View Single Post
Old 02-27-2016, 07:58 AM   #39
chucky5p
Human being with feelings
 
Join Date: Apr 2011
Location: Canada
Posts: 454
Default

TheWhistler,

Most likely, your problem was fixed by behringer's latest driver (v3.29.0). Back in September 2015, I tried 4 USB interfaces, among which, a UMC204HD and a UMC404 (the others were Tascam US2x2 and Zoom's UAC-2) Both UMCs had the same behavior; clean sound for a while, then interrupted by a short gap/silence, and this independently of buffer size, driver version (tried all of the available ones), USB protocol (2 or 3) and PCs (tried on 3 different PCs; same issue) So I returned both UMCs back to the store. This latest driver must have fixed the issue. I noticed it's a universal driver for all UMC audio interfaces. This fix is great news for all UMCs owners

Aside from that issue, the UMCs had good sound but I would hardly call them "great" audio interfaces. Both had much more noise and distortion than either the US2x2 or UAC-2 and while their round-trip latency was "low", it was no where near as low as Zoom's UAC-2, but more worrisome is that the UMC driver "reports" artificially low latency results which can "throw-off" any DAW's Plugin Delay Compensation mechanism(PDC), which in turn may produce "misaligned" tracks when tracking (unless you "offset" for it by the proper amount in your DAW). This is caused by the UMC driver which has a relatively large "hidden" buffer that add ~3ms to whatever you see in your DAW, so the REAL latency is always more than what's shown in your DAW, but if you use a round-trip latency test utility like Centrance you'll see the UMCs true latency (look at screen caps below). Note that these results are based on the old drivers and it is possible that the latest drivers from Behringer have fixed this "false" latency reporting. However, since I don't have the UMCs anymore, I cannot confirm/deny this, but it is easy to verify yourself by using the Centrance utility:

https://centrance.com/downloads/ltu/

FWIW, if you like the "low" latency of your UMC404HD, you'd LOVE the "amazingly low" latency of the UAC-2 (less than half of the UMC404!). However, the UAC-2 is relatively much more expensive than the UMC404HD(~250$ vs 100$), and more importantly, if you're happy with your UMC404, it's all that really matters

Test results of 5 audio interfaces below (UA-1G is Roland's. It's a 1st gen USB 1.1 audio interface):



Below: Latency results of UMC-204HD (same driver/performance as UMC404) at 48Khz and buffer-size of 64 samples (lowest size available). Notice Reaper showing ~2.3/2.3ms (~4.6ms total) which IF was accurate, would be VERY good, vs real latency of 7.73ms as measured by Centrance, so about 3ms of "hidden buffers". Still, 7.7ms is unnoticeable by most people. Only e-drummers would notice that kind of latency and they would prefer <5ms.



Below: Latency results of UAC-2 at 48Khz and buffer-size of 24 samples (lowest size available) as measured by Centrance. Since the UAC-2 driver doesn't have any "hidden" buffers, Reaper is showing ~0.5/2.8ms (~3.3ms total), which is the same result as Centrance (3.35ms) so in this case, the Zoom has ~4.3ms less latency than the UMC! Also, the UAC-2 can operate at much higher sample-rate (i.e. 96k, 192k) and still use low buffer-size (32spl) with "extremely" low resulting round-trip latencies (2.2ms and 1.2ms respectively) as long as your projects are not too complex (a few VSTs).


Note that the UAC-2 is far from perfect. It has ground-loop problems, it's input jacks are WAY too tight (you need to use insane amount of force to plugin a 1/4in cable) but by far it's biggest issue is a "bump" of ~+0.5db in frequency response around 450hz that will "color" anything going through it. Although 0.5db doesn't seem much, in 2016, where < +/- 0.1db is common, IMO it's totally unacceptable. Even more so since most people can hear > 0.3db of coloration, therefore 0.5db is clearly detectable. So I would never use the UAC series as a "reference" converter (the UAC-8 has the same issue). However, this bump doesn't affect the inputs, only the outputs. So if you use the UAC-2 to track guitars, vocals, etc, it wont affects the recordings nor the renderings, only what you are monitoring so you can always use Reapers's monitor FX with ReaEQ to offset this playback bump. Look at the chart below. The UAC-2 is in pink:


If you want similar low-latency performance with even better specs all-around than the UAC-2, (lower noise, no freq bump) you would need to go with RME's BabyFace-PRO interface. However, it is MUCH more expensive than the UAC-2 (~3x more) IMO, within its limitations, currently the UAC-2 represent the best-bang-for-your-buck and it is a Low-Latency Monster

Hope this helps,

Chuck
chucky5p is offline   Reply With Quote