Lowest latency Firewire audio interface for Logic... which should I choose?

I've used several MOTU interfaces and the Apogee Ensemble... if you're looking for low-latency with BFD, forget the Ensemble. It sounds great, but latency is a weak point. I can run a 64 sample buffer with BFD and MOTU Firewire interfaces at 96k, but with the Ensemble, 96k required a 256 sample buffer with BFD on the same machine (older Mac Pro 2.6 quad). Not to mention that Ensemble has significantly higher latency at any given buffer setting.

On the other hand, I'm not a big fan of the MOTU interfaces because 1) they have flaky S/PDIF inputs at 96k (confirmed this problem with tech support at Universal Audio), and 2) MOTU has an aliasing bug with SMUX ADAT at 96k (I've reported this to MOTU multiple times with no reply). However, for low-latency Firewire, MOTU is definitely a contender.

But I've never compared directly with the Fireface 800--I'd be very interested if anybody has directly compared the latency of a MOTU device with RME using BFD as the VI.

(Seeing as how the V.P. is such a V.I.P., shouldn't we keep the P.C. on the Q.T.? 'Cause if it leaks to the V.C. he could end up M.I.A., and then we'd all be put out in K.P.)

-James

I'm a very happy user of several RME 800's and I can only recommend this. Monitoring through the RME itself gives you 0 latency (could be one sample or so) and the Total Mix Software is extremly versatile for all routings you can think of. The Mic Preamps are great as well.
Tracking Software Instruments can be easily done at a buffer of 64/128 samples, while mixing you can then turn the Buffer up to 1024.

I'm using a Euphonix MC Mix to control Logic but the RME's Total Mix Software I control with the mouse. You can setup presets for various situations and just recall those in an instance.

Yes, it has an option in the preferences "Enable Midi Control" and you can choose individual Controllers for Midi Input and Mid Output

yeloop wrote:
Interesting to hear that the Ensemble is so ordinary when it comes to latency. For some users this would not be important, but for something like BFD, latency is a massive issue.

I spoke with a tech at Apogee about the Ensemble's latency--I was surprised there was so much and wondered if there was a problem with my system. The answer I got was that Apogee doesn't write the driver for the Ensemble--it uses Apple's generic (tweaked for Apogee ?) OS X Firewire audio driver, and Apogee is aware the latency is high. I was told there were no immediate plans for Apogee to write their own driver, and they weren't aware of any impending improvements on Apple's side. Their official word is to use Maestro for low-latency audio, and if you need low-latency VIs, buy the Symphony system instead.

Is your experience of the MOTU devices the latest 896Mk3, or the Mk2 versions?

The Mk2 versions. Would be surprised if the latency is significantly different (the Mk3s appear to use the same drivers), but can't confirm that...

-James

Sampleconstruct wrote:
Monitoring through the RME itself gives you 0 latency (could be one sample or so) and the Total Mix Software is extremly versatile for all routings you can think of.

Thanks for the info on the Fireface 800, and I'm sure it's a great piece of gear!

But, I think you're probably measuring the Total Mix latency incorrectly--the Fireface converters almost certainly have more than a few samples of latency. For instance, I did some latency measurements comparing the MOTU Traveler Mk2, Metric Halo 2882, and Apogee Ensemble, using "direct monitoring" on all three devices:

2882 (MIO): 75 samples (1.70ms at 44.1k, 0.78ms at 96k)
Traveler (CueMix): 73 samples (1.66ms at 44.1k, 0.76ms at 96k)
Ensemble (Maestro): 22 samples (0.50ms at 44.1k, 0.23ms at 96k)

The Ensemble's 22 samples is the lowest analog in->analog out latency I've seen on any digital device, so for direct analog audio monitoring (i.e. not monitoring through software), it's great. When I compared using external A/D/A converters routing via ADAT or S/PDIF, I got similar values for all three devices, so it appears the Ensemble's converters have lower latency than the other two devices, and the latency of the mixing itself is similar.

For comparison, ProTools HD reports 105 samples of latency at 44.1k (software monitoring through ProTools). AFAIK direct monitoring is not available with ProTools HD.

Check out the RME's latency again, if you get a chance--I'm guessing when you measured the latency, you maybe weren't getting the full roundtrip analog in to analog out path.

-James

If yeloop is using BFD, then direct monitoring isn't going to help him. That is for audio in.

FWIW, on MOTU Ultralites, I get a little (less than 1 ms) less latency with the Mk 1 rather than the Mk 3.

Perhaps the built-in DSP slows the Mk 3 down?

It is still 4.3 ms at 32 samples with the Mk 3. For MIDI controlled VIs, it would be less than that.

Thomjinx

Thomjinx wrote:
If yeloop is using BFD, then direct monitoring isn't going to help him. That is for audio in.

Exactly.

FWIW, on MOTU Ultralites, I get a little (less than 1 ms) less latency with the Mk 1 rather than the Mk 3.

Perhaps the built-in DSP slows the Mk 3 down?

Interesting--thanks for the info (one more reason not to upgrade MOTU, frankly...)!

It is still 4.3 ms at 32 samples with the Mk 3. For MIDI controlled VIs, it would be less than that.

I think I measured 3.9ms with my Traveler mk2 at 44.1k with a 32 sample buffer. So, it's a bit faster than the mk3. At 96k I can use a 32 sample buffer to get 2.5ms latency, but I can't run BFD2 with a 32 sample buffer. Best I can do with BFD2 is a 64 sample buffer at 96k... 3.16ms.

But, you're exactly right, Thomjinx: with a VI, you won't hear that full amount of latency. Assuming the MIDI latency is negligible, the VI response time will be roughly 1/2 the full analog audio latency. (Logic reports both of the these figures--output latency and full roundtrip latency--if you click on the latency info in the Audio preferences. Be aware, though, that some devices report this information correctly, while others do not. For instance, with the Apogee and Metric Halo products I've used, the numbers reported by Logic are exactly correct, while with MOTU, the numbers shown in Logic are lower than reality: with a 64 buffer at 44.1k the Traveler mk2 indicates 4.3ms roundtrip latency with Logic, but I measured it as 5.6ms... so you may be misled if you try to use the latency values reported by Logic to compare products... in that regard, MOTU seems to cheat, while Apogee and Metric Halo follow the rules 🙂

-James

Thanks for that info, James.

I was not sure how accurate Logic was reporting, but I was assuming it would be at least be consistent. But apparently not.

In any case, if one needs more than two ins and/or outs, the MOTUs offer a lot of bang for the buck.

What I would love is an Apogee Mobile Symphony! Yeah!

Currently I am running the Ultralite in a small 3 space rack, with two or three extra mic pres when I need to run more than two mics. For most of my live gigs, I just use two mics and use MainStage to change the settings on the mics. That way I can use one mic for Sax and harmonica, and the other for flute, pennywhistle, cornet and steel pan. EWI runs into a 1/4 inch line in. With the two-mic setup, I don't even need the rack.

Thomjinx

Thomjinx wrote:
Thanks for that info, James.

I was not sure how accurate Logic was reporting, but I was assuming it would be at least be consistent. But apparently not.

Technically, I think Logic itself is probably being consistent... the problem is that Logic relies on the latency reporting from the device drivers, and the MOTU drivers seem to underestimate their latency. On a related note, this is why Logic is able to place recorded audio from the Apogee and Metric Halo devices with sample accuracy automatically, while the MOTU devices will record audio slightly late (~40-100 samples depending on device) when you first plug them in. The MOTU devices can record perfectly in Logic, but you'll need to do a loopback test and set the delay compensation manually...

Currently I am running the Ultralite in a small 3 space rack, with two or three extra mic pres when I need to run more than two mics. For most of my live gigs, I just use two mics and use MainStage to change the settings on the mics. That way I can use one mic for Sax and harmonica, and the other for flute, pennywhistle, cornet and steel pan. EWI runs into a 1/4 inch line in. With the two-mic setup, I don't even need the rack.

Cool! I'm using the MOTU Traveler for live instruments, and it's working well, too. Definitely not the best-sounding box you can buy, but for size, features, and latency, it's a good match for live performance.

-James

This RME UFX looks Nice:

http://www.rme-audio.de/enproducts_firefaceufx.php

jnashguitar wrote:
Cool! I'm using the MOTU Traveler for live instruments, and it's working well, too. Definitely not the best-sounding box you can buy, but for size, features, and latency, it's a good match for live performance.

-James

As a side note...

Some of the touring Broadway shows, have been using MBP's, with Mainstage and a MOTU Traveler for the orchestra pit keyboards. This is what I would call mission critical as the show depends on these keyboard parts for efx, intros, musical cues and main backing keyboards.

pancenter-