latency between devices when using Audio Midi Setup's Aggregate Devices?

Question

latency between devices when using Audio Midi Setup's Aggregate Devices?

Any body using Aggregate devices? If so, do you experience any latency between the two (or more) devices? I have a tascam fw1082 and a presonus firepod working together, with the former as the master clock.

If i mic a snare with mic in's from both units, you can see the latency of the presonus unit.

Any solutions to get them to sync?

iBookG4 & DP-g5, Mac OS X (10.4)

Posted on Mar 14, 2006 5:33 AM

Reply

Answer 1

jord

Level 5

5,230 points

Mar 14, 2006 5:58 AM in response to SpartanDelicious

I'm aggregating a MOTU 24I/O and a ProjectMix I/O. Both of them are sync'd by the word clock from my MTP/AV. I'm not experiencing latency issues as this. I did a session last night using two mics on the same source (one going into the ProjectMix and the other going to the 24I/O). Both sources were in sync.

Did you set your 1082's clock source in it's own control panel (setting it in AMS is not enough)? As well, did you specifically set your Firepod as a slave?

jord

Reply

Answer 2

Mar 14, 2006 8:01 AM in response to jord

You can't assume that two interfaces in an aggregate device will be in perfect sync, even if they're running off the same master clock.

Why? Because the converter chips themselves have a small amount of latency, which is not handled by Core Audio. This latency is different for different models of converter chips.

So, if you have two identical interfaces merged together as an aggregate device, you can expect them to be in perfect sync. Otherwise, maybe, maybe not, it's not guaranteed.

Reply

Answer 3

jord

Level 5

5,230 points

Mar 14, 2006 8:07 AM in response to tele_player

This is true to a point. The devices may need to be wired up physically to a word clock source to be in sync. This is often done using a BNC cable. Appointing one as a master clock in AMS is not enough.

jord

Reply

Answer 4

Mar 14, 2006 8:28 AM in response to jord

No, even that won't do it if the conveter chips themselves have different internal latencies.

Reply

Answer 5

jord

Level 5

5,230 points

Mar 14, 2006 9:01 AM in response to tele_player

Considering that I have mine wired up as such, I can't agree with you. Not to mention that the original poster didn't seem to be complaining about the latency of a single device. This sounds a lot like a sync situation.

jord

Reply

Answer 6

Mar 14, 2006 10:42 AM in response to jord

hey guys, thanks for your responses.

what's MTP/AV? it's generating the clock signal for both your projectmix and motu?

yup, i set the 1082's clock source as internal in the tascam driver manager control panel, and also in AMS (audio midi setup).

The firepod is set for clock external in AMS (there is a choice to either be internal or mac), where mac is external i presume.

Unfortunately, neither the Tascam 1082 nor the Presonus firepod have BNC wordclock (though the tascam 1804 rack mount does).

I thought about also doing the following...do you think it will work? The tascam can get it's clock from either internal or SPDIF in. So let's say i set it to SPDIF in. Then perhaps the firepod can be the clock source and i can use it's SPDIF out to go into the 1082's SPDIF in?

anybody sync firewire devices with SPDIF?

Reply

Answer 7

Mar 14, 2006 1:16 PM in response to jord

jord - you're misunderstanding me. Re-read it. It may or may not work correctly. Your combination does, other combinations may not.

Reply

Answer 8

Mar 14, 2006 1:26 PM in response to tele_player

I've never heard of this before. What's your source on this?

Reply

Answer 9

Mar 14, 2006 2:02 PM in response to Mike Connelly

Go read data sheets on some converter chips, and you'll see that they each have a latency. I'm pretty sure this latency isn't handled in CoreAudio.

Do a Google search for 'converter latency', and read the data sheets for some A/D converters.

Reply

Answer 10

Mar 14, 2006 2:10 PM in response to tele_player

It looks like you can have a difference in latency, but generally only in the range of 30-50 samples, which is about 1.5 milliseconds.

So you could have latency, but it likely wouldn't be audible. You wouldn't get clocking issues if you sync clock via coax. Sounds like there isn't anything worth worrying about, am I missing something?

Reply

Answer 11

Mar 14, 2006 2:44 PM in response to Mike Connelly

The point I mentioned is really only an issue where the same signal passes through two converters with different latency, such as the bleed on a multi-mic'ed drum set. But you're already going to have similar issues just from the distances involved - hence, it's a small point.

I brought it up since it didn't seem incorrect clocking would cause the latency he observed.

Reply

Answer 12

Mar 14, 2006 2:47 PM in response to tele_player

do you guys aggregate using AMS as well? how have your experiences been?

Reply

Answer 13

Mar 14, 2006 3:33 PM in response to SpartanDelicious

I've done it, but only to see that it worked.

Reply

Answer 14

Mar 14, 2006 3:40 PM in response to SpartanDelicious

I don't use the Aggregate device function, but I'm pretty sure the two units DO have to be connected by clock. Obviously BNC is the best way to do this, but yes, S/PDIF will work.

Reply

Answer 15

Mar 14, 2006 4:10 PM in response to Mark the Williams

Yes, they need to be clocked together, onw way or another.

Reply