Motu 16a
 

These new AVB interfaces are amazing. If you are recording with a Mac and have Yosemite, I highly recommend checking this out. Previously on PT12 with firewire device, I was getting 20 tracks stable and noise free at 256 samples 48khz. With the MOTU 16A, I just did a benchmark test, 24 tracks @ 96khz with setting of 32 samples, less than 3ms round trip and nary an issue. Call me impressed in the AVB protocol. :)

How did you measure the round trip latency?

Hi Amack,

Not the most accurate method, but I use the Logic Pro X audio interface settings panel. Probably not 100% accurate. But I have always used that for my round trip reading for all my interfaces because at least it's consistent. :)

If I record x tracks without pops and crackling, I consider it a pass, and I then use the setting readouts. Whether or not it's 100% accurate does not change the fact that this interface, through AVB, is measurably better than firewire interfaces.

I'd be happy to use whatever methods you guys use to provide some useful measurements. I'm extremely excited these AVB interfaces and their possibilities. The fact alone that I can get 128 channels in and 128 channels out simultaneously with AVB or thunderbolt, is amazing!

Just wanted to share my experiences. Happy to do some examples for people. :)

Hi ShiftStudios,

Thanks for sharing - that does sound like an awesome device. And at <1/3 the price of HD_IO, it sounds like the better performer (though its hard to tell because, unlike Motu, Avid apparently doesn't provide any real specs for the HD_IO).

How does it work on your PC?

Are you using Logic Pro for recording and mixing, or just recording?

V.R. :-)
Amack

ProTools Vanilla or HD will only go to buffer 32/64 at 44.1/96k

RTL will be A/D +(2X buffer)/SampleRate + D/A + any Thunderbolt overhead

MrMiller from MOTU has a lot of posts on this starting here - Somewhere in that thread there is an update from him about where they ended up. Great interfaces for sure but I would not totally trust what Logic says and sadly the Pro Tools answer may be different...

https://www.gearslutz.com/board/musi...erface-10.html

Audio over Ethernet is amazing!

Hi Amack,

I only used Logic Pro for the test, I generally use Pro Tools for all my record/mixing. :)

In regards to real world scenario I did the following last night:

45 tracks, lots of heavy plugins (H-Reverb, SSLChannel, RVerb, iZotope oZone 6 Advanced. Created 2 x stereo and 7 x mono Audio, but input, fed output from drum tracks to busses, armed and hit record/play. At 64 samples there was no perceivable latency in the drum tracks being recorded against the tracks playing and being recorded, not a stutter or crackle either. This was not possible previously for me, so it's definitely a marked improvement. :)

Cheers.

Hi propower,

You are right, it was 64, so apologies for that, the latency time is still correct though. :) As I told Amack in my last post, the test I ran last night proved a perceivable improvement to me. :)

Cheers!

Okay, so today I connected the 16A via thunderbolt to my MBP and did some testing. You can click the link below to see screenshot of Logic Pro X reporting latency and settings. Also I added what it appears like in Audio Midi Setup. 32 bits of 192kHz, very nice.

https://www.dropbox.com/sh/cwcxhpcou...xFONf5VGa?dl=0

I really feel that these AVB/Thunderbolt devices may be a game-changer. The noise floor on them is so immaculately low, it's almost surreal.

From the MOTU website

http://www.motu.com/products/avb/16a

State-of-the-art latency performance
Every digital audio workflow has some latency. The only question is, how much? In the case of the 16A, the answer is, not much at all. In the 1248, latency is measured in a handful of samples, from the time an audio signal first goes digital to its arrival somewhere else (to an output, the mixer, or the computer). By keeping latency this low in the hardware, the 16A helps minimize latency introduced by your host software. Over Thunderbolt, round trip host latency is an impressive 1.4 ms, from analog in, through the host, to analog out (@ 96 kHz with a 32-sample host buffer)

Yup, and my brief noob tests would seem to confirm their statement, if even at an amateur level. Regardless, I was just showing that what they state is basically confirmed, by me. ;)