Jump to content
Kevin White

Anyone using Dante Drivers? (Latency discussion)

Recommended Posts

My earlier discussion of attempting to load multiple ASIO drivers from various devices into a CW project eventually led me to "audio over ethernet" explorations into Dante protocols.  At issue is multiple distinct recording stations in a house (during covid times), and I wanted to set folks up in their own space and then port into the DAW over the LAN. Dante solves this, but in doing so, other challenges arise.

Dante VIA is a software solution that translates various audio drivers on a device into Dante "language" for transmission over ethernet. It allows for 16 in/16 out with its drivers in CW.

Dante Virtual Soundcard (DVS) is a software program that is essentially a switching station. It doesn't translate, but allows up to 64 channels in/64 out into a DAW for Dante Enabled devices. (More on this in a moment)

Dante Controller is the LAN's Grand Central Station ... where all traffic (save preset VIA porting) can be routed between Dante Enabled Devices.

Due to limitations with an incompatibility w/ my laptop, (which I'm sorting out w/ customer support), I've been backed into running VIA on two high powered computers. Normally I'd run my Soundcraft Ui24r and port that through USB to my laptop - sending everything to the DAW via DVS. Until I sort out the incompatibility, I've got to connect the mixer to the same rig as the DAW. 

Here's where the oddity of the latency issue arrives. I'm using the Dante drivers in CW. I COULD choose the ASIO drivers of the mixer directly in the DAW, but I'd rather just leave everything in place and have it all work right - so I can add other stations at will into a common language system (and port those Dante Enable Devices into the DAW). 

In Sonar, Dante Via's drivers are enabled. The ASIO driver is set to 128. I tried 64 and 32, but the audio drivers failed. The performance level in Via is set to "High" (this is an AMD Ryzen 8 core CPU machine). 

I'm running a microphone into a channel on the Soundcraft, which is then translated by VIA and routed into Sonar through the VIA drivers (with aforementioned low latency settings). 

As we listened to real time performance, sync was perfect, but as the tracks recorded the audio data, on playback, they were almost an eighth note out of sync. Once I realigned the tracks (moved them back into sync) they played fine ... but I don't understand why so far out of sync in the tracking.

Any/all thoughts/fixes/points in right direction warmly appreciated!!

 

Edited by Kevin White

Share this post


Link to post
Share on other sites

Further study in different approach reveals the same issues.

I took the mixer and translated its drivers through Via into Dante environment ... controlling the mixer with my laptop. I then switched the DAW rig over to Dante Virtual Soundcard - which can be set as low as 4 ms.

These (DSV) are the drivers that loaded into Cakewalk. The benefit of using DVS versus VIA is instead a 16x16 configuration, you can expand up to 64x64.

I played/tracked a new guitar part to the source material, listening back on headphone through the mixer. 

The SAME thing happened in the tracking lag ... with the new track slightly off beat behind. 

It all recorded fine ... just out of time.

The original track clip began at 04:01:045 but had to be forwarded to 02:04:430 to sync up correctly.

I'm aware that this is not actually a Cakewalk issue. It's dutifully recording the data at the time it is receiving it.

My wonderment is merely whether any here have any experience dealing with the topic, and/or insights towards fixes they'd be willing to share.

Thanks again!!

Edited by Kevin White

Share this post


Link to post
Share on other sites

I got a mental cramp as I tried to translate your description into a topology diagram.  If you'd draw one for me, I can be more helpful.

Nevertheless, it seems to me that CbB can't see the reported latency of two "layers" of drivers, so its 'Recording Offset' setting needs to be changed to achieve alignment.  

Get an out-to-in jumper cable on the out-of-time recording input, and run a pre-recorded click track to it. Record the input click track coming through the loopback and adjust the Recording Offset value to make is all happy.  That's the best I can offer without a more clear understanding of what's connected to what.

Edited by MediaGary

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×
×
  • Create New...