View Single Post
Old 9th November 2015 | Show parent
Registered User
First thing's first; Hi Gearslutz forum members! This is my very first post on GS

Originally Posted by daxmaestro ➡️
I no longer have either one. I've never recorded anything through the zoom's inputs, but on the da side of things, I can say it held its own against the quartet. Definitely in the mid-end range of gear.
It's unfortunate that you no longer have the UAC-2, I just bought one myself a few days ago (partially because of your comments -and Tui's- regarding its low latency capabilities) and I would have liked to ask you a few questions regarding your UAC-2 performance compared to mine, and that could have involved doing some simple tests on your UAC-2, but this won't be possible anymore since you no longer have yours.

In any case, I want to thank you for sharing your experience with the UAC-2 Overall I'm satisfied with its performance, even though I was not able to achieve the same low RTL that you did. My RTL results (as measured by Oblique, Centrance) are a bit higher than yours, probably because of our computers USB 3 chipset differences (Intel USB 3 chipset in your case, vs the Renesa USB 3 chipset in my case). BTW, Zoom really puts an emphasis in using the Intel USB 3 chipset over other USB3 chipsets, and they do not guarantee compatibility with other chipsets, but it seems that my notebook's Renesa chipset also works fine.

My RTL results are as follow (tested in Reaper 5.1pre3, i7 notebook's USB3 port*, windows 7 64bit):

At 48Khz sampling rate: 24 samples/3.3 ms, 32 samples/3.6 ms, 64 samples/4.9 ms, 128 samples/7.6ms

At 96Khz sampling rate: 24 samples/NA, 32 samples/2.1 ms, 64 samples/2.8 ms, 128 samples/4.1ms

FWIW, even at 48Khz/24 samples buffer, I could play a small project with a few VSTs without obvious "clicks and pops". The combination of sampling rate and buffer that gave me the smallest RTL while still being able to play 3 instances of Pianoteq 5 (a piano VST that can be quite demanding) + 2 live vocal tracks with multifx on each, all without any hiccups whatsoever, was at 96Khz/64 samples/2.8ms...Nice! Of course, if you have a high track count project, you'll need to raise the buffers to 128 or 256 but for small projects, this UAC-2 is FAST!

* Note that when I tested my UAC-2 with a USB2 port instead of a USB3, the RTL increased slightly at certain sample rates. For example: 96k/32 samples -->2.1ms(USB3) vs 2.4ms(USB2). So it seems that using USB3 is making a real difference. And using a USB3 intel chipset, an even bigger difference (based on your tests results vs mine).

From what I've been reading about USB3 vs USB2 implementation (point-to-point links in USB3 vs multiplexing/polling in USB2) , in theory USB3 can achieve not only much higher speed transfer/throughput (~x10) but also much lower latencies (~ 50usec) vs USB2's minimum "streaming buffer" of 1ms, which would explain why the UAC-2 is capable of such low latencies. Just for fun, I tested how low the UAC-2 could go and at [email protected] I got 1.1ms RTL! Granted, it was unusable (heavy clicks and pops) but in theory this would be impossible to achieve with USB2 (USB2 has a minimum processing time of 2ms (1ms send/1ms receive)

So if this is all true (I need to analyze this further) it means that IF audio-card manufacturers implemented their USB3 interfaces properly, they could achieve much better round-trip latencies than current USB2 interfaces (i.e. RTL similar to their PCIexpress/ThunderBolt card) We will see...Interesting times ahead

BTW daxmaestro, is there any special reason you got rid of your UAC-2?