First attempts using Dante via Digiface Dante and Logic I/O Utility went swimmingly. Latency as reported in the I/O plugin amounted to 3.5 ms @48kHz and a tad less @96kHz. Sent 2 mono tracks to 2 inputs into a reverb algorithm (mix at 40%) and bounced that down. 

Before doing the above, I tried using ADAT. Didn’t sound so nice. Latency was atrocious.

I’ll use headphones to compare the 48 and 96 recordings. 

This is odd. My Adat latency in and out of studio one is reported never above 200 samples at 48k (Thats less than 5ms). It’s often only 30 samples on most of the algos I gravitate to. 

Are you using ADAT in send/return fashion from within Studio One? How do you determine latency in Studio One? 

 

When you use pipeline it pings your external hardware and gives you a number. Pipeline will sync everything up if you hit the auto button. Analog is behind around 130 samples while ADAT is typically 30 or so. That’s for one algo and some algos are a little more. When you string them together it goes up but not by double.

Thanks for the explanation! I hadn’t paid much attention to Studio One. Poking around, I found one video about Pipeline when using a UAD interface that clarified why my previous trial use of H9000 ADAT in Logic sounded bad. I should have muted the ADAT channel in UAD’s Console app. The result was a normal-sounding signal. The round trip latency dropped to 21 samples, which is in line with what you reported.

Regarding 48 vs 96, I think I should compare mixes that involve a lot of tracks and fx. 

 

 

Hi there, glad you got it sorted. I too use A UAD Apollo and keep console muted, especially everything having anything to do with my Eventide, Also, mute any sends to the FX channel if you’re bussing. Ive seen some really wonky numbers that way, too. I actually made a Studio One Macro which I split up into 2 separate macros. With one macro I can populate 8 stereo FX channels with pipeline properly routed using all my ADAT and Analog IO for the session. Then I made macros giving me 4 stereo channels of each. Saves a LOT of time. As I populate an algorithm the first thing I do is ping it so I dont forget. Everything lines up perfect. This workflow took some time and really took learning pipeline and S1 better.