Remote with cascaded Metric Halo ULN-8

[Adebar]

Is there anyone using more than one ULN-8 for recording live concerts?

I would like to go with 32 channels. For that 2 solutions together with a MacBookPro are on my list:
Apogee Symphony mobile with 2 AD16X or
4 Metric Halo ULN-8 on Firewire.

In terms of stability I would trust the Symphony - in theory more reliable than cascaded Firewire interfaces. On the other side I like the converter and also the pres of the MH.

Would you consider the Metric Halo with 4 interfaces on a fw hub as relaible for live recording?

[rumleymusic]

Just by the limitations of firewire, you would be stuck with 48kHz max on the Metric halo if you were to chain 4 of them together.

Is it even possible to use a firewire hub for 4 separate interfaces? I can't imagine that being a good idea even if it were possible.

[Adebar]

Originally Posted by rumleymusic

Just by the limitations of firewire, you would be stuck with 48kHz max on the Metric halo if you were to chain 4 of them together.

How did you calculate that? I thought 32 channels with 24/96 shouldn´t be the limit for firewire.

[Allen Rowand]

Originally Posted by Adebar

Is there anyone using more than one ULN-8 for recording live concerts?

I would like to go with 32 channels. For that 2 solutions together with a MacBookPro are on my list:
Apogee Symphony mobile with 2 AD16X or
4 Metric Halo ULN-8 on Firewire.

In terms of stability I would trust the Symphony - in theory more reliable than cascaded Firewire interfaces. On the other side I like the converter and also the pres of the MH.

Would you consider the Metric Halo with 4 interfaces on a fw hub as relaible for live recording?

There's no need to fear FW; one of our beta testers has been tracking Neil Young with 9(!) ULN-8s and a Mac Pro for a few years.

For your situation, the key is to daisychain the interfaces. Take the AES output of box 2 into box 1, and box 4 into box 3. You record the the 8 analog and 8 digital channels from boxes 1 & 3. This means that you're only using the CoreAudio transport of two boxes instead of 4. You will still have full "online" control of all 4 boxes. This should work at 44.1-96kHz; the ULN-8 only transports 8 channels at 176.4/192kHz.

Best,
Allen

[Dave Davis]

I've definitely done 3 boxes (24 channels) with no hiccups with ULN8s. One such recording even got a Grammy nomination.

As to hubs, actually that's the DESIRED way to hook up multibox rigs, as opposed to daisy chaining. Seems counterintuitive, but that's the recommended approach. One obvious benefit is it protects the host mac and attached uln8s from one another when things go bad-wrong. A powerbook with a dead FW port is a doorstop, and direct connections are one shortcut to killing your FW card (which is itself a hub!).

I'm not sure what the limits and sample rates are, but I'm pretty sure I've done 3 box 96K recordings, and definitely 2 boxes at 192. Some variables are the number of DAW and FW return channels you're using. In practice I hang the ULN8s off the built in FW port, and put my drives on an eSata Express Card. I've found running drives on the same port as the ULN8s to be a limiting factor for high res, so I've always used additional ports (before my MBP/eSata solution, I used a PCMCIA FW card).

Lemme see... 2 channels @ 48K ~ 1.5mbits/sec. At 96K you get 1 channel at that rate. 32 channels x 1.5 = 48mbits/sec. 4 ULN8s have potentially 64 inputs and more than 64 outputs using all digital and analog i/o., which would consume roughly 200 mbits/sec, in other words STILL under FW400's 400mbit speed limit with room to spare at 96K. Given that, I don't see it a big stretch to handle 32 channels at 192 (as per Allen's note you'd need to use 3 individual boxes due to 8 channel limit at 192). Regardless, the key is to use a different (preferably FASTER) bus for drives. When using eSata, there's enough bandwidth left over to let Console write captures/recordings to 2 separate drives at the same time.

-d-

[rumleymusic]

How did you calculate that? I thought 32 channels with 24/96 shouldn´t be the limit for firewire.

There are theoretical limits of these data transfer mediums and there are real limits. But the protocol doesn't always conform to that. You are usually lucky to get half of that, and if you use a hub I doubt you would get a quarter of that.

It is not impossible, and a mac book pro will probably be the only laptop that could do it. My concern would be stability. But the guys a MH would know. It is their interface.

[sonare]

FWIW I used 3 ULN-8s (44.1) for a Houston Symphony CD without a hiccup. Used the AES bus to connect the three, and then a FW800 optical repeater to carry the FW data over LCLC fiber to the control room. Granted, at 44.1 it was not breaking a sweat with 24 channels, but the important point is that I saw absolutely no problems!

Rich

[Adebar]

Originally Posted by Allen Rowand

For your situation, the key is to daisychain the interfaces. Take the AES output of box 2 into box 1, and box 4 into box 3. You record the the 8 analog and 8 digital channels from boxes 1 & 3. This means that you're only using the CoreAudio transport of two boxes instead of 4. You will still have full "online" control of all 4 boxes. This should work at 44.1-96kHz; the ULN-8 only transports 8 channels at 176.4/192kHz.

Best,
Allen

Thanks Allen,

Isn´t it easier for the setup to just use the CoreAudio transport of all single boxes witout connecting AES out to box 2 and 3 into 4?

Would 4 boxrs at 192k still run without problems on the FW bus?

Somewhere I read instead of daisy chaining the 4 boxes it ould be better to use a FW hub. The new MacBook Pros only have FW800. Is there something to consider with FW800?

Roland

[Allen Rowand]

Originally Posted by Adebar

Thanks Allen,

Isn´t it easier for the setup to just use the CoreAudio transport of all single boxes witout connecting AES out to box 2 and 3 into 4?

Would 4 boxrs at 192k still run without problems on the FW bus?

Somewhere I read instead of daisy chaining the 4 boxes it ould be better to use a FW hub. The new MacBook Pros only have FW800. Is there something to consider with FW800?

Roland

Hi Roland,
Due to the way Core Audio works, every box that is transporting audio is always allocating 20 channels to and from the computer. So, 4 boxes actively transporting would eat 80 channels of bandwidth. By cascading two boxes into the other two boxes you have 32 channels of audio occupying 40 channels of bandwidth instead of 80. This lightens the load significantly.

I'll have to get back to you about 4 boxes at 192k. It might require a second FW bus via ExpressCard.

FW800 vs. 400 makes no difference. I daisy chain my boxes currently, but am looking for a hub to mount in my rack at some point.

Best,
Allen

[Adebar]

Originally Posted by Allen Rowand

Hi Roland,
Due to the way Core Audio works, every box that is transporting audio is always allocating 20 channels to and from the computer. So, 4 boxes actively transporting would eat 80 channels of bandwidth. By cascading two boxes into the other two boxes you have 32 channels of audio occupying 40 channels of bandwidth instead of 80. This lightens the load significantly.

Thanks for that clarification, didn´t consider that.

I'll have to get back to you about 4 boxes at 192k. It might require a second FW bus via ExpressCard.

My idea was to connect the audio interfaces on the firewire bus and the external harddisc on the ExpressCard slot, assuming it is better not to have an external HD drive on the same bus as the audio interfaces.
Does that make sense or is that not a good idea?

[DanRock101]

Wow! Good to know! Thanks!

Originally Posted by Allen Rowand

Hi Roland,
Due to the way Core Audio works, every box that is transporting audio is always allocating 20 channels to and from the computer. So, 4 boxes actively transporting would eat 80 channels of bandwidth. By cascading two boxes into the other two boxes you have 32 channels of audio occupying 40 channels of bandwidth instead of 80. This lightens the load significantly.

I'll have to get back to you about 4 boxes at 192k. It might require a second FW bus via ExpressCard.

FW800 vs. 400 makes no difference. I daisy chain my boxes currently, but am looking for a hub to mount in my rack at some point.

Best,
Allen

[Musician]

I often got the remark: "why on earth would you want to record at 192 khz ??"

So could you guys pls enlighten me ?
Also, reading above posts, most of you still dont higher than 96 khz.
Is then the ULN8 still that much better in sound ?

Additional questions:
- what FW repeater allows to use optical transport for long distances ?
- what device can be used for using cat5 cable for long distances ?
- what FW hub is recommended ? Since there are the asian cheaper ones and such...

Thanks!

[ghellquist]

Originally Posted by Musician

I often got the remark: "why on earth would you want to record at 192 khz ??"

Very good question.

I have tried different sample rates. Now I always record at 44.1 kHz.
The reason, for me with my sound card, have tested, is that the difference is small or even tiny when going at higher speeds.

I preferr spending the energy on getting mics placed just right ( well, I do try, not saying I succeed ). Moving the main pair mics a foot makes more difference that going from 44.1 to 96 in my experience.

Please do a proper test and validate with your equipment.

Your mileage may vary.

[Allen Rowand]

Sorry to take so long to get back to this thread!

Here's the deal, with the current driver:
To do 4 boxes @192k you need to have 2 FW busses, with 2 boxes on each. You should be able to use a FW800 disk on one of the busses.

For 96k or lower, you can digitally cascade boxes together so that you have 2 boxes on one bus (that also get signal from the other 2 boxes via AES), and have your record drive on the other FW bus.

If you need to do 32 channels at 192k regularly, please get in touch with support- we may be able to build you a custom driver that would better fit your needs.

Best,
Allen

[HughH]

Hello Allen,
Thanks for your input in keeping this thread informative. One suggestion for a driver update - the RME boxes can limit the bandwidth to only use certain channels of the box, handy for multi-box configs. Just a thought.
H

[KBOY]

You would do a custom driver for someone? That's pretty amazing, and a mindset that truly warms my soul.

I am not anywhere near buying anything. But I am really contemplating this box. The people behind it are sealing the deal. Seems like great people, making great products.

I mean, what exactly is the downside?

[HFTobeason]

We are researching the purchase of 2 ULN-8s and 2 LIO-8s to be used in a live performance situation. We hope that the abundant experience and prowess of the GearSlutz community may be able to answer some key questions as follows:

We have two 2.4GHZ 17" MacBook Pros, each running Apple MainStage 2, plus a Mac Mini (can't remember the specs, but it's brand new) running Logic Pro 9 (although we don't think that Logic will come into play, as you will see).

Laptop "A" gets one stereo feed (Keyboard A) plus two mono feeds (Violin and Sequence A) from MainStage.

Laptop "B" gets one stereo feed (Keyboard B) plus one mono feed (Sequence B) from MainStage.

Laptop "A" also gets a single mono line input (Violin) which is routed into MainStage.

The band also generates 15 MIC level signals plus another 12 LINE level signals.

So -- in total, we've got 15 MIC signals going into the 2 ULN-8s and 13 (12 band + violin) LINE signals going into the 2 LIO-8s.

We plan on connecting the 2 LIO-8s to the 2 MBPs (each via direct FireWire) and the 2 ULN-8s to the Mac Mini (cascaded via FireWire and then through a Glyph FW800). Then, using the AES ports, running the LIO-8 outs into the ULN-8 ins.

Now, we've got the 6 LINE signals from LIO#1 plus the Keyboard A stereo pair going out via AES into ULN#1. ULN#1 also has 7 MIC signals from the band, plus (KEY QUESTION!) the LIO#1 analog out signal (post MainStage) of the processed violin.

Same setup, basically, for LIO#2 - 6 LINE signals plus Keyboard B stereo pair going AES into ULN#2. ULN#2 also has 8 MIC signals from the band.

The two Sequence outputs come out analog outs from LIO#1 and LIO#2 respectively and go to the stage mixer only.

So -- now we have 16 inputs into each ULN-8 (8 AES and 8 MIC into each). We assume (KEY QUESTION!) that the Record Panel will be able to record all 32 inputs onto the Mac Mini/Glyph. We assume (KEY QUESTION!) that we will be able to record at 48K at least - and perhaps at 96K.

We expect to be able to route the two keyboard stereo pairs, mixed down to a composite stereo pair, and the mono violin signal out of the analog outputs of ULN#1 which go to stage mix and FOH.

At the same time, we hope to be able to generate a stereo monitor mix from the 32 inputs for the keyboardist/violinist (and possibly others) using the v.5 Mixer and outputting that mix via ULN#2 analog outputs.

The attached diagram shows what we're planning. Any advice, corrections, help, contradictions, etc, would be very much appreciated before we spend almost $20K. Thank you.

[Allen Rowand]

The Record Panel can record all 32 channels from the ULN-8s at 96k.

The only issue is that you can't route audio directly between the two ULN-8s to create your Keys and IEM mixes. Your options are:
1) Use Logic on the Mini- create an aggregate device with the ULN-8s and do your mixing there. The downside is that this will add latency.
2) Free up 2 AES inputs on each ULN-8, so that you can route audio between the units over AES.

I don't understand what your setup is with the drums; perhaps there's a way to simplify them? Feel free to email me directly and I'll help you get this worked out.

Allen

[HFTobeason]

For anyone following this thread, here's an edited version of the advice I got from the extremely helpful folks at Metric Halo:

The Record Panel can record all 32 channels from the ULN-8s at 96k. The only issue is that you can't route audio directly between the two ULN-8s to create your Keys and IEM mixes.

I suppose I could split the Keys outputs (i.e. not mix them down) and output Keys A directly from LIOA and Keys B directly from LIOB, likewise output Violin direct from LIOA as well. I still need all those feeding into the ULNs, though, in order to record them, don't I?

Yes, you'll need the signals in the ULN-8 to record them to one computer.

Could I simply mult two sets of IEM mixes, one out of each ULN, using a "Y" cable? I assume I can still "mix" everything coming into each ULN in one v.5 Mixer interface, no?

Yes, your best bet might be to use a padded "Y" cable. It would have to be padded since you shouldn't directly connect two outputs this way. We can help you with the specifications; it's something that most techs would have no trouble building or you may be able to find one premade. And yes, you'll see both boxes in one mixer.

I don't understand what your setup is with the drums; perhaps there's a way to simplify them?

The issue is that we have dual drummers - we can't afford enough ULN boxes (nor the space to house them in our touring rig) to record all drum mics individually (one drummer uses 14 mics, the other 11). Thus, our plan is to take mixdowns of various mics from the stage mixer (as line level inputs) and record those - that's what the "Hats A L"/"Hats A R", etc, indicate. Since we're recording for potential release, we can't really combine the drum mixdowns any further - it would limit the subsequent mixing process too severely.

You didn't shoot down my plan to "passthru" the Violin output from LIOA to ULNB (line level to mic level) - can I assume that will work?

That is no problem; the mic inputs can handle line level signal.

I have attached a diagram of the final routing.

Again, thank you to Allen from MH for all the help.