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

Auralic LEO GX DAC Clock Review

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Master clocks were relevant in complex setups that needed to synchronise audio with video running on separate machines or cope with digital routing between various devices using SPDIF, ADAT, etc. Here is an example of a master clock device made for that purpose:

https://apogeedigital.com/products/big-ben
View attachment 46709

Post-production editing of video and audio is increasingly being done "in-box" on a single non-linear editor, so the problem of synchronising outboard devices is becoming obsolete.

I assumed as much, the problem I have with the unit reviewed, is it seems to be simply for a single device, to which I can't fathom even a reason in principle to use it, let alone practice >_>
 

lszomb

Member
Forum Donor
Joined
Jan 17, 2020
Messages
41
Likes
63
Location
Singapore
Btw: from what I heard. Arualic plans to release 4 pieces as a suit: a streamer / network player, a dac, a masterclock, plus ???. Maybe technically, it makes sense in that way if did not look at the price tag. Yes, I am poor.

But in China, they had sold quite some number of clocks as people always recommend this setup.
 
Last edited:

hugodlc

Member
Joined
Jun 26, 2019
Messages
18
Likes
42
House clocks are still used in post to generate a common clocking signal to time align audio and video equipment, the most expensive ones use GPS as their source clock, others use different crystals and multipliers to get to the different rates they need to output.

If you have 2 separate devices going into a third device and they all connect digitally, they need to be sample locked to the same clock or there will be nasty consequences.

Some of those atomic clocks were sold as snake oil to some smaller studios because the people that operate them have been to big studios where they are needed and they thought to achieve the same results, you need the same gear, ok maybe not quite so simple, but anyways, they were sold something they didn't need, and this is exactly the same here.

As has been previously mentioned, having a local clock next to the dac is the best performance you will get timing wise, as even if you manage to not degrade said clock over a long cable, the dac will have to resolve to this other clock and afaik that is never 100% as good as the internal clock which is usually already in superclock speed, and doesn't need to be multiplied, like the external clock, and any aberration will be much smaller compared to an outside clock.
 

milosz

Addicted to Fun and Learning
Forum Donor
Joined
Mar 27, 2019
Messages
583
Likes
1,643
Location
Chicago
how309-handbook-snake-oil-560.jpg
 

Xulonn

Major Contributor
Forum Donor
Joined
Jun 27, 2018
Messages
1,828
Likes
6,311
Location
Boquete, Chiriqui, Panama
Btw: from what I heard. Arualic plans to release 4 pieces as a suit: a steamer...

A steamer? Can they beat Red Pro's $70 Amazon price?

Hair Steamer.jpg
 

milosz

Addicted to Fun and Learning
Forum Donor
Joined
Mar 27, 2019
Messages
583
Likes
1,643
Location
Chicago
The reason Amirm didn't see any improvement in results is because he was not using the $11,000 AC line cord and he doesn't have Bybee Quantum Purifiers in his test gear. However, he could just call Machina Dynamica and they can send tones over the phone that will fix everything.

FYI not making any of this up. I am absolutely convinced there has been some kind of dimensional rift and that we are now living in a continuum where truth still exists but just doesn't matter.
 

Xulonn

Major Contributor
Forum Donor
Joined
Jun 27, 2018
Messages
1,828
Likes
6,311
Location
Boquete, Chiriqui, Panama

Anmol

Member
Forum Donor
Joined
Apr 24, 2019
Messages
89
Likes
109
Great idea ;), trying PTB:
PING www.ptb.de(www.ptb.de (2001:638:610:bd02::12)) 56 data bytes
^C
--- www.ptb.de ping statistics ---
6 packets transmitted, 0 received, 100% packet loss, time 5107ms
Doesn't work. Lets try a server we trust: :)
ping audiosciencereview.com
PING audiosciencereview.com(2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d)) 56 data bytes
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=1 ttl=59 time=10.3 ms
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=2 ttl=59 time=9.64 ms
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=3 ttl=59 time=9.60 ms
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=4 ttl=59 time=9.74 ms
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=5 ttl=59 time=9.37 ms
64 bytes from 2606:4700:3037::681b:b60d (2606:4700:3037::681b:b60d): icmp_seq=6 ttl=59 time=9.64 ms
^C
--- audiosciencereview.com ping statistics ---
6 packets transmitted, 6 received, 0% packet loss, time 5008ms
rtt min/avg/max/mdev = 9.369/9.719/10.324/0.293 ms

Works - if you can live with 0.293ms jitter...:oops:
Ping uses icmp protocol which you are showing above. Network Time syncronization uses ntp which is a different protocol. Different beast , different standard. Ntp will synchronize perfectly and accurately if configured correctly in a data center environment. There is more on following link.
https://www.eecis.udel.edu/~mills/ntp/html/warp.html
 

lszomb

Member
Forum Donor
Joined
Jan 17, 2020
Messages
41
Likes
63
Location
Singapore
Apart from the joke, I'd like to ask @amirm for more.

Reference: https://hometheaterhifi.com/reviews/dac/auralic-vega-g2-dac-leo-gx-master-clock-review/

Let me try to keep this short. Under the name of science review, we have to show how serious we are.
  1. Leave the high price tag for now.
  2. With the assumption of masterclock will help to "correct" or "fix" the jitter coming along with the input data and also the intrinsic jitter on different stages,
  3. With the assumption of the final output is still limited by the digital processor itself,
  4. The current test was not able to demonstrate the value of a masterclock as the input data is almost perfect.
  5. Under current test setup, DAC + Masterclock supposed to show a minimum improvement and the test had verified that.
The above reference I gave went a step ahead by adding jitter to the input data and the masterclock did fix it according to the link. So in that case, there is value for adding a masterclock to the system.

However, the story should not stop just here. We shall also discuss:
  1. What are possible input jitter sources and how to measure those?
  2. Is masterclock designed (in Hi-Fi scenarios) to solve a real problem or just try to address a hypothesis?
  3. If there was a real problem, how severe it could be?
At last, we shall look into the different solutions if we potentially need to have a masterclock in our system and compare the solutions with their price tags.

For example, if this masterclock can address the last mile issue within a digital Hi-Fi system and it costs 5k to make, we cannot blame Auralic sells it at current price tag unless we can prove that the same thing we can achieve with only 1k. Or we can look at how much other company sell the solution for, for example, dCS.

About 8 years ago, I made an external clock with FE-5680A Rubidium Atomic Clock and I can conclude that was just piece of juke and worth nothing. However, I would not jump to conclusions too fast to say Leo Gx is the same.

Cheers,
 

pierre

Addicted to Fun and Learning
Forum Donor
Joined
Jul 1, 2017
Messages
962
Likes
3,047
Location
Switzerland
On the bright side you could install these in geographically disperate data centers to try to coordinate globally unique transaction locks or mount them in the trunk of your autonomous vehicle research mule to synchronize your sensors for your fusion algorithms.

Maybe add vibration testing to your gauntlet to see how it holds up?

If you are wondering why distributed databases use atomic clocks you can read this Research paper.
what kinf of atomic clock do they put it inside? Rubidium? That could explain the price. I think they are also used for the GPS system.

Ping uses icmp protocol which you are showing above. Network Time syncronization uses ntp which is a different protocol. Different beast , different standard. Ntp will synchronize perfectly and accurately if configured correctly in a data center environment. There is more on following link.
https://www.eecis.udel.edu/~mills/ntp/html/warp.html

ntp synchronization between data centers which are phisically complicated is difficult to get accuratly. Some at least use atomic clocks to get the time right (like the gps system).
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
The above reference I gave went a step ahead by adding jitter to the input data and the masterclock did fix it according to the link.
Well, this is what he did:
1579581663523.png


That is half a millisecond of jitter which is HUGE! I measured 0.5 nanoseconds in recent measurement of the minidsp udio-8 and called that bad. This is 1000 times more jitter! Even with that, he got a tiny, tiny blip.

Here, we are talking about network and USB playback anyway. In both cases the input is a data pipe and its jitter is immaterial. THe clock would be locally generated by the DAC since the input is not synchronous.
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
3,427
Likes
3,981
Location
SoCal
How can an external clock help to correct external jitter anyway when clearly the internal clock/distribution is clean enough to not show any artifacts on the normal jitter test? Those test results look fishy.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
That is half a millisecond

0.5μS ... still a lot.

I suspect the device needs to warm up and stabilize and that might be responsible for the long boot period.

picture not optimized nor clarified... just stolen elsewhere:
LEO-GX-Inner-Top.jpg
 
Top Bottom