Hello, I plan to expand my babyface interface with an ADAT expander to connect external hardware synths, drums machine and sequencers. Is there something I need to pay attention to when choosing the ADAT expander for everything to run tight with Cubase playback ?
Cubase always need to be the master clock I believe right ? what I am afraid of is having the analog drum machine not perfectly in sync with whatever is already sequencer in Cubase, or in the hardware sequencer.
Is it best practice to first jam with the hardware only, and use Cubase then only to record everything instead of mixing playback from the gear+the daw at the same time ?
I believe most recent ADAT expanders will have a very limited impact on the overall latency. As for the clock, I would use the RME as the master clock and sync both devices through the ADAT connection as, if my memory is correct, there is no BNC clock connector on the RME?
With the Babyface, current drivers and current computer, latency should be low enough for simultaneous playback and recording I’d say.
You’re talking about two different clocks for different purposes. One is the sample rate clock used to sync the sample rate of your ADAT device and your RME interface. The other is likely MIDI beat clock used to sync your drum machine to Cubase.
Sample rate clock sync should occur over the optical cable if the RME is master.
You have to be aware that the chosen ASIO Device can only report one latency. However, if the ADAT expander has AD/DA converters (which I assume is the whole point of adding it) then their latency will be most likely different from the latency of the RME converters.
This should result in slightly shifted positions from the ADAT recordings compared to the recordings done through the RME converters.
A few (dozend) samples most likely.
I use the Babyface pro fs with a Behringer ada8200. The bfp is the master and the ada is slave. To accomplish this you need two optical cables for in and out. As far as Cubase is concerned it is one device.
Yes, perfectly good for line-level signal such as from synths and drum machines. Connect via optical ADAT cables, set the sync on the unit to slave, and you see 8 more ins and outs in TotalMix at 44.1kHz or 48kHz, or 4 at 88.2kHz or 96kHz . It just works. I’ve been using the older model ADA8000 for years, just set it and forget it.
You’re right. The external converters can have a different latency than the ones on the “main” interface that will not be compensated for. The clock sync ensures no drifting takes place.
No offence but if the RME-Audio guys say one thing and you say another thing - I put my trust in the RME guys.
Basically their woe was that an ASIO device can only report a single latency value to the host. They would rather like to be able to report a latency value per port.
Audio clock does not affect this at all. It is a design flaw in the ASIO protocol.
I don’t know if you saw my reply just above, but I was basically saying that my initial comment was incorrect and you were right. The AD/DA latency of an auxiliary ADAT device will not be compensated for and that the sample clock is only to ensure that there will be no drift. In practice it means that there is likely a small discrepancy between the internally connected I/O and those of an external AD/DA. But it will be constant.
Ah, okay. Now I understand you post better. Thanks for the clarification.
It is actually the oversampling and filtering around the converters that cause the latency. Just to drop a value for @bensmir.hbs 's guidance: My RME interface introduces 0.9ms latency on its built-in AD conversion at 48kHz sample rate. The interface knows this value and can report it to Cubase. However, it does not know the latency of the converters attached via ADAT. and even if it did it could not report it to Cubase as it is allowed to only report one value.
This issue also occurs if you don’t use external AD converters via ADAT but simply a digital connection (maybe a DAT machine or another computer). Since there is no latency induced by a converter that signal will be a bit too early in Cubase.
Synth 1 is connected to the ADAT expander via TRS jack cable, a first A/D conversion is happening in the ADAT Expander, with its own timing
Synth 2 doing the same on a different port, a different timing
then the ADAT expander send digital audio with through ADAT to the babyface with a single timing, which send audio to Cubase digitally, Cubase send it back, and then the babyface does a final D/A conversion with another timing
So 2 A/D conversion are happening right now? Which timing get compensated for exactly ? I am still trying to understand how the MIDI clock and the audio clock are interacting
Midi and audio clock do not interact at all.
The midi clock is basically a pulse (in reality a midi message) send from Cubase to your synth in regular intervals in regards to musical time. Let’s pretend a pulse is send every 1/16ths note. If the bpm tempo in Cubase is raised more pulses per minute will be send, if the tempo is lowered less pulses will be send. Since every midi device knows how often per bar to expect such a pulse they can calculate the bpm. Thusly both devices can run at the same tempo.
The audio clock is used to synchronize the data stream of digital audio of two or more devices with each other. It is all about the sample rate. While to us 44100 sample per second seem to be a lot, but the amount of time during such a second where no sample is taken is much bigger.
The audio clock tells the slave device when a sample is taken, so that that device can take a sample at the same moment.
If you use a different audio interface for the AD conversion, yes. If you use the same device, just another input, then the conversion latency is the same.
The one that the RME engineers know about. The Babyface’s AD converter.
For every day use it is nice to know these things, in case you wonder why one audio recording is slightly offset to another audio recording e.g., but on many occasions you won’t really notice anything.
However, if you record a drumset and you use the Babyface’s AD and an expander’s AD you could create phase issues. In this case one set of audio recordings’ position must be manually adjusted by a few samples. It’s no the end of the world, it just requires a manual correction.
The Oxford drum gate 2 will phase align if you do go across a different adat interface to the host interface. May not be of any use to you but maybe for others.