D-1624 Lightpipe to PC Via RME

MiXit-G

New member
Sorry to interupt all the MR-8 threads but i have a question about lightpipe on my D-1624.

When i owned an adat and the adat/edit card (god bless the little thing...) i needed to use the sync cable and set the adat or the PC as master.

Now i'm just about to part with some cash for a RME digi96/8 pad but i need to know if i need the optional word clock module or if the 8 channel lightpipe does all the sync work without the need for another cable/connection?

And if cubase 5 is a program that will facilitate these transfers.

Also if i have 16 tracks of audio to transfer will tracks 9-16 sync up with the first 8?

I'm sure if i buy the card all these little mysteries will unfold but any reassurance or knowledge on the subject is much appreciated.
 
In short, yes. I have the D1624 with an RME Hammerfall and Cubase (both 5 and SX). You can clock the whole rig two different ways: one, use the Fostex as the master, and set the card to slave its clock to the ADAT inputs from the Fostex. This works quite well for flying in tracks. The Fostex is not as happy being the slave- it can occasionally hiccup in deriving clock from the ADAT inputs, in my experience. But letting it be the master and slaving the RME to it is rock solid. The implicit clock in the ADAT stream works just fine.

I've taken it one step further, and gone to step 2: I have a Lucid GenX6 as a house clock generator. I set up everything to sync to word clock in (Fostex, RME, everything else), and it truly is rock solid. But that's further than most will want (or need) to go.

All 16 tracks transfer in perfect sample-accurate sync with the Hammerfall (if you fly them in all at once, in other words with 2 ADAT cables to the RME)- I do this all the time. If you fly in groups of tracks at *different* times for some reason (like only having 8 channels of ADAT available on the digi96/8), you can get accuracy to the limit of MIDI timecode: which I don't really trust all that much, since it depends on the accuracy and consistency of the MIDI interface on the computer. So I avoid the problem outright and always just bring 'em all over at once.

I also have an RME ADI-8 standalone converter box. I use it for critical A/D work, because it really is very quiet and nice. But the cool thing comes at mixdown time: I can put the Fostex in input-monitor mode and run 16 tracks down to it from Cubase (using it just as a D/A converter box), and 8 more tracks to the ADI-8- and thus mix 24 tracks in the analog domain, which is the only way I like to do it...

That was one big benefit in getting the Hammerfall, with 24 channels of simultaneous ADAT I/O. You can use that in a lot of unexpected ways, once you get up and running... I don't know enough about the digi96/8 to really comment, but if I were you, I'd probably go for more channels of ADAT digital I/O, and skip the on-card A/D and D/A: you have 8 channels of A/D in the Fostex, and 16 channels of D/A. On-card converters (even RMEs!) are going to suffer in the noise floor department, simply because of the electrical environment- I hate the pitched wheezes that get stuffed into the bottom few bits.

I like the Midiman Flying Cow converter box that I drive with the SPDIF from the Hammerfall for monitoring _much_ better than I've liked any onboard sound card, simply because of the noise floor issues. I could do better, but that'll do for me for now...

IMNSHO, the Hammerfall is the perfect mate for the Fostex. Your call, though.
 
Thanks skipp thats what i needed to hear.

Ive thought about the hammerfall but the conveinience of analog in is to tempting because i have a mobile rig i do the bulk of my recording to it, then little overdubs i like to go straight in to cubase
Plus how can i monitor my mixes without going back through the fozzie?

With transferring 16 tracks via 8 channel lightpipe i thought as long as i start the transfer at the start of the song it should sync up ok??

Thanks again
 
Well, like I said, I use an external converter box (the Flying Cow) for monitoring. If you are mostly doing mobile work, are you rack-based? You'd have a place to put a half-rack deal like that...

My mobile rig is the 1624 and an Alesis Studio 32 for preamping/monitoring, and that goes straight into/out of the Fostex. I have no computer hardware at all in the road rack: location recording is not the time for the Blue Screen Of Death, in my opinion. When I get back to the studio, I just pull the disk carrier out of the road 1624, plug it into the studio unit, and fly the tracks down to Cubase via the Hammerfall. So my location monitoring takes place entirely in the analog domain, and my studio monitoring of Cubase stuff takes place via the Cow, which listens to the Hammerfall's SPDIF outputs.

I didn't want to compromise either the road or the studio units: the road unit has to be absolutely stone reliable as its primary characteristic, because the bulk of the location work I do does not offer the opportunity for a second chance. And the studio rig is set up for convenience, ease of use, and mostly _speed_ of use.

Like I said- I hate internal converters on soundcards. But you might not be able to get away from having them, in which case you'll have to compromise how many tracks you can fly down at once. Think very carefully about making that compromise, though: syncing up tracks perfectly is not as easy as simply pressing go at the start of the song. You'll need to develop the habit of putting slate/alignment tones at the start of takes, for example, so that you'll have a stable reference for checking the alignment...

To sync up multiple tracks flown in at different times, you'll need to make use of the MTC (MIDI timecode) out on the Fostex. At least it comes for free, so you don't have to worry about striping timecode and using up a track! You slave Cubase to the Fostex's timecode, and that will do the rough alignment for you to *subframe* accuracy: ideally 1/98 of a frame, or about 18 samples of uncertainty at 25fps. If you want absolutely dead-nuts perfect alignment down to the sample (which I suspect that you do), you'll then have to nudge the tracks manually the last few samples.

In the real world, that 18-sample inaccuracy is probably worse, because the MIDI hardware on the computer will probably not be perfect- there may be jitters or latencies in there that will let things move around a few more samples. Anyway, all that seemed like _way_ too much of a hassle for me to deal with: I'm basically lazy. So I just avoided the problem altogether by getting enough ADAT channels to start with.

Timecode was really intended for syncing sound to image (which can tolerate a lot of slop, really), not sample to sample. So I think that you may be making more work for yourself- but only you can make that call. Anyway, something like the Cow in a half-rack unit solves that problem.

Shoot, if you wanted to avoid having a separate box, you could even get something like an Audiophile 2496 and set it up inside the machine as a standalone A/D-D/A using its control panel (SPDIF to analog and vice versa, without talking to the PCI bus). Connect it to the Hammerfall's SPDIF I/O, and have it just do the conversion without actually talking to Cubase at all (since you can only have one flavor of soundcard, and that'd be the Hammerfall). I did this very thing at the beginning with my DAW: that worked fine too, but I just flat could not stand the digital noise down in my reverb tails- so out it went. That'd give you usable MIDI I/O as well, of course: Cubase _can_ talk to the MIDI on the Audiophile while routing the audio to the Hammerfall...

Your call, though. I'd check out the realities of syncing tracks before plunking the cash down: it may be less work than I think it is, and you might not mind it at all. Or it could drive you completely nuts and make you take up golf instead... (;-)
 
Back
Top