Following a discussion about E-Mu and its unsure future (Emu ASIO input - Cubase - Steinberg Forums), I was wondering if a RME RayDat could be a solution for replacing in a more or less far future what I already have (see my sig) and if so, how. As the subject is hardware related, it was suggested that I should start a topic about this here, so…
Considering my needs (10 line inputs, 2 mics and 2 Hi-Z inputs, I would probably have to purchase one more ADAT unit to have the necessary connections. Thus the question : how to sync properly 3 units (Raydat and 2 preamp/lines devices together via TOSLink, and which unit would be better as master ?
Basically, keep the RayDAT as master, connect the ADA8000 ADAT in and out optically to a RayDAT ADAT pair and set it to external (slave) sync to ADAT and it will just work. I’ve done this with multiple devices and mixed with MOTU devices (828MkII’s). If a preamp has an S/PDIF or AES-EBU input it should also be possible to sync it to the RayDAT … for example, if you retained the EMU 1616M on a separate computer you could send either ADAT or S/PDIF to it and set it to slave to the RayDAT. Don’t forget also that you can use the RayDAT’s S/PDIF and AES as two separate and independent stereo channels (2 stereo ins and 2 stereo outs), so you could have a 2-channel preamp coming in on S/PDIF, a stereo monitor amp on the S/PDIF output, another 2-channel preamp coming in on AES (XLR) and finally a stereo headphone channel going out on AES. After that, you still have 32 in and 32 outs available via ADAT
The main problems are when you want to connect a preamp that only has an S/PDIF output … this is common with many guitar effects and pedals. There’s no way to get these to sync to another device; the only option there is to sync the RayDAT to the device, in which case you lose out on the quality of the clock from the RayDAT. One notable exception in the guitar world for example is the Behringer V-Amp Pro (the 19" rack version) which has a word clock input – which brings me neatly to word clock.
If the budget allows, I’d strongly recommend getting the WCM along with the RayDAT as this adds word clock in and out, however it could always be added later. I’m still kicking myself for not getting the EMU word clock module when I bought my EMU 1212M years ago, while they were still on the market!
In a nutshell, as far as sync in concerned, with the RayDAT as master, what you need to be able to do is send an output to the input of each device that you want to operate in sync – and each of those devices must allow itself to be set to external sync as well! A device which has no digital inputs at all cannot be slave-synced to anything.
I’ve had a look at your thread over on Gearslutz and it’s all pretty good advice as far as I can see, although I wouldn’t worry too much about the recommendation to use the clock in the converters as opposed to the RayDAT (unless of course, you have very, very good converters!). Why that advice is given (and this applies to A/D conversion only, i.e. recording) is explained by Hugh Robjohns in an SOS article. Note however that if you use word clock to the converter (ADA8000 in your case), this “interface jitter” won’t happen; another argument for word clock. Hugh writes a more in-depth article on the subject here and if you want some real food for thought then have a look at his article on whether or not anyone benefits from a dedicated digital master clock at all!
Finally, don’t forget that RME’s TotalMix is simply awesome (not a word I use lightly, as is the fashion!). Think PatchMIX on steroids.
Well, considering the jitter problem, things are as complicated as I was expecting. if I understand well the SOS articles, it’s better to use the internal clock of an A/D converter to avoid jitter issues. Problem is : what happend if you have 2 of them connected to Raydat via TOSLink ? At this point, it seems that the word clock is indeed the solution to keep the Raydat as master without problem, which means, as you said, purchasing the WCM (100 €).
There is a thing that I don’t get, though : let’s say we set the Raydat as master and the ADA8000 as slave. If a TOSLink cable (a good quality one) is connected from Raydat (out) to the ADA8000 (in) there is no digital audio transfer on it, only a clock signal : how can jittering arise in this case, as there is no blending of audio and clock signals ?
Didn’t have time to read the last article, yet. I’ll post questions if I have some more after reading.
Anyway, a great thank for your help and feel free to correct me if I am wrong about this.
It’s important to remember that there is only one optical cable each way so there is only one signal consisting of a mixture of the clock and data signals; an ADAT Lightpipe receiver must recover the clock signal if it is to be able to sync to it. Now, even when pure digital silence is transmitted (all zeros), this will not be a square wave; and as soon as any audio signal arrives, this will appear to be an almost random signal to the untrained eye. Nevertheless, given a perfect, theoretically clean input signal, it would still be possible to recover a jitter-free clock. The problem is that this must first travel through the physical medium of the fibre optic cable, so when it comes out the other end the rising and falling edges are smeared and this is where the jitter comes from.
In the case of word clock, it’s always a square wave, and although it too can be degraded when passing through a cable, the nature of the degradation does not change and therefore the period (“distance”) between transitions will not vary, making it possible to recover a very clean and stable clock at the receiver.
Here’s another Hugh Robjohns article, “Digital interfacing”
Check out the section on jitter, particularly the first diagram:
There’s some more science about this type of encoding here.
After all that, IMHO unless your requirement is extremely critical, for example recording multiple mic channels on cymbals where phase alignment is important, the difference is unlikely to be noticeable. If you just want to record e.g. several analogue synths and a few electric guitars simultaneously, having the ADA8000 sync via ADAT will be just fine. You can always start out that way, do a few tests and make a few measurements, and if there is some clear evidence that jitter may be a problem, then add the WCM later.
If you do decide to go with word clock, I’d recommend a T-piece and 75-ohm terminator at each end of the cable.
N.B. the WCM word clock input is high-impedance, so if you are ever sending a word clock signal into the WCM (which I understand you are not proposing to do), you just push in the switch on the WCM and take the feed for the next device from a WCM output – in this case will not need need a T-piece and a terminator at the WCM input and the WCM acts as a word clock “refresher” – the regenerated output word clock is of higher quality.
Correction: what I meant to say was “at the end of each cable” as you’re using both WCM outputs in Picture2, one to each device. At the device end, you put a T-piece and a 75-ohm terminator, unless the specific device has internal termination – the ADA8000 doesn’t, so you will need it there. You don’t fit a T-piece or terminator at the WCM outputs in that case.
One other thing comes to mind: I spent a lot of time looking for 75-ohm cables and terminators until it dawned on me that this stuff actually originated in the video world, so ask some video guys and they’ll get you sorted. Don’t be tempted to use the type of cable that was common in old computer networks … the centre pin is too big and will damage the WCM connectors, as well as being the wrong impedance. The correct cable is RG59 75-ohm.
