View Single Post
Old 01-16-2018, 09:54 AM   #15
serr
Human being with feelings
 
Join Date: Sep 2010
Posts: 12,560
Default

I've asked a few loaded questions here...

I get the sense that you never actually ran loopback tests at different block size settings and at every sample rate to initially shake out your system and see where the true limits are. I suspect it just worked well enough out of the box the first day and you just ran with that. (Not the worst thing. The success part speaks for itself. Just that this means you might have been closer to some edge than you realized.)

This is how I'd approach this:

1. Boot from an install of OSX 10.12 or 10.11 (or whatever older known stable system you've used without error in the past). Yep, we're shotgunning OSX right off the bat because of all the badmouthing of it going on!

2. Do that initial system shakedown to find your baseline round trip latency for every sample rate. Your replies above suggest you believe this is a sound quality discussion. It is pointedly not! This is purely about most efficient operation with the attitude that ANY sample rate that is more efficient for the system is just fine for sound quality.

3. Understand that a slow drive connected by USB is a CPU hit. Not just from the drive speeds or the USB speeds either. USB requires CPU cycles to manage. There's always an initial lag from both USB and seek time of slow HDDs that could easily shut you down in a live system. Always try a faster drive if you have problems. (Like that screamin' pci connected SSD! Since you have this style drive that is so many magnitudes faster than literally any other available technology, basing everything - OS, samples, backing tracks, live multitrack recording - on it would be higher performance than introducing any slower drive into the system.)

You have mentioned doing some post work with the system in some of this conversation. (Or I misread.) Let's keep live work and post work separate! You can always (and should) set the block size to 1024 samples for post work since you're never monitoring live real time events in sync with a live mix/processing. The issue is with the low latency live sound settings, so keep these jobs separate. It's typical to switch between live settings and post work settings FYI. Running the system with a low block size for post work would be like driving on the freeway in first gear (for a bad car analogy).

I hope some of this helps!

PS. I don't see a direct clue anywhere or I'd come right out and say so. All I can offer is to go through the system methodically and to actually measure your latencies at different settings so it isn't a mystery variable.

My guess is you are going to find two things. Probably a bug or incompatibility of some 3rd party plugin with OSX 10.13 and then possibly a more efficient setup - and thus more headroom - for your system.
serr is online now   Reply With Quote