• Registered: 2023-06-08
  • Posts: 17

Topic: Is reported round trip latency correct when using DANTE/MADI/ADAT?

I've recently bought a UCX II and are thinking in investing more into the RME ecosystem in my semi-pro studio. I have previously used MOTU, and connected other MOTU converters via AVB and ADAT.

With MOTU I noticed that the measured round trip latency is different if I use the main interface's analog inputs and outputs compared to if I use analog inputs and outputs on the converters that go into my main interface via AVB.

AVB, DANTE, ADAT and MADI, by nature have additional latency I assume. That's just how it works. Please correct me if wrong.

With my RME UCX II the drivers report 100% correct RTL values according to RTL Utility when measured.

My question is if I use a RME UFX III as my main interface and connect an M-32 AD and an M-32 DA via MADI to my UFX III, will the RTL value from the driver report correct values for when running audio out and in via the M-32 converters?

  • Jeff
  • Jeff
  • Moderator
  • Offline
  • From: USA
  • Registered: 2007-05-21
  • Posts: 4,665

Re: Is reported round trip latency correct when using DANTE/MADI/ADAT?

will the RTL value from the driver report correct values for when running audio out and in via the M-32 converters?

Unfortunately, no. The unit can only report one latency value for the entire interface, that will be based on the built-in analog I/Os of the UFX III. There is no way to determine the latency of any converters added by digital connections except by ping testing.

Regards,
Jeff Petersen
Synthax Inc.

Jeff's Website

3 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 25,599

Re: Is reported round trip latency correct when using DANTE/MADI/ADAT?

stephan_martin wrote:

My question is if I use a RME UFX III as my main interface and connect an M-32 AD and an M-32 DA via MADI to my UFX III, will the RTL value from the driver report correct values for when running audio out and in via the M-32 converters?

In this specific case the difference of RTL for internal and external AD/DA will be only a few samples, so nothing to worry about.

Regards
Matthias Carstens
RME

MC's Website

4 Reply by stephan_martin (edited by stephan_martin 2023-10-30 19:17:31)

  • Registered: 2023-06-08
  • Posts: 17

Re: Is reported round trip latency correct when using DANTE/MADI/ADAT?

Ok, as I expected. But being so positively surprised how many well thought through features the RME products have, I thought you might have come up with a really smart solution for this as well!

With AVB or DANTE I guess it get's worse though? And perhaps noticeable is some cases. But there are workarounds if you are aware of the problem.

In the contact with MOTU they said that AVB have at least 0.7 msec latency by default. not sure if that is AVB or their implementation of AVB . When running the MOTUs at 44.1kHz and 128 sample buffer size I get 2 msec longer RTL than with the primary audio interface's analog in and outs.


One way to solve this I guess is to insert a delay in some stage in the faster parts of the system (the main interface hooked up to the computer). But perhaps that not really feasible today.
Maybe something that could be included in the AVB or DANTE protocol in the future. Similar to the feature that TVs connected via HDMI audio receivers communicate their image processing delay so that the receiver can sync the audio to the video.