Didier, could you do the test Babyface A/D --> FF400 D/A ?

I also posted a link to RME forums, the RME support answers are quite interesting...

_ben 20th November 2011 06:23 PM

Quote:

Originally Posted by didier.brest (Post 7250318)
3) My correlated depth figures are got with the default setting of Audio diffMaker that limits the measurement to 100 Hz-12 kHz. This may explain the 2 dB discrepancy between both our figures for the TC.

The result will be lower :
>> ********************************
>> (not configured for EQ response correction)
>> ---Beginning difference extraction---
>> Files are monophonic, at 44,1kHz sample rate, about 121,1 seconds long.
>> The Reference track is 16 bits; the Compared Track is 16 bits
>> Using Large memory algorithms
>> Finding the spectra of the "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 8388608]
>> Correlating the "Compared" with the "Reference"
>> [inverse fourier transform: size 8388608]
>> Coarse Time Offset = -1,678msec.
>> Move track subsections to freqency domain
>> Transforming overlapped portions of "Reference" and "Compared" tracks
>> [dual fourier transform: size 2x 131072]
>> [dual fourier transform: size 2x 131072]
>> [fourier transform: size 131072]
>> [fourier transform: size 131072]
>> Calculating energy in Reference track
>> Beginning fine determination of time offset between tracks
>> Finished fine determination of time offset between tracks
>> Beginning fine determination of gain difference between tracks
>> Finished fine determination of gain difference between tracks
>> Aligning the "Compared" track to the "Reference" track
>> Sample Rate error is -0,0062 ppm, adjustment unnecessary
>> [inverse fourier transform: size 8388608]
>> Saving the aligned COMPARE file and calculating the Difference track
>> Calculating Correlated null depth.
>> [fourier transform: size 8388608]
>> [fourier transform: size 8388608]
>> Complete: saving the "Difference" WAV file
>> Parameters: -1,673msec, -3,076dB.Corr Depth: 29,1 dB


Reference file mono/normalized :
originalmono.wav

Result with TC Impact :
originaltc.wav

Result with EMU1616M :
originalEMUVovox.wav

didier.brest 20th November 2011 08:26 PM

All the previous results have been got from this stereo file.
Please could you do a stereo loop from this file ?diddlydoo

didier.brest 20th November 2011 09:20 PM

Quote:

Originally Posted by farnouts (Post 7250438)
Didier, could you do the test Babyface A/D --> FF400 D/A ?

I also posted a link to RME forums, the RME support answers are quite interesting...

You rather mean FF400 D/A --> Babyface A/D ?
I do not have any Jack/XLR male cable.wworried

Thank you for your notice about the RME forum. I am also a member of this forum (didier) but I prefer answering here to some objections:

1) Is my Babyface OK ?
Another member of this forum did a loop with its own Babyface and got results worse than mine. The test file is available to anybody. So testing any Babyface is quite easy. I am of course interested in evaluating a loop from another Babyface. By the way, I think my Fireface is OK and I enjoy its sound everyday.

2) Impact of power supply
I tested both the USB power supply and an external one and got similar results.

3) Audio DiffMaker gives only one figure
Not exact: Audio DiffMaker provides also the difference file, the RMS level of which I measure in Wavelab. Here below the frequency spectra of these differences for the Fireface 400 (white) and the Babyface (red). One can see that the gap does not depend much on frequency, which, as an addition to the low performance from the BabyfaceD/A ---> Fireface A/D loop, makes doubtful the LF response from the preamp being the main cause of the lower performance from the Babyface.

http://www.mediafire.com/imgbnc.php/...7b983e156g.jpg

_ben 20th November 2011 09:53 PM

Quote:

Originally Posted by didier.brest (Post 7250885)
All the previous results have been got from this stereo file.
Please could you do a stereo loop from this file ?diddlydoo

I used one side of this file (normalized) it's more simple.
Can't see the interest of doing stereo.

Timur Born 20th November 2011 10:48 PM

Quote:

Originally Posted by didier.brest (Post 7251040)
Here below the frequency spectra of these differences for the Fireface 400 (white) and the Babyface (red).

I still have to wrap my head around DiffMaker, but your "difference" file makes me wonder if source and destination were aligned correctly?

Have a look at my manual phase-inversion test above (orange graph). Doing a BF -> BF loopback got me close to your "white" FF400 measurement as soon as the low frequency filter part was taken out of the loop. Interestingly once the low frequency filter is active (content present in that band) the high frequency areas' differences increase, too (harmonics I guess).

didier.brest 20th November 2011 11:09 PM

The time alignment is determined by Audio Diiffmaker. I do not see why this measurement would be bad only (and repetitively) for the Babyface.

Result from comparison of the test file with its copy to which I applied -3 dB gain and added -62 dBFS RMS white noise:

78,77nsec, 3,001dB (L), 3,000dB (R)..Corr Depth: 76,4 dB (L), 89,9 dB (R)

The accuracy of the time alignment in this case is about the sample duration divide by 300 and the accuracy of the gain correction is 0.001 dB: not so bad...

nms 21st November 2011 12:14 AM

I hope no one is surprised by differing results if they're not running the same Diffmaker settings as didier. This obviously means testing in stereo amongst other things.
I run my tests with the range set to 40k-18k which is more accurate for what we're doing since performance below 100hz and above 12k is important to us. Aside from that though it shouldn't be too shocking if some units test lower. Much of the expectation of these units comes from comments passed around here from people who are testing with bias of preconceived notions, subjectivity, and an approach that's generally not all that reliable. You can read pages and pages of praise for the MR816 yet I've never seen one person mention the low frequency roll off or fact you get a different frequency response depending on whether you use 1/4" or XLR's on it!

didier.brest 21st November 2011 12:31 AM

I think that the frequency-range setting in Audio DiffMaker is not only the one in which the correlated depth is evaluated but also the one for determining the time alignment and gain corrections, the accuracy of which is much critical for the result. This is likely the reason why the default setting is 100 Hz-12 kHz: below 100 Hz and above 12 kHz, the signal-to-noise ratio is generally lower so that limiting the mesurement to 100 Hz - 12 kHz may make the time alignment and the gain correction more accurate.

Anyway, whatever is this setting, the wav difference file provided by Audio DiffMaker is full band. So my measurements of the difference RMS level, on which I rely for the ranking, are not impacted directly by (or, because of the more accurate time alignment and gain correction, get benefit from) this frequency bounds.

Corr Depth: 29,1 dB (L), 29,1 dB (R) Difference -48,5 dBFS (L)

New list icluding TC Impact and E-MU 1616m tests from ben

Evaluating DA-AD loops from this wav file by means of Audio DiffMaker.


ADDA converters


Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Impact Twin (left channel only, ben)
Corr Depth: 29,1 dB (L) Difference: -48,5 dBFS (L)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

E-MU 1616m (left channel only, ben)
Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Please give me the link for downloading an AD-DA loop achieved by any AD-DA converter, even if it is already in the list: it is is interesting for checking the consistency of the performance. Lavry, UA, RME UFX especially welcome.

_ben 22nd November 2011 07:40 AM

_En stéréo 24 bits normalisé dans Ableton, le fichier source :
OriginalNorm.wav

_Tous les enregistrements sont normalisés avec Ableton.

_1-2-3 fois le même enregistrement :
OriginalTcStereo24-1.wav
http://www.mediafire.com/?6fg1uieo2gnf00j
OriginalTcStereo24-3.wav


_ j'ai interverti les câbles :
OriginalTcStereo24-Swap.wav

_ avec le driver Asio4All :
OriginalTcStereo24-Asio1.wav
OriginalTcStereo24-Asio2.wav

Certains parlent d'éviter le patchbay, je ne vois pas comment faire avec la TC ou l'EMU.

didier.brest 22nd November 2011 07:34 PM

I will ckeck these new files on next week-end.

david1103 24th November 2011 07:33 PM

1 Attachment(s)
hi didier,

i had big problems getting this sorted. diffmaker was crashing all the time on my win7 x64 system. (hope they update it)

it would not record, so i played and recorded in reaper. i assume this should work just the same? all bitrates etc are the same, faders all at 0.

the result for my echo audiofire 4 seems too bad!!! anyone else got one to test?

>> Parameters: -1.509sec, 0.513dB (L), 0.412dB (R)..Corr Depth: 25.6 dB (L), 27.4 dB (R)

i played out of output 3_4 into input 3_4. the files seem ok on listening.

can it be i made a mistake? could you have a check of the files?

i also tried output from my benchmark DAC1 phono outs into the audiofire 4 input 3_4. diffmaker won't have this file, it just crashes!

here is a link to the 2 .wav files 'diffmaker audiofire 4 test.zip'

diffmaker audiofire 4 test.zip

i feel i may have wasted a day on this, and don't want to do any more, but if you think these file are valid please add them to the list. hope it was of some help! this is a GREAT thread to see the quality of convertors, hope it keeps growing. can the audiofire4 be this bad?

_ben 24th November 2011 10:49 PM

Parameters: -4,012sec, 0,510dB (L), 0,412dB (R)..Corr Depth: 25,6 dB (L), 27,4 dB (R)

didier.brest 25th November 2011 12:44 AM

Quote:

Originally Posted by _ben (Post 7255731)

Merci! heppy

Quote:

Originally Posted by david1103 (Post 7263872)
Corr Depth: 25.6 dB (L), 27.4 dB (R)
(...)
can the audiofire4 be this bad?

Not so bad. Equivalent within 1 dB to audiofire 8.

Results list updated with the TC Impact twin stereo loop from ben instead of its mono loop. I will check the other files from ben and the ones from david1103 later on.


ADDA converters

Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (ben)
Corr Depth: 29,0 dB (L), 30,7 dB (R) Difference: -48,6 dBFS (L), -49,8 dBFS (R)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

E-MU 1616m (left channel only, ben)
Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

didier.brest 26th November 2011 02:52 PM

Correlated depth figures from ben's Impact Twin loops

1: 29,0 dB (L), 30,7 dB (R)
2: 29,1 dB (L), 30,8 dB (R)
3: 29,1 dB (L), 30,8 dB (R)
swap: 29,1 dB (L), 30,8 dB (R)
Asio1: 29,1 dB (L), 30,8 dB (R)
Asio2: 29,1 dB (L), 30,8 dB (R)

Results list updated with the TC Impact twin stereo loop Asio2 from ben instead of his loop 1. Correction of the Babyface-Fireface400 loop correlation depth figures that were mistakingly copied-pasted from the Duet2-Babyface loop result.
I will check the files from david1103 later on.

DA-AD loops from this wav file evaluated by means of Audio DiffMaker.


ADDA converters

Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (ben)
Corr Depth: 29,1 dB (L), 30,8 dB (R) Difference: -48,7 dBFS (L), -49,7 dBFS (R)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

E-MU 1616m (left channel only, ben)
Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 16,6 dB (L), 18,4 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)

didier.brest 27th November 2011 12:28 AM

Quote:

Originally Posted by david1103 (Post 7263872)
i had big problems getting this sorted. diffmaker was crashing all the time on my win7 x64 system. (hope they update it)

it would not record, so i played and recorded in reaper. i assume this should work just the same?

I did never use Audio DiffMaker for recording a loop. I am using it only for evaluating loops. Loop recording can be dona in any DAW software. So, yes Reaper is fine! (I'm using Reaper also.kfhkh)

Quote:

Originally Posted by david1103 (Post 7263872)
i also tried output from my benchmark DAC1 phono outs into the audiofire 4 input 3_4. diffmaker won't have this file, it just crashes!

The left and right channels were swapped. But even after correction, the correlated depth is only 10 dB. Something should come wrong because Benchmark DAC1 performing much lower than any DAC tested here up to now is not expected.

Results list updated with Audiofire loop from david1103.

DA-AD loops from this wav file evaluated by means of Audio DiffMaker.


ADDA converters

Mytek 8 x 192 (MainTime)
Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -54.2 dBFS (L) -55.1 dBFS (R)

TC Electronics Impact Twin (AndG)
Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

Prism Orpheus (MainTime)
Corr Depth: 30,8 dB (L), 32,4 dB (R) Difference: -51.1 dBFS (L) -52.4 dBFS (R)

RME Fireface 400 (didier.brest)
Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (ben)
Corr Depth: 29,1 dB (L), 30,8 dB (R) Difference: -48,7 dBFS (L), -49,7 dBFS (R)

Echo Audiofire 12 (guze)
Corr Depth: 28,8 dB (L), 30,1 dB (R) Difference: -48.4 dBFS (L) -49.2 dBFS (R)

E-MU 1616m (left channel only, ben)
Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Echo Audiofire 4 (david1103)
Corr Depth: 25,6 dB (L), 27,4 dB (R) Difference: -45,2 dBFS (L), -46,3 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Edirol FA-66 (didier.brest)
Corr Depth: 15,6 dB (L), 17,1 dB (R) Difference: -34.4 dBFS (L) -35.4 dBFS (R)

RME Babyface (didier.brest)
Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)


Sets of different converters

Aurora8 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 34,6 dB (L), 36,3 dB (R) Difference: -55.3 dBFS (L) -56.5 dBFS (R)

Echo Audiofire 8 with ADAT clocked by RME HDSP 9632 (schtim)
Corr Depth: 25.9 dB (L), 27.5 dB (R) Difference: -45.5 dBFS (L) -46.5 dBFS (R)

M-Audio Profire2626 clocked by RME HDSPe RayDAT (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.1 dBFS (L) -45.0 dBFS (R)

M-Audio Profire2626 clocked by Lynx Aurora8 (cylens)
Corr Depth: 24,4 dB (L), 25,9 dB (R) Difference: -44.0 dBFS (L) -44.9 dBFS (R)

RME Babyface --> RME Fireface 400 (didier.brest)
Corr Depth: 16,6 dB (L), 18,4 dB (R) Difference -29,4 dBFS (L), -28,8 dBFS (R)

Apogee Duet 2 --> RME Babyface (isma)
Corr Depth: 13,6 dB (L), 14,8 dB (R) Difference: -20.8 dBFS (L) -19.4 dBFS (R)