• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Lynx Aurora 8 Review (8-channel DAC/ADC)

Rate this DAC:

  • 1. Poor (headless panther)

    Votes: 6 5.9%
  • 2. Not terrible (postman panther)

    Votes: 35 34.7%
  • 3. Fine (happy panther

    Votes: 54 53.5%
  • 4. Great (golfing panther)

    Votes: 6 5.9%

  • Total voters
    101

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
2,220
Likes
2,117
Location
.de
Sold to a mastering studio in Paris 7 odd years ago.
Which arguably is the kind of place a holy grail converter should be at. In 1998, it was absolute state of the art. These reportedly will hit over 120 dB in loopback, absolutely unheard of at the time. You won't find many things with no less than 8 (!) Motorola 56009 DSPs either. (Mind you, they cost as much as a decent car and never turned a profit either...)

Speaking of insanely good converters from 1998, another good candidate for Amir to be measured would be the Prism Sound Dream AD-2 ADC. Its dynamic range spec is 130dB unweighted (!), though that's at a whopping +28 dBu. Still, we're talking "capture the entire dynamic range of a large-diaphragm condenser" kind of terrain.

Coming back to the Aurora 8, its specs are as follows:

Analog I/O
[...]
Level+4 dBu nominal / +20 dBu max. or
-10 dBV nominal / +6 dBV max
Input ImpedanceBalanced mode: 24k Ω
Unbalanced mode: 12k Ω
Output ImpedanceBalanced mode: 100 Ω
Unbalanced mode: 50 Ω
Output Drive600 Ω impedance, 0.2 μF capacitance

Analog In Performance
Frequency Response20 Hz - 20 kHz, +0/-0.1 dB
Dynamic Range117 dB, A-weighted
Channel Crosstalk-120 dB maximum, 1 kHz signal, -1 dBFS
THD + N-108 dB (0.0004%) @ -1 DBFS
-104 dB (0.0006%) @ -6 DBFS
1 kHz signal, 22 Hz - 22 kHz BW

Analog Out Performance
Frequency Response20 Hz - 20 kHz, +0/-0.1 dB
Dynamic Range117 dB, A-weighted
Channel Crosstalk-120 dB max., 1 kHz signal, -1 dBFS
THD + N-107 dB (0.00045%) @ -1 DBFS
-106 dB (0.00050%) @ -6 DBFS
1 kHz signal, 22 Hz - 22 kHz BW

Converters used are
TypeDR nomTHD+N nom
ADCCS5381120 dB(A)-110 dB @ -1 dBFS
DACCS4398120 dB(A)-108 dB @ 0 dBFS

I don't think either chip is good for much more than about 117 dB(A) in real life, so the Aurora is pretty much maxing them out. The new Focusrite Clarett+ 2Pre with CS5381 makes it to 117.5 dB(A) according to @Julian Krause.
 
Last edited:

manisandher

Senior Member
Joined
Nov 6, 2016
Messages
407
Likes
320
Location
Royal Leamington Spa, UK
Speaking of insanely good converters from 1998, another good candidate for Amir to be measured would be the Prism Sound Dream AD-2 ADC.

Not quite the same level, but I have an AD-124 in my 'vinyl studio'. I've compared it to the RME ADI-2 Pro, and it's really hard hearing a difference between needle drops made with each. The Prism is perhaps a little bit 'softer' sounding than the RME. The RME (new) cost about twice what the Prism (used) did, but has an excellent DAC and headphone amp too (and loads of other neat functionality, of course).

I think it's hard justifying these old converters when new ones are so good.

But still, they were things of utility and beauty in their time.

Its dynamic range spec is 130dB unweighted (!), though that's at a whopping +28 dBu. Still, we're talking "capture the entire dynamic range of a large-diaphragm condenser" kind of terrain.

I guess it very much depends on the mic preamp you use too, no? I've never tried my Prism with one, as I just use the MOTU UltraLite-mk5 when I want to record with a mic, which seems to have a pretty good preamp built in.

Mani.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
2,220
Likes
2,117
Location
.de
I think it's hard justifying these old converters when new ones are so good.

But still, they were things of utility and beauty in their time.
Indeed. Some features like fancy dithering in the ADC you just don't need any more these days, and it takes a very Legit Studio for analog input levels as high as +28 dBu. At least pro-level ADCs should generally have digital filters good enough to reject the assault of high-frequency vinyl noise.

Do you happen to know what sort of converter tech the AD-124 uses? This model seems to be from about 1996, and nothing available at the time would have been capable of -120 dB (unwtd) worth of idle channel noise, assuming it's not a composite ADC with a bunch of CS/AK5390 as pioneered a few years earlier.

I guess it very much depends on the mic preamp you use too, no?
Sure. Getting 130 dB through one is a substantial challenge. Probably not a standard topology.
 

hollis

Member
Joined
Sep 3, 2018
Messages
96
Likes
68
One other note. If one wants to do immersive audio, either via Atmos or Auro3d, from a PC, then the Aurora16 might be your only USB option. 7.1.4 would take up quite a few channels, especially if one runs multiple subs.

One could buy 2x Okto Dac8 and run them as an aggregate device in OSX, but they would not stay sample accurate, no way to tell how well they would stay in time.
Rather than USB, you could run them via AES off the 16ch Lynx PCIE AES card.
But if that $400-700 card is in needed, then you could simply drop the usb reqirement and get 2x $700-1000 Aurora8s.
The Aurora16, with a usb card, seems to be the best bet for 16ch and usb. There are pleny of other 16 cards to consider though.

OSX now has native support for Atmos, and win10 can play Atmos after jumping through a few hoops. Auro3d is pretty much dead, for that one would likely run a 7.1 bedlayer, then use the Auromatic upmixer ($500-600) as a VST for heights.
 

GTAXL

Member
Joined
May 15, 2020
Messages
32
Likes
36
Location
Mount Vernon, Ohio
Sold to a mastering studio in Paris 7 odd years ago.
Figures, everyone seems to let it loose. I have a friend that has a PM2 in his studio. I've tried to convince him to send it to be measured but he can't part with it even temporarily, he uses it daily in his studio and knows what a gem he has and ain't gonna be letting it loose.
 

L-Train

Member
Joined
Nov 17, 2018
Messages
22
Likes
46
I looked at my measurements in post #56 again and realized in my haste I incorrectly used a plain sine wave for the 12 kHz jitter measurement (duh). Using the J-Test FLAC files found at https://hydrogenaud.io/index.php?topic=108927.0, I re-ran the measurement. Again, the testing device is the Lynx E44 and all screenshots are in this order horizontally:
  1. Aurora 8 internal clock
  2. Aurora 8 clocked by AES with SynchroLock OFF
  3. Aurora 8 clocked by AES with SynchroLock ON

J-Test (48 kHz, 24 bit)

Aurora 8 to E44 - 48 kHz 24 bit - J-Test - Aurora internal clock.png Aurora 8 to E44 - 48 kHz 24 bit - J-Test - E44 AES master - Aurora SynchroLock OFF.png Aurora 8 to E44 - 48 kHz 24 bit - J-Test - E44 AES master - Aurora SynchroLock ON.png

Zoomed:
Aurora 8 to E44 - 48 kHz 24 bit - J-Test - Aurora internal clock zoomed.png Aurora 8 to E44 - 48 kHz 24 bit - J-Test - E44 AES master - Aurora SynchroLock OFF zoomed.png Aurora 8 to E44 - 48 kHz 24 bit - J-Test - E44 AES master - Aurora SynchroLock ON zoomed.png


The jitter components are only a few dB above -140 dB which is pretty good, and is consistent between internal and external clocking (with SynchroLock enabled and locked). As with my previous measurements, there's a slightly wider skirt at the base of the fundamental when using an external clock and SynchroLock compared to the internal clock.
 

GTAXL

Member
Joined
May 15, 2020
Messages
32
Likes
36
Location
Mount Vernon, Ohio
I looked at my measurements in post #56 again and realized in my haste I incorrectly used a plain sine wave for the 12 kHz jitter measurement (duh). Using the J-Test FLAC files found at https://hydrogenaud.io/index.php?topic=108927.0, I re-ran the measurement. Again, the testing device is the Lynx E44 and all screenshots are in this order horizontally:
  1. Aurora 8 internal clock
  2. Aurora 8 clocked by AES with SynchroLock OFF
  3. Aurora 8 clocked by AES with SynchroLock ON

J-Test (48 kHz, 24 bit)

View attachment 211881 View attachment 211882 View attachment 211883

Zoomed:
View attachment 211884 View attachment 211885 View attachment 211886


The jitter components are only a few dB above -140 dB which is pretty good, and is consistent between internal and external clocking (with SynchroLock enabled and locked). As with my previous measurements, there's a slightly wider skirt at the base of the fundamental when using an external clock and SynchroLock compared to the internal clock.
Great stuff! Can you do measurements of the ADC?
 

L-Train

Member
Joined
Nov 17, 2018
Messages
22
Likes
46
Great stuff! Can you do measurements of the ADC?

Sure. Here are a few measurements of 1 kHz sine waves at 4 different generator levels at 44.1 kHz, 24 bit. The screenshots are in this order horizontally:
  1. Lynx E44 loopback
  2. Lynx E44 to Aurora 8 (internal clock)
  3. Lynx E44 to Aurora 8 (AES external clock with SynchroLock ON)

-1 dBFS

E44 loopback - 44.1 kHz 24 bit - 1 kHz sine at -1 dBFS.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -1 dBFS - Aurora internal clock.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -1 dBFS - Aurora SynchroLock ON.png


-3 dBFS

E44 loopback - 44.1 kHz 24 bit - 1 kHz sine at -3 dBFS.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -3 dBFS - Aurora internal clock.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -3 dBFS - Aurora SynchroLock ON.png


-6 dBFS

E44 loopback - 44.1 kHz 24 bit - 1 kHz sine at -6 dBFS.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -6 dBFS - Aurora internal clock.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -6 dBFS - Aurora SynchroLock ON.png


-12 dBFS

E44 loopback - 44.1 kHz 24 bit - 1 kHz sine at -12 dBFS.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -12 dBFS - Aurora internal clock.png E44 to Aurora 8 - 44.1 kHz 24 bit - 1 kHz sine at -12 dBFS - Aurora SynchroLock ON.png


The Aurora's ADC is just as clean as its DAC section with low noise and no spurious tones or 60 Hz harmonics. Again, slightly wider skirt with the external clock. The results here appear to be limited by the E44's DAC.
 

L-Train

Member
Joined
Nov 17, 2018
Messages
22
Likes
46
Alright, here's another set of measurements for the Aurora 8. This time it's THD vs. frequency at 4 different generator levels. The test setup is the same: the Aurora 8 is loopbacked with an E44 (i.e. outputs connected to inputs and vice versa), both set to +20 dBu FS, but this time at 96 kHz, 24 bit. This sample rate was chosen because the noise shaping of the CS5381 ADC in the E44 beyond ~50 kHz when using a 192 kHz sample rate didn't make for clean and easily comparable measurements. The Aurora 8 is using its internal clock since the THD performance is the same as with an external clock. The screenshots are in this order horizontally:
  1. Aurora 8 output -> E44 input (THD vs. frequency of the Aurora 8 DAC)
  2. E44 direct loopback (reference)
  3. E44 output -> Aurora 8 input (THD vs. frequency of the Aurora 8 ADC)

-1 dBFS

Aurora 8 to E44 - 96 kHz 24 bit - THD vs. frequency -1 dBFS - Aurora internal clock.png E44 loopback - 96 kHz 24 bit - THD vs. frequency -1 dBFS.png E44 to Aurora 8 - 96 kHz 24 bit - THD vs. frequency -1 dBFS - Aurora internal clock.png


-3 dBFS

Aurora 8 to E44 - 96 kHz 24 bit - THD vs. frequency -3 dBFS - Aurora internal clock.png E44 loopback - 96 kHz 24 bit - THD vs. frequency -3 dBFS.png E44 to Aurora 8 - 96 kHz 24 bit - THD vs. frequency -3 dBFS - Aurora internal clock.png


-6 dBFS

Aurora 8 to E44 - 96 kHz 24 bit - THD vs. frequency -6 dBFS - Aurora internal clock.png E44 loopback - 96 kHz 24 bit - THD vs. frequency -6 dBFS.png E44 to Aurora 8 - 96 kHz 24 bit - THD vs. frequency -6 dBFS - Aurora internal clock.png


-12 dBFS

Aurora 8 to E44 - 96 kHz 24 bit - THD vs. frequency -12 dBFS - Aurora internal clock.png E44 loopback - 96 kHz 24 bit - THD vs. frequency -12 dBFS.png E44 to Aurora 8 - 96 kHz 24 bit - THD vs. frequency -12 dBFS - Aurora internal clock.png


The harmonics of the E44 in loopback are consistent across the frequency range so that makes it a decent reference. The Aurora 8 DAC has a slight uptick in THD at the ends of the frequency range mostly due to the 3rd harmonic. There's also an interesting dip in the 2nd harmonic between 2 kHz and 10 kHz, with a subsequent rise. Also, at -6 dBFS the 2nd harmonic appears to overtake the the 3rd. The Aurora 8's ADC performance is almost identical to the E44 loopback except for a slight rise in the 2nd harmonic at the upper end of the frequency range. We're pretty much limited by the E44's DAC here, but we can be confident that the Aurora 8's ADC performance is at least at that level.
 
Last edited:

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
2,220
Likes
2,117
Location
.de
If you have any line-level attenuators at hand, those tend to be quite useful in more accurately determining output distortion levels. Once you're say 20 dB down, the output tends to be dominating by far.

Conversely, an amplifier of known gain can be used to give a better estimate of output noise levels. (You can even recruit a little mixer with mic inputs for this task. Have REW generate a -31 dBFS tone, set input gain to maybe +40 dB, channel to 0 dB and main out to about -10 dB, fine-tune gains until ADC input level hits -1 dBFS, et voilà, a 30 dB amplifier.)
If you have a very good amplifier - one of the usual very well-measuring headphone amp models with balanced I/O comes to mind (e.g. Topping A30Pro or something) - you can also use that to amplify an output with hopefully benign distortion levels until the ADC side breaks a sweat.
 

Scrappy

Member
Joined
Aug 4, 2022
Messages
40
Likes
14
I agree that it's not the best design, but it does appear to be standardized by AES for audio too (AES59) , thanks to the likes of Tascam/Yamaha and DigiDesign. These converters you get the breakouts direct from the manufacturer and they will all be properly labelled. You can get custom lengths of such cables made by Redco Audio, which made my breakout cable for my Lynx AES16e, I just had a single AES out and AES in made to make the breakout efficient. Yes, they label each connector for the cables.
>Tascam/Yamaha and DigiDesign.

Ho boy did I learn this few years ago. IIRC was for faancy analog summing for touring. Guy had a Rivage too!!
 

Scrappy

Member
Joined
Aug 4, 2022
Messages
40
Likes
14
Wait, really? Wouldn't the two inputs be from separate PDU's fed from separate circuits in the rack, one being general mains and the other being an essential UPS/(+Generator) supply, so in a power outage, you don't lose all your processing, much like i dunno, every data center in existence?

To me, nothing would scream professional louder than processor with a redundant power supply setup.
More PSU more better. 100% correct with UPS feeding second supply. That’s been standard for touring for decades.
 
Top Bottom