The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Evaluating AD/DA loops by means of Audio Diffmaker
Old 10th December 2013
  #361
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by bolide View Post
In theory a 1-bit conversion might score exceptionally well but the audio quality would suffer a tad

A 1-bit conversion would score very low because it would result in 1-bit file the difference of which with respect to the 24-bit original would be very large, providing you would have first defined how to transform this 1-bit file in a 24-bit file for computing its difference with respect to the 24 bit original
(for instance 0 would be transformed in -FS/2 and 1 in +FS/2 if the 24 bit format represents 2 x FS integer values from -FS to FS-1, FS = 2 at the power of 23).
Old 10th December 2013
  #362
Lives for gear
 
BrianVengeance's Avatar
 

Quote:
Originally Posted by swartzfeger View Post
Maybe a wrong assumption, but I'm assuming you're talking about the pres. My main concern is the quality of the AD/DA. I've never read much about the MOTU's converters and most reviews seem to gloss over this... and after going to MOTU's site, they don't post actual specs like Apogee, Focusrite etc. THD+N, A-weighted dyn range, etc.
Yes and no... the MOTU preamps are notoriously lamented for those reasons, but same can be said for the converters.

Without getting too far into the rabbit hole about the subjectiveness of quality, there isn't anything terribly special about MOTU conversion. 2408s score well due to solid clocks, good converter chips, helpful filters and non-distorting signal paths both in and out... to mix a metaphor, the photocopy of the photocopy looks (sounds) very much like the first photocopy that was made.

If this is what you are looking for in a converter, then "nothing special" equates to "bang on the mark".

On the other hand, Burl straps transformers across their inputs for the express purpose of adding (pleasant) signal distortion and in all likelihood, this same test would produce a larger delta / diff for the Mothership than the 2408, Hilo or other similarly high-ranked converter... yet there are few complaints about Burl conversion quality.

I agree that there's not a lot out there on MOTU conversion quality, but I stand by my original statement of it being clean, workhorse stuff. You aren't going to get gobs of mojo goo smeared on your signals using MOTU, but they will get you into the range of converters that step enough out of the way for you to worry about other problems.
Old 10th December 2013
  #363
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by 6strings View Post
I don't think you really understand this thread. The better the score, the clearer and more transparent the sound. i.e.: what you put in is what you get out. If the specs were as bad as folks here make them out to be, then they wouldn't score well.
I hate to spoil the party, but these AD/DA loop tests say nothing about the quality of the convertors in question. Note that all of them show terribly bad difference measurements anyway, as you'd expect near perfect cancellation (at least +90dB cancellation).

This is because any AD and DA convertor will change the phase-spectrum a lot (and very extreme actually, it goes all over the place). By itself, you can't hear this (we don't 'hear' phase as such on a signal by itself), but you will notice when summing/comparing back with the original like this test does.

The different measurements for the convertors merely show that the phase-spectrum response is different for each convertor, but again, this says nothing about quality.

I think relying on your ears to decide which convertor sounds better to you is the only relevant 'measurement'.
Old 11th December 2013
  #364
Lives for gear
 

Quote:
Originally Posted by DrDeltaM View Post
I think relying on your ears to decide which convertor sounds better to you is the only relevant 'measurement'.
I would ordinarily agree with this statement in all other cases, except that when it comes to converter quality. The cream of the crop should not reveal any audible difference between "looped" audio and the original sound source. If that is the case, how can our wildly subjective hearing honestly determine which unit is truly the most transparent sounding converter?

Unless, of course, converter "mojo" is purposely meant to be included as part of your overall workflow... To each their own, I guess.

My point I tried to make earlier, however, is that unless a converter is really, really poorly built, most users are better off making purchasing decisions on other aspects. ie: What good is the best converter in the world if the developer has a poor reputation for keeping up with his driver updates for your platform? Do you need quality preamps on board, or just line inputs? Etc., etc...
Old 11th December 2013
  #365
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by 6strings View Post
the best converters you should not reveal any audible difference between "looped" audio and the original sound source.
I actually agree with this, but I'm saying you should listen to the original and the AD-DA looped each by itself after each other. Summing them together (like Audio Diffmaker comparision) will only reveal lots of phaseshifting, but doesn't say anything about the quality.
So only your ears can tell you if you find there to be a signal deterioration in the AD-DA looped version or not.
Old 20th December 2013
  #366
Here for the gear
 

Here are a Few DAC-ADC combinations that I had in the shop.

https://www.mediafire.com/folder/qq8...io%20DiffMaker

They are primarily audiophile, not studio, gear but It might be interesting to compare them to the studio gear you have measured so far.

I used your original ~2 minute File for playback and as the music reference as requested.

In using Audio DiffMaker I have noticed that it seems to fudge up the beginning of the difference track for about 215ms.
I highly suggest you delete this tiny portion before you ask Wavelab to compute the RMS average or it throws off the result quite a bit.
It doesn't Make much difference for the lower resolution equipment but does with Higher rez equipment.

All these tests use a Windows 7 Computer with the [email protected] XTe soundcard used as digital I/O and or Analog I/O.
All Tests used Balanced Analog Interconnects.

#1
--
ReQuest Beast Soul DAC - MSB Studio ADC orig.wav - not processed by Audio DiffMaker
ReQuest Beast Soul DAC - MSB Studio ADC.wav - preprocessed Difference file
This file is the new and absurdly expensive ReQuest Beast Music Server DAC (designed by MSB Technology) connected to the equally expensive MSB Studio ADC.
This Test has the Beast Soul as the clock master and the MSB Studio ADC as the clock slave. (The Beast only has a clock output)

Parameters: -440.2msec, 4.485dB (L), 4.449dB (R)..Corr Depth: 63.1 dB (L), 65.0 dB (R)
Wavelab RMS with 215ms noise burst at the beginning -69.70 (L), -65.87 (R)
Wavelab RMS without 215ms noise burst at the beginning -73.72 (L), -72.66 (R)
--

#2
--
MSB Platinum IV Plus DAC - MSB Studio ADC orig.wav - not processed by Audio DiffMaker
MSB Platinum IV Plus DAC - MSB Studio ADC.wav - preprocessed Difference file
This file is the MSB Platinum IV plus DAC connected to the MSB Studio ADC.
This Test has the Platinum DAC plus as clock slave and the MSB Studio ADC as clock master.

Parameters: -590.3msec, 4.879dB (L), 4.853dB (R)..Corr Depth: 63.8 dB (L), 65.5 dB (R)
Wavelab RMS with 215ms noise burst at the beginning -69.55 (L), -66.35 (R)
Wavelab RMS without 215ms noise burst at the beginning -74.36 (L), -74.61 (R)
--

#3
--
Juli DAC - MSB Studio ADC orig.wav - not processed by Audio DiffMaker
Juli DAC - MSB Studio ADC.wav - preprocessed Difference file
This file is the [email protected] Soundcards Onboard DAC connected to the MSB Studio ADC.
This test has the [email protected] soundcard clock Slaved to the MSB Studio ADC.

Parameters: -426.5msec, 1.610dB (L), 1.579dB (R)..Corr Depth: 47.2 dB (L), 49.8 dB (R)
Wavelab RMS with 215ms noise burst at the beginning -61.61 (L), -61.95 (R)
Wavelab RMS without 215ms noise burst at the beginning -61.99 (L), -63.36 (R)
--

#4
--
Juli DAC - Juli ADC orig.wav - not processed by Audio DiffMaker
Juli DAC - Juli ADC.wav - preprocessed Difference file
This file is the [email protected] Soundcard Onboard DAC connected to the [email protected] Soundcard onboard ADC.
The [email protected] Soundcards internal clock was used.

Parameters: -535.2msec, -0.315dB (L), -0.350dB (R)..Corr Depth: 33.9 dB (L), 36.1 dB (R)
Wavelab RMS with 215ms noise burst at the beginning -48.98 (L), -49.96 (R)
Wavelab RMS without 215ms noise burst at the beginning -48.99 (L), -50.03 (R)
--

Last edited by DUJS; 20th December 2013 at 04:27 AM.. Reason: Forgot Original Files
Old 20th December 2013
  #367
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by DUJS View Post
In using Audio DiffMaker I have noticed that it seems to fudge up the beginning of the difference track for about 215ms.
I highly suggest you delete this tiny portion before you ask Wavelab to compute the RMS average or it throws off the result quite a bit.
It has been part of my evaluation method from the beginning of this test.

I have udpated the list with the results from your file got with the combination of The Beast DAC from ReQuest Audio and the MSB Platinum Studio ADC.

I will add the results from the three other loopback files later on.

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

ADDA converters


Lynx Hilo (CoolColJ)

-0,1 dB (L), -0,1 dB (R) Corr Depth: 41,8 dB (L), 43,7 dB (R), Difference: -56,8 dBFS (L), -57,2 dBFS (R)

MOTU 2408mk3 (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,9 dBFS (R)

MOTU 24I/O (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,8 dBFS (R)

MOTU 828mk2 (laurend)
6,1 dB (L), 6,2 dB (R) Corr Depth: 36,4 dB (L), 37,9 dB (R) Difference: -55.7 dBFS (L) -57.0 dBFS (R)

Mytek 8 x 192 (MainTime)
0,4 dB (L), 0,5 dB (R) Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
-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)

-0,432dB (L), 0,029dB (R) Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
0,5 dB (L), 0,5 dB (R) Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

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

MOTU Track16 (EvgenyStudio)
7,0 dB (L), 6,9 dB Corr Depth: 30,0 dB (L), 31,5 dB (R) Difference: -50,4 dBFS (L), -51,6 dBFS (R)

RME Fireface 400 (didier.brest)
0,3 dB (L), 0,3 dB (R) Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (Coucou)
Corr Depth: 29,1 dB (L), 30,4 dB (R) Difference: -48,9 dBFS (L), -49,2 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)
0,7 dB (L), 0,7 dB (R) 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)
-0,4 dB (L), Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Apogee Rosetta 800 (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Symphony I/O (Ajantis)
Corr Depth: 43,4 dB (L), 45,6 dB (R) Difference: -45,9 dBFS (L), -45,5 dBFS dB (R)

Echo Audiofire 4 (david1103)
0,5 dB (L), 0,4 dB Corr Depth: 25,6 dB (L), 27,4 dB (R) Difference: -45,2 dBFS (L), -46,3 dBFS (R)

Metric Halo ULN-2 (DownSideUp)
Corr Depth: 25,3 dB (L), 26,9 dB (R) Difference: -44,9 dB (L), 46,0 dB (R)

Universal Audio Apollo Quad (davidbayles)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Multiface II - Unbalanced, -10 dBV (juki, Audiofanzine)
Corr Depth: 22,7 dB (L), 24,4 dB (R) Difference: -42,4 dBFS (L) -43,5 dBFS (R)

RME Fireface 800 (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dB (L), -43,0 dB (R)

RME Multiface II - Balanced, 4 dBu (juki, Audiofanzine)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L) -43,0 dBFS (R)

MOTU Traveler MK1 (sinebox)
Corr Depth: 22,2 dB (L), 23,9 dB (R) Difference: -41,8 dBFS (L), -42,9 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
0,8 dB (L), 0,9 dB (R) Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
1,0 dB (L), 0,9 dB (R) Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Avid Mbox 3 (Coucou)
Corr Depth: 21,8 dB (L), 23,6 dB (R) Difference: -39,7 dBFS (L), -40,3 dBFS (R)

Focusrite Saffire Pro 24 DSP (CoolColJ), first 45s part disregarded
5,2 dB (L), 5,2 dB (R) Corr Depth: 20,2 dB (L), 22,0 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to TRS inputs 3-4 (didier.brest)
0,8 dB (L), 0,7 dB (R) Corr Depth: 18,5 dB (L), 20,1 dB (R) Difference: -38,9 dBFS (L), -40,3 dBFS (R)

Antelope Orion (davidbayle)
Corr Depth: 19,2 dB (L), 20,7 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, TRS outputs 5-6 to TRS inputs 3-4 (didier.brest)
0,7 dB (L), 0,7 dB (R) Corr Depth: 18,4 dB (L), 20,1 dB (R) Difference: -38,8 dBFS (L), -40,3 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
1,1 dB (L), 0,6 dB (R) Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Focusrite Forte (Coucou)
Corr Depth: 18,3 dB (L), 19,8 dB (R) Difference: -38,3 dBFS (L), -39,4 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to XLR preamp inputs 9-10 (didier.brest)
0,5 dB (L), 0,4 dB (R) Corr Depth: 14,7 dB (L), 16,2 dB (R) Difference -35,03 dBFS (L) -36,30 dBFS (R)

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

Focusrite Scarlett 2i2 (EvgenyStudio)
15,5 dB (L), 15,5 dB (R)Corr Depth: 19,3 dB (L), 23,6 dB (R) Difference: -31,8 dBFS (L), -31,1 dBFS (R)

RME Babyface (didier.brest)
0,6 dB (L), 0,7 dB (R) Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)



External clocking


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)

Apogee Rosetta 800 clocked by Universal Audio Apollo (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Rosetta 800 clocked by Antelope Orion (davidbayle)
Corr Depth: 39,6 dB (L), 41,6 dB (R) Difference: -45,9 dBFS (L), -45,8 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)

Universal Audio Apollo clocked by Apogee Rosetta (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

Universal Audio Apollo clocked by Antelope Orion (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Fireface 800 clocked by Yamaha 02R (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L), -43,0 dBFS (R)



Sets of different converters


Synchronised mode (DA master unless otherwise specified)

ReQuest Audio, The Beast DAC (master clock) ---> MSB Platinum Studio ADC (DUJS)
Corr Depth: 62,0 dB (L), 63,4 dB (R) Difference: -76,9 dBFS (L), -76,3 dBFS (R)

Atoll DAC200 --> RME Babyface (didier.brest)
Corr Depth: 10,1 dB (L), 11,9 dB (R) Difference -30,0 dBFS (L) -31.3 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)

Benchmark DAC1 phone output ---> Echo Audiofire 4 (david1103)
Corr Depth: 10,4 dB (L), 10,2 dB (R) Difference: -22,6 dBFS (L), -23,1 dBFS (R)

Unsynchronized mode. The DA and AD converters run each on its own clock. Measurement performed with sample rate drift correction activated in Audio DiffMaker and limited to the first half of the sample. Because the RMS level of this first half is 1 dB lower than the RMS level of the full-size sample, one must increase the difference RMS level figures by 1 dB for comparison with the synchronized mode.

ATOLL DAC 200 --> RME Fireface UFX (didier.brest)
2,3 dB (L), 2,4 dB (R). rate adj=8,8289 ppm.Corr Depth: 23,2 dB (L), 25,3 dB (R) Difference: -44,0 dBFS (L), -45,4 dBFS (R)

ATOLL DAC 200 --> RME Babyface (didier.brest)
Rate adj=-4,3193 ppm.Corr Depth: 10,8 dB (L), 13,0 dB (R) Difference: -31,5 dBFS (L), -33,1 dBFS

For contributing to this test :
- download the test 2 x 44 kHz x 24 bit wav file (31 Mb),
- cable link the 2 DAC outputs to 2 ADC inputs,
- play back the test file to the DAC outputs and record from the ADC inputs,
- upload to mediafire or any other file sharing site the 2 x 44 kHz x 24 bit wav file from this recording and
- provide the download link here with information about ADC and DAC identities.
Your contribution is welcome even if your converter is already in the above list.
Thank you!
Old 20th December 2013
  #368
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by didier.brest View Post
It has been part of my evaluation method from the beginning of this test.
Sorry to say that this test isn't really giving any useful information, please read my post concerning this:
https://www.gearslutz.com/board/9665979-post363.html
Old 20th December 2013
  #369
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by DrDeltaM View Post
Summing them together (like Audio Diffmaker comparision) will only reveal lots of phaseshifting, but doesn't say anything about the quality.
AudiodiffMaker compensates for any time delay between the original and the copy. Such a delay is a phase shift varying linearly as a function of the frequency. Any other phase shift function would result in transients smearing that would be possibly audible. This is why high end gear (microphone, preamp, converter) manufacturers provide specifications for the tranfer function phase of their products.

It is not only by chance that in general the more sophisticated (hence expensive) converters have got the best results in this test.
Old 20th December 2013
  #370
Lives for gear
 
fossaree's Avatar
Quote:
Originally Posted by DrDeltaM View Post
Sorry to say that this test isn't really giving any useful information, please read my post concerning this:
https://www.gearslutz.com/board/9665979-post363.html
Quote:
Originally Posted by didier.brest View Post
AudiodiffMaker compensates for any time delay between the original and the copy. Such a delay is a phase shift varying linearly as a function of the frequency. Any other phase shift function would result in transients smearing that would be possibly audible. This is why high end gear (microphone, preamp, converter) manufacturers provide specifications for the tranfer function phase of their products.

It is not only by chance that in general the more sophisticated (hence expensive) converters have got the best results in this test.
I see truth in both of you - can you get to agree on something ?
Old 20th December 2013
  #371
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by didier.brest View Post
AudiodiffMaker compensates for any time delay between the original and the copy. Such a delay is a phase shift varying linearly as a function of the frequency.
Correct, but this phase shift is very different per frequency, and not a linear function of the time delay. So what AudiodiffMaker does is compensating the average intersample delay that occurs due to the extreme phase shifts over the whole spectrum. It does not correct the phase per frequency.

Quote:
Originally Posted by didier.brest View Post
Any other phase shift function would result in transients smearing that would be possibly audible.
You'd think so, but that doesn't seem to be the case. Humans can't hear phase very well by itself. And of course, it will depend on how exactly the phase shift is, on whether or not transients sound smeared. A phase-cancel test like done by AudiodiffMaker will not put a quality-value on that however.

Quote:
Originally Posted by didier.brest View Post
It is not only by chance that in general the more sophisticated (hence expensive) converters have got the best results in this test.
I fear the cancellation test result are actually quite random, and represent no 'quality-value' at all.
Please have a look at the phase-response of an AD-DA loop of a convertor, in attachment to this post. In testing several convertors this way, similar graphs were always seen. There were some differences in those phase-reponse figures, where the freqs were of the different phase zero-crossings, but all showed phase going all over the place in a similar way as the graph below.

And that's exactly why you can't phase-cancel the original with the convertor-looped version very well, and why this phase-cancel test it says nothing about the quality of the convertor.
Attached Files
File Type: pdf A_D D_A Phase.pdf (49.5 KB, 232 views)
Old 20th December 2013
  #372
Lives for gear
 
fossaree's Avatar
thanks !
Old 20th December 2013
  #373
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by DrDeltaM View Post
Correct, but this phase shift is very different per frequency, and not a linear function of the time delay. So what AudiodiffMaker does is compensating the average intersample delay that occurs due to the extreme phase shifts over the whole spectrum. It does not correct the phase per frequency.
AudioDiffmaker compensates for a delay which by nature is a phase shift varying linearly vs. frequency according to the relation:

unwrapped phase in radian = 2 x pi x frequency in Hz x time delay in second.

We agree that it does not compensate for any deviation with respect to this ideal law. We disagree whether such a deviation is audible or not. My point is that it may result in transient smearing, which is a well known effect of non linear phase filtering.


Quote:
Originally Posted by DrDeltaM View Post
Please have a look at the phase-response of an AD-DA loop of a convertor, in attachment to this post. In testing several convertors this way, similar graphs were always seen. There were some differences in those phase-reponse figures, where the freqs were of the different phase zero-crossings, but all showed phase going all over the place in a similar way as the graph below.
This graph shows the wrapped phase modulo 2 x pi (360°). For making appear the deviation with respect to a linear phase function, one has to unwrapp this phase curve and to present it with a linear frequency scale rather than a logarithmic scale.
Then one would observe that this curve is very close to a sloping-up (is this correct English ?) straight line.

Let us do that on some points of your curve:

f = 20 Hz ---> phase ~ 15°
f = 100 Hz ---> phase ~ 70°
f = 1000 Hz ---> phase ~ 180° + 360° + 180° = 720° = 2 x pi rd
f = 2000 Hz ---> phase ~ 720° + 720° = 1440° = 4 x pi rd

Looks much like a pure delay on the order of 2 ms. For a time delay exactly equal to 2 ms, the phase shift to frequency relationship is:
f = 20 Hz ---> phase = 14.4 °
f = 100 Hz ---> phase = 72°
f = 1000 Hz ---> phase = 720°
f = 2000 Hz ---> phase = 1440°



Anyway we agree that the best way for evaluating these converters is listening to the wav files, which is what I already wrote in this thread.
Old 20th December 2013
  #374
Lives for gear
 
fossaree's Avatar
Quote:
Originally Posted by DrDeltaM View Post
Correct, but this phase shift is very different per frequency, and not a linear function of the time delay. So what AudiodiffMaker does is compensating the average intersample delay that occurs due to the extreme phase shifts over the whole spectrum. It does not correct the phase per frequency.



You'd think so, but that doesn't seem to be the case. Humans can't hear phase very well by itself. And of course, it will depend on how exactly the phase shift is, on whether or not transients sound smeared. A phase-cancel test like done by AudiodiffMaker will not put a quality-value on that however.



I fear the cancellation test result are actually quite random, and represent no 'quality-value' at all.
Please have a look at the phase-response of an AD-DA loop of a convertor, in attachment to this post. In testing several convertors this way, similar graphs were always seen. There were some differences in those phase-reponse figures, where the freqs were of the different phase zero-crossings, but all showed phase going all over the place in a similar way as the graph below.

And that's exactly why you can't phase-cancel the original with the convertor-looped version very well, and why this phase-cancel test it says nothing about the quality of the convertor.
Quote:
Originally Posted by didier.brest View Post
AudioDiffmaker compensates for a delay which by nature is a phase shift varying linearly vs. frequency according to the relation:

unwrapped phase in radian = 2 x pi x frequency in Hz x time delay in second.

We agree that it does not compensate for any deviation with respect to this ideal law. We disagree whether such a deviation is audible or not. My point is that it may result in transient smearing, which is a well known effect of non linear phase filtering.




This graph shows the wrapped phase modulo 2 x pi (360°). For making appear the deviation with respect to a linear phase function, one has to unwrapp this phase curve and to present it with a linear frequency scale rather than a logarithmic scale.
Then one would observe that this curve is very close to a sloping-up (is this correct english ?) straight line.

Let us do that on some points of your curve:

f = 20 Hz ---> phase ~ 15°
f = 100 Hz ---> phase ~ 70°
f = 1000 Hz ---> phase ~ 180° + 360° + 180° = 720° = 2 x pi rd
f = 2000 Hz ---> phase ~ 720° + 720° = 1440° = 4 x pi rd

Looks much like a pure delay on the order of 2 ms. For a time delay exactly equal to 2 ms, the phase shift to frequency relationship is:
f = 20 Hz ---> 14.4 °
f = 100 Hz ---> 72°
f = 1000 Hz ---> 720°
f = 2000 Hz ---> 1440°



Anyway we agree that the best way for evaluating these converters is listening to the wav files, which is what I already wrote in this thread.
What they both say ! Thanks !
Old 20th December 2013
  #375
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by didier.brest View Post
AudioDiffmaker compensates for a delay which by nature is a phase shift varying linearly vs. frequency according to the relation:

unwrapped phase in radian = 2 x pi x frequency in Hz x time delay in second.

We agree that it does not compensate for any deviation with respect to this ideal law.
Indeed, and this deviation is defenately there for any convertor (hence the very bad phase-cancelling values). The phase-cancel test AudiodiffMaker does is however not a value of how good or how bad this devation is for the actual sound, the value has no direct relation to the quality of this convertor.

Quote:
Originally Posted by didier.brest View Post
Anyway we agree that the best way for evaluating these converters is listening to the wav files, which is what I already wrote in this thread.
I think that's indeed the most important conclusion
Old 20th December 2013
  #376
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by DrDeltaM View Post
Indeed, and this deviation is defenately there for any convertor (hence the very bad phase-cancelling values).



There is undoubtedly a phase deviation with respect to the ideal linear phase vs. frequency law. It is very low as shown by your own graph.

It is a distorsion of the audio signal as produced by a non linear phase filtering (the reason why linear phase filtering is preferred by some people especially for mastering).

Hence if it impacts significantly the results from AudioDiffMaker, which is your unproved allegation (which I doubt about because it is so low and generally peaks at the very low end where the energy is small), it is quite legitimate.
Old 20th December 2013
  #377
Lives for gear
 
DrDeltaM's Avatar
 

Quote:
Originally Posted by didier.brest View Post
It is a distorsion of the audio signal as produced by a non linear phase filtering (the reason why linear phase filtering is preferred by some people especially for mastering).
A distortion by non linear phase filtering = a phase shift

Quote:
Originally Posted by didier.brest View Post
Hence if it impacts significantly the results from AudioDiffMaker, which is your unproved allegation (which I doubt about because it is so low and generally peaks at the very low end where the energy is small), it is quite legitimate.
I wonder how else you can explain cancellation of only up to 40-50dB. That's FAR away from the +90dB cancellation one would expect.
Old 21st December 2013
  #378
Lives for gear
 
didier.brest's Avatar
 

List updated with the results from the 4 looback file from DUJS

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

ADDA converters


Lynx Hilo (CoolColJ)

-0,1 dB (L), -0,1 dB (R) Corr Depth: 41,8 dB (L), 43,7 dB (R), Difference: -56,8 dBFS (L), -57,2 dBFS (R)

MOTU 2408mk3 (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,9 dBFS (R)

MOTU 24I/O (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,8 dBFS (R)

MOTU 828mk2 (laurend)
6,1 dB (L), 6,2 dB (R) Corr Depth: 36,4 dB (L), 37,9 dB (R) Difference: -55.7 dBFS (L) -57.0 dBFS (R)

Mytek 8 x 192 (MainTime)
0,4 dB (L), 0,5 dB (R) Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
-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)

ESI [email protected] XTe (DUJS)
Corr Depth: 32,7 dB (L), 34,3 dB (R) Difference: -52,3 dBFS (L), -53,5 dBFS (R)

TC Electronics Impact Twin (AndG)

-0,432dB (L), 0,029dB (R) Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
0,5 dB (L), 0,5 dB (R) Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

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

MOTU Track16 (EvgenyStudio)
7,0 dB (L), 6,9 dB Corr Depth: 30,0 dB (L), 31,5 dB (R) Difference: -50,4 dBFS (L), -51,6 dBFS (R)

RME Fireface 400 (didier.brest)
0,3 dB (L), 0,3 dB (R) Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (Coucou)
Corr Depth: 29,1 dB (L), 30,4 dB (R) Difference: -48,9 dBFS (L), -49,2 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)
0,7 dB (L), 0,7 dB (R) 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)
-0,4 dB (L), Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Apogee Rosetta 800 (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Symphony I/O (Ajantis)
0,0 dB (L), 0,0 dB (R) Corr Depth: 43,4 dB (L), 45,6 dB (R) Difference: -45,9 dBFS (L), -45,5 dBFS dB (R)

Echo Audiofire 4 (david1103)
0,5 dB (L), 0,4 dB Corr Depth: 25,6 dB (L), 27,4 dB (R) Difference: -45,2 dBFS (L), -46,3 dBFS (R)

Metric Halo ULN-2 (DownSideUp)
Corr Depth: 25,3 dB (L), 26,9 dB (R) Difference: -44,9 dB (L), 46,0 dB (R)

Universal Audio Apollo Quad (davidbayles)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Multiface II - Unbalanced, -10 dBV (juki, Audiofanzine)
Corr Depth: 22,7 dB (L), 24,4 dB (R) Difference: -42,4 dBFS (L) -43,5 dBFS (R)

RME Fireface 800 (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dB (L), -43,0 dB (R)

RME Multiface II - Balanced, 4 dBu (juki, Audiofanzine)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L) -43,0 dBFS (R)

MOTU Traveler MK1 (sinebox)
Corr Depth: 22,2 dB (L), 23,9 dB (R) Difference: -41,8 dBFS (L), -42,9 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
0,8 dB (L), 0,9 dB (R) Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
1,0 dB (L), 0,9 dB (R) Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Avid Mbox 3 (Coucou)
Corr Depth: 21,8 dB (L), 23,6 dB (R) Difference: -39,7 dBFS (L), -40,3 dBFS (R)

Focusrite Saffire Pro 24 DSP (CoolColJ), first 45s part disregarded
5,2 dB (L), 5,2 dB (R) Corr Depth: 20,2 dB (L), 22,0 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to TRS inputs 3-4 (didier.brest)
0,8 dB (L), 0,7 dB (R) Corr Depth: 18,5 dB (L), 20,1 dB (R) Difference: -38,9 dBFS (L), -40,3 dBFS (R)

Antelope Orion (davidbayle)
Corr Depth: 19,2 dB (L), 20,7 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, TRS outputs 5-6 to TRS inputs 3-4 (didier.brest)
0,7 dB (L), 0,7 dB (R) Corr Depth: 18,4 dB (L), 20,1 dB (R) Difference: -38,8 dBFS (L), -40,3 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
1,1 dB (L), 0,6 dB (R) Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Focusrite Forte (Coucou)
Corr Depth: 18,3 dB (L), 19,8 dB (R) Difference: -38,3 dBFS (L), -39,4 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to XLR preamp inputs 9-10 (didier.brest)
0,5 dB (L), 0,4 dB (R) Corr Depth: 14,7 dB (L), 16,2 dB (R) Difference -35,03 dBFS (L) -36,30 dBFS (R)

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

Focusrite Scarlett 2i2 (EvgenyStudio)
15,5 dB (L), 15,5 dB (R)Corr Depth: 19,3 dB (L), 23,6 dB (R) Difference: -31,8 dBFS (L), -31,1 dBFS (R)

RME Babyface (didier.brest)
0,6 dB (L), 0,7 dB (R) Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)



External clocking


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)

Apogee Rosetta 800 clocked by Universal Audio Apollo (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Rosetta 800 clocked by Antelope Orion (davidbayle)
Corr Depth: 39,6 dB (L), 41,6 dB (R) Difference: -45,9 dBFS (L), -45,8 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)

Universal Audio Apollo clocked by Apogee Rosetta (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

Universal Audio Apollo clocked by Antelope Orion (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Fireface 800 clocked by Yamaha 02R (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L), -43,0 dBFS (R)



Sets of different converters


Synchronised mode (DA master unless otherwise specified)

MSB Platinum IV Plus DAC ---> MSB Studio ADC, master clock (DUJS)
Corr Depth: 62,8 dB (L), 63,4 dB (R) Difference: -77,7 dBFS (L), -77,4 dBFS (R)

ReQuest Audio, The Beast DAC ---> MSB Platinum Studio ADC (DUJS)
Corr Depth: 62,0 dB (L), 63,4 dB (R) Difference: -76,9 dBFS (L), -76,3 dBFS (R)

[email protected] XTe DAC ---> MSB Studio ADC, master clock (DUJS)
Corr Depth: 46,0 dB (L), 48,0 dB (R) Difference: -65,3 dBFS (L), -66,7 dBFS (R)

Atoll DAC200 --> RME Babyface (didier.brest)
Corr Depth: 10,1 dB (L), 11,9 dB (R) Difference -30,0 dBFS (L) -31.3 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)

Benchmark DAC1 phone output ---> Echo Audiofire 4 (david1103)
Corr Depth: 10,4 dB (L), 10,2 dB (R) Difference: -22,6 dBFS (L), -23,1 dBFS (R)

Unsynchronized mode. The DA and AD converters run each on its own clock. Measurement performed with sample rate drift correction activated in Audio DiffMaker and limited to the first half of the sample. Because the RMS level of this first half is 1 dB lower than the RMS level of the full-size sample, one must increase the difference RMS level figures by 1 dB for comparison with the synchronized mode.

ATOLL DAC 200 --> RME Fireface UFX (didier.brest)
2,3 dB (L), 2,4 dB (R). rate adj=8,8289 ppm.Corr Depth: 23,2 dB (L), 25,3 dB (R) Difference: -44,0 dBFS (L), -45,4 dBFS (R)

ATOLL DAC 200 --> RME Babyface (didier.brest)
Rate adj=-4,3193 ppm.Corr Depth: 10,8 dB (L), 13,0 dB (R) Difference: -31,5 dBFS (L), -33,1 dBFS

For contributing to this test :
- download the test 2 x 44 kHz x 24 bit wav file (31 Mb),
- cable link the 2 DAC outputs to 2 ADC inputs,
- play back the test file to the DAC outputs and record from the ADC inputs,
- upload to mediafire or any other file sharing site the 2 x 44 kHz x 24 bit wav file from this recording and
- provide the download link here with information about ADC and DAC identities.
Your contribution is welcome even if your converter is already in the above list.
Thank you!
Old 21st December 2013
  #379
Lives for gear
 
fossaree's Avatar
Coincidentally or not , the most expensive of the whole list presents the best figures . (MSB Platinum IV Plus DAC)
Not saying it's better nor worse . Just saying .
Old 23rd December 2013
  #380
Lives for gear
 
didier.brest's Avatar
 

The list includes now the results from the 4 files provided by DUJS.
Old 27th December 2013
  #381
Lives for gear
 

Quote:
Originally Posted by DrDeltaM View Post
I actually agree with this, but I'm saying you should listen to the original and the AD-DA looped each by itself after each other. Summing them together (like Audio Diffmaker comparision) will only reveal lots of phaseshifting, but doesn't say anything about the quality.
So only your ears can tell you if you find there to be a signal deterioration in the AD-DA looped version or not.
Good luck telling them that. I have been trying to do that since... 6 months ago and this pointless practice is still going on.

I don't think ears are a good evaluator here either (unless you have lots of music to listen through the converters). In my opinion we need to test the actual claims of the manufacturers. The THD, aliasing and frequency/phase response. From there you could draw more subjective conclusions also.


Quote:
AudioDiffmaker compensates for a delay which by nature is a phase shift varying linearly vs. frequency according to the relation:

unwrapped phase in radian = 2 x pi x frequency in Hz x time delay in second.

We agree that it does not compensate for any deviation with respect to this ideal law. We disagree whether such a deviation is audible or not. My point is that it may result in transient smearing, which is a well known effect of non linear phase filtering.
Delay compensation is mostly useless when trying to eliminate the nonlinear phase shift. Furthermore, I would expct the converter with the phase differences to sound better, because the phase difference could mean greatly reduced aliasing and other artifacts. The "phase smearing" is not very audible compared to how much it affects the results of the null test, or perhaps the latency of the unit (since "perfect" filters introduce more latency).


Quote:

Let us do that on some points of your curve:

f = 20 Hz ---> phase ~ 15°
f = 100 Hz ---> phase ~ 70°
f = 1000 Hz ---> phase ~ 180° + 360° + 180° = 720° = 2 x pi rd
f = 2000 Hz ---> phase ~ 720° + 720° = 1440° = 4 x pi rd

Looks much like a pure delay on the order of 2 ms. For a time delay exactly equal to 2 ms, the phase shift to frequency relationship is:
f = 20 Hz ---> phase = 14.4 °
f = 100 Hz ---> phase = 72°
f = 1000 Hz ---> phase = 720°
f = 2000 Hz ---> phase = 1440°
This is a joke, right? Anyway, would be nice to see a picture where the delay is compensated for (of course, I believe stickyfingers posted those here earlier), but no converter is going to ship without AA/reconstruction filters and DC offset filters as well. Unless you prefer a converter that blows your speakers...
Old 30th December 2013
  #382
Gear Addict
 
lostinmusic's Avatar
 

As a highly subjective aside.... it's my opinion that MOTU have made some of the worst and also the best converters I've used.

The 1st gen 2408 especially was a bit small and muddy sounding and the early 828 and recent 192 were only ok to my ears. The 1224 was great though and I still use one and I've heard the PCI/mk3 2408 and can't find fault with it (for conversion only).

My point is.... I can't help wondering if people here have had their perfectly valid, bad experiences with a MOTU converter and react when reading how well another MOTU model performs, not aware that the range varies so much and not always in relation to cost?
Old 20th January 2014
  #383
Lives for gear
 

Is this still an active project? I could test a Focusrite Liquid Saffire 56
Old 20th January 2014
  #384
Lives for gear
 
didier.brest's Avatar
 

Yes it is and your test is welcome !
Old 22nd January 2014
  #385
Lives for gear
 

didier.brest, I have PMed you a link to my test. I used Focusrite Liquid Saffire 56. Unfortunately I did not have balanced 1/4" TRS to 1/4" TRS to use the Line In jacks to the ADC, so I had to use TRS to XLR and go through the mic pres. I had to use the PAD to avoid distortion, but amazingly the audio did not pass -0.1dBFS on the AD side. Regardless, the Line Ins on this interface go through the preamp anyway, so it should not be much if any difference. I also shifted the recorded file back in time to line up with the original wav to account for the converter delay before I exported. Hopefully this test is ok. Thanks!
Old 22nd January 2014
  #386
Lives for gear
 
didier.brest's Avatar
 

Thank you much for your effort !
The result from your Liquid Saffire test is quite close to the the one from the Focusrite Forte. Lists of results updated herebelow.

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

ADDA converters


Lynx Hilo (CoolColJ)

-0,1 dB (L), -0,1 dB (R) Corr Depth: 41,8 dB (L), 43,7 dB (R), Difference: -56,8 dBFS (L), -57,2 dBFS (R)

MOTU 2408mk3 (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,9 dBFS (R)

MOTU 24I/O (ram75)
0,1 dB (L), 0,1 dB (R) Corr Depth: 36,3 dB (L), 37,9 dB (R) Difference: -55,80 dBFS (L), -56,8 dBFS (R)

MOTU 828mk2 (laurend)
6,1 dB (L), 6,2 dB (R) Corr Depth: 36,4 dB (L), 37,9 dB (R) Difference: -55.7 dBFS (L) -57.0 dBFS (R)

Mytek 8 x 192 (MainTime)
0,4 dB (L), 0,5 dB (R) Corr Depth: 35,3 dB (L), 36,9 dB (R) Difference: -55.2 dBFS (L) -56.0 dBFS (R)

Aurora 8 (cylens)
-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)

ESI [email protected] XTe (DUJS)
Corr Depth: 32,7 dB (L), 34,3 dB (R) Difference: -52,3 dBFS (L), -53,5 dBFS (R)

TC Electronics Impact Twin (AndG)

-0,432dB (L), 0,029dB (R) Corr Depth: 32,0 dB (L), 33,6 dB (R) Difference: -51.7 dBFS (L) -52.7 dBFS (R)

RME HDSP 9632 (schtim)
0,5 dB (L), 0,5 dB (R) Corr Depth: 31.4 dB (L), 33.0 dB (R) Difference: -51.2 dBFS (L) -52.1 dBFS (R)

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

MOTU Track16 (EvgenyStudio)
7,0 dB (L), 6,9 dB Corr Depth: 30,0 dB (L), 31,5 dB (R) Difference: -50,4 dBFS (L), -51,6 dBFS (R)

RME Fireface 400 (didier.brest)
0,3 dB (L), 0,3 dB (R) Corr Depth: 30,2 dB (L), 31,8 dB (R) Difference: -49.8 dBFS (L) -50.8 dBFS (R)

TC Electronics Impact Twin (Coucou)
Corr Depth: 29,1 dB (L), 30,4 dB (R) Difference: -48,9 dBFS (L), -49,2 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)
0,7 dB (L), 0,7 dB (R) 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)
-0,4 dB (L), Corr Depth: 28,9 dB (L) Difference: -45,9 dBFS (L)

Apogee Rosetta 800 (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Symphony I/O (Ajantis)
0,0 dB (L), 0,0 dB (R) Corr Depth: 43,4 dB (L), 45,6 dB (R) Difference: -45,9 dBFS (L), -45,5 dBFS dB (R)

Echo Audiofire 4 (david1103)
0,5 dB (L), 0,4 dB Corr Depth: 25,6 dB (L), 27,4 dB (R) Difference: -45,2 dBFS (L), -46,3 dBFS (R)

Metric Halo ULN-2 (DownSideUp)
Corr Depth: 25,3 dB (L), 26,9 dB (R) Difference: -44,9 dB (L), 46,0 dB (R)

Universal Audio Apollo Quad (davidbayles)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Multiface II - Unbalanced, -10 dBV (juki, Audiofanzine)
Corr Depth: 22,7 dB (L), 24,4 dB (R) Difference: -42,4 dBFS (L) -43,5 dBFS (R)

RME Fireface 800 (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dB (L), -43,0 dB (R)

RME Multiface II - Balanced, 4 dBu (juki, Audiofanzine)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L) -43,0 dBFS (R)

MOTU Traveler MK1 (sinebox)
Corr Depth: 22,2 dB (L), 23,9 dB (R) Difference: -41,8 dBFS (L), -42,9 dBFS (R)

Echo Audiofire 8 non adat CS4272 (guze)
0,8 dB (L), 0,9 dB (R) Corr Depth: 26,4 dB (L), 28,2 dB (R) Difference: -40.6 dBFS (L) -40.6 dBFS (R)

Apogee Duet2 (isma)
1,0 dB (L), 0,9 dB (R) Corr Depth: 19,8 dB (L), 21,4 dB (R) Difference: -40.2 dBFS (L) -41.6 dBFS (R)

Avid Mbox 3 (Coucou)
Corr Depth: 21,8 dB (L), 23,6 dB (R) Difference: -39,7 dBFS (L), -40,3 dBFS (R)

Focusrite Saffire Pro 24 DSP (CoolColJ), first 45s part disregarded
5,2 dB (L), 5,2 dB (R) Corr Depth: 20,2 dB (L), 22,0 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to TRS inputs 3-4 (didier.brest)
0,8 dB (L), 0,7 dB (R) Corr Depth: 18,5 dB (L), 20,1 dB (R) Difference: -38,9 dBFS (L), -40,3 dBFS (R)

Antelope Orion (davidbayle)
Corr Depth: 19,2 dB (L), 20,7 dB (R) Difference: -38,9 dBFS (L), -39,9 dBFS (R)

RME Fireface UFX, TRS outputs 5-6 to TRS inputs 3-4 (didier.brest)
0,7 dB (L), 0,7 dB (R) Corr Depth: 18,4 dB (L), 20,1 dB (R) Difference: -38,8 dBFS (L), -40,3 dBFS (R)

Yamaha Steinberg MR816X (didier.brest)
1,1 dB (L), 0,6 dB (R) Corr Depth: 19,1 dB (L), 21,0 dB (R) Difference: -38.6 dBFS (L) -40.0 dBFS (R)

Focusrite Forte (Coucou)
Corr Depth: 18,3 dB (L), 19,8 dB (R) Difference: -38,3 dBFS (L), -39,4 dBFS (R)

Focusrite Liquid Saffire 56, Line output to microphone input + pad (Melodeath)
0,2 dB (L), 0,2 dB (R) Corr Depth: 17,9 dB (L), 19,3 dB (R) Difference: -37,9 dBFS (L) -38,9 dBFS (R)

RME Fireface UFX, XLR outputs 1-2 to XLR preamp inputs 9-10 (didier.brest)
0,5 dB (L), 0,4 dB (R) Corr Depth: 14,7 dB (L), 16,2 dB (R) Difference -35,03 dBFS (L) -36,30 dBFS (R)

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

Focusrite Scarlett 2i2 (EvgenyStudio)
15,5 dB (L), 15,5 dB (R)Corr Depth: 19,3 dB (L), 23,6 dB (R) Difference: -31,8 dBFS (L), -31,1 dBFS (R)

RME Babyface (didier.brest)
0,6 dB (L), 0,7 dB (R) Corr Depth: 9,2 dB (L), 10,8 dB (R) Difference: -28.8 dBFS (L) -29.8 dBFS (R)



External clocking


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)

Apogee Rosetta 800 clocked by Universal Audio Apollo (davidbayle)
Corr Depth: 39,7 dB (L), 41,7 dB (R) Difference: -45,9 dBFS (L), -45,8 dBFS (R)

Apogee Rosetta 800 clocked by Antelope Orion (davidbayle)
Corr Depth: 39,6 dB (L), 41,6 dB (R) Difference: -45,9 dBFS (L), -45,8 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)

Universal Audio Apollo clocked by Apogee Rosetta (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

Universal Audio Apollo clocked by Antelope Orion (davidbayle)
Corr Depth: 26,5 dB (L), 28,0 dB (R) Difference: -42,7 dBFS (L), -43,0 dBFS (R)

RME Fireface 800 clocked by Yamaha 02R (genelec79)
Corr Depth: 22,3 dB (L), 23,9 dB (R) Difference: -41,9 dBFS (L), -43,0 dBFS (R)



Sets of different converters


Synchronised mode (DA master unless otherwise specified)

MSB Platinum IV Plus DAC ---> MSB Studio ADC, master clock (DUJS)
Corr Depth: 62,8 dB (L), 63,4 dB (R) Difference: -77,7 dBFS (L), -77,4 dBFS (R)

ReQuest Audio, The Beast DAC ---> MSB Platinum Studio ADC (DUJS)
Corr Depth: 62,0 dB (L), 63,4 dB (R) Difference: -76,9 dBFS (L), -76,3 dBFS (R)

[email protected] XTe DAC ---> MSB Studio ADC, master clock (DUJS)
Corr Depth: 46,0 dB (L), 48,0 dB (R) Difference: -65,3 dBFS (L), -66,7 dBFS (R)

Atoll DAC200 --> RME Babyface (didier.brest)
Corr Depth: 10,1 dB (L), 11,9 dB (R) Difference -30,0 dBFS (L) -31.3 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)

Benchmark DAC1 phone output ---> Echo Audiofire 4 (david1103)
Corr Depth: 10,4 dB (L), 10,2 dB (R) Difference: -22,6 dBFS (L), -23,1 dBFS (R)

Unsynchronized mode. The DA and AD converters run each on its own clock. Measurement performed with sample rate drift correction activated in Audio DiffMaker and limited to the first half of the sample. Because the RMS level of this first half is 1 dB lower than the RMS level of the full-size sample, one must increase the difference RMS level figures by 1 dB for comparison with the synchronized mode.

ATOLL DAC 200 --> RME Fireface UFX (didier.brest)
2,3 dB (L), 2,4 dB (R). rate adj=8,8289 ppm.Corr Depth: 23,2 dB (L), 25,3 dB (R) Difference: -44,0 dBFS (L), -45,4 dBFS (R)

ATOLL DAC 200 --> RME Babyface (didier.brest)
Rate adj=-4,3193 ppm.Corr Depth: 10,8 dB (L), 13,0 dB (R) Difference: -31,5 dBFS (L), -33,1 dBFS

For contributing to this test :
- download the test 2 x 44 kHz x 24 bit wav file (31 Mb),
- cable link the 2 DAC outputs to 2 ADC inputs,
- play back the test file to the DAC outputs and record from the ADC inputs,
- upload to mediafire or any other file sharing site the 2 x 44 kHz x 24 bit wav file from this recording and
- provide the download link here with information about ADC and DAC identities.
Your contribution is welcome even if your converter is already in the above list.
Thank you!
Old 23rd January 2014
  #387
Lives for gear
 

Thanks for analysis! "Disappointing" numbers considering this is Focusrite's flagship interface still! It is older than the Forte and Saffire Pro 24, though. I wonder if Focusrite updated the converters.

I can also provide test results for Profire 2626 soon.

I honestly couldn't hear a difference between the Babyface and Hilo files when blind, though.
Old 24th January 2014
  #388
Lives for gear
 
didier.brest's Avatar
 

Quote:
Originally Posted by Melodeath View Post
I can also provide test results for Profire 2626 soon.
It would be great !


Quote:
Originally Posted by Melodeath View Post
I honestly couldn't hear a difference between the Babyface and Hilo files when blind, though.
We are many who can't...
Old 30th January 2014
  #389
Here for the gear
 

Echo Indigo IO, in/out loop:
0,616dB (L), 0,602dB (R)..Corr Depth: 26,1 dB (L), 27,8 dB (R) Difference: -40.5dBFS (L) -40.52dBFS (R)
Old 30th January 2014
  #390
Lives for gear
 
didier.brest's Avatar
 

Thank you ! Please could you upload your loopback file at a file sharing site, for instance at Simple File Sharing and Storage. and give here the download link ?
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
kraku / Gear Shoot-Outs / Sound File Comparisons / Audio Tests
261
esaias / Remote Possibilities in Acoustic Music and Location Recording
0

Forum Jump
Forum Jump