Does this interface exist ? 8 SPDIF in 1 ADAT out (containing 8 channels)

[bluzee]

If you use Linux you can forget Dante. Alternatives are Ravenna which is an open standard, and AVB, also an open standard.

Still networked audio doesn't solve the problem of getting a sync signal into a device.

Say we have a synth that has an SPDIF output only. It has to be the master clock as there is no input to feed it a clock from elsewhere. If this synth had an SPDIF input and output it could be set to slave mode and take its clock from the SPDIF input. So adding networked audio doesn't solve the problem of a device having only an output.

I'm not sure if that matters. The stream gets reclocked to the network master in the converter. None of the converters I've seen for SPDIF have an SPDIF output. They are one direction. SPDIF in to Dante.

Regardless, not something I would do. Ridiculously expensive stuff. The Z Systems unit might work if you can find one, but it still seems like a flaky idea. If you can't get good audio from a keyboard using a balanced connection something is not right.

[merlyn]

The Digiface may still work, I haven't found info yet which explicitly rules it out.

How are you proposing sync would work? If there are four synths and an interface, that's five clocks to sync.

[bluzee]

It won't work. He needs 4 in and one ADAT out.

[asbak]

The Digiface may still work, I haven't found info yet which explicitly rules it out.

How are you proposing sync would work? If there are four synths and an interface, that's five clocks to sync.

That's a question for RME support, but why couldn't it work the same or similar as on the Z-series which apparently solves that problem?

[tavasti]

The Digiface may still work, I haven't found info yet which explicitly rules it out.

How are you proposing sync would work? If there are four synths and an interface, that's five clocks to sync.

That is really good question. Would it need to do resampling for all inputs?

[merlyn]

Would it need to do resampling for all inputs?

That would work and I can't think of any other way to do it.

... why couldn't it work the same or similar as on the Z-series which apparently solves that problem?

You've maybe missed the significance of sample rate conversion. Re-sampling from 44.1kHz to 44.1kHz doesn't change the sample rate but it does sync the streams.

The basic idea is that in a digital audio system there can only be one sample clock. Pro setups use wordclock, which distributes one clock between multiple bits of gear. SPDIF and ADAT have the clock embedded in the signal. When SPDIF and ADAT send silence the way they're coded mean the signal is not a stream of zeros and the clock can be extracted.

If you have a 9652 you will have encountered this. What is plugged into your 9652? Say it's a Behringer ADA8200. The input and output of the 8200 are connected to the 9652. One device has to be the master device, say the 9652. The 8200 is set to slave and takes its clock from the ADAT input. All good. More 8200 units can be connected and synced in the same way -- the whole system is running from the clock in the 9652.

Now take a device with no digital input. It has to be the master. Add another device without a digital input and there are two masters -- this can't be done.

How do we sync multiple devices that don't have a digital input? Typical of engineers the answer is "we don't". It's impossible. So the solution is to use one device as the master clock and resample all other streams to that clock.

This is also the way alsa-in works to allow multiple interfaces to be used at the same time with JACK.

[asbak]

It's still as clear as muck to me.

So let's assume the OP has 4x synths all set at 48kHz and he plugs their electrical SPDIFS into a Z-Systems Input #8.
According to the Z-Systems manual it does Sample Rate Conversion.

I make the assumption that what they're getting at (unfortunately I don't have enough gear or a Z-Systems to put this to the test) is that different sources (keyboards, ADAT machines, whatever) are all producing output that isn't synchronized, and obviously they cannot possibly be.

So to solve this, the Z-series apparently sends an ADAT optical signal to the SPDIF destinations to use a clock reference.
The manual is vague and doesn't go into any detail about what it expects you to connect to what for this to possibly work.

The best guess I can come up with based on limited and vague info is that on the Z-Series one would have to send keyboard SPDIF out to the the SPDIF electrical inputs on the Z-Series. It doesn't explicitly mention anything in the manual about connecting the Z-Series SPDIF outputs to the keyboard SPDIF inputs. Maybe they assume the end user will do this, maybe it's not necessary, whatever the case this part is unclear and the manual isn't great.

Assuming then that this is the only way it can be made to work then well ... that's the end of it.

If however it does work somehow without hooking up the keyboard SPDIF inputs to the Z-Series SPDIF outputs things get interesting.

[merlyn]

... So to solve this, the Z-series apparently sends an ADAT optical signal to the SPDIF destinations to use a clock reference.

I think this is where things are becoming as clear as muck. That is not what is happening. Any attempt to do anything with the clock on devices attached to the z-systems box has been abandoned. The clock on the e.g. synth is left to drift out of sync and resampling brings it back into sync. Does this help?

[asbak]

Muchos gracias for the diagrams merlyn, I'm going to have to sleep and read a lot more on this one before the lights go on.
It was so much simpler when I just rammed ADAT into ADAT and clocked it all off a MTP AV.