13-Oct-2017, 11:13
(This post was last modified: 13-Oct-2017, 11:16 by Jean-Marie.
Edit Reason: Fat fingers on my phone
)
I don’t own a Phantom either, so what follows is speculation based on reasoning:
I don’t believe that the latency is coming from the DSP treatment. What makes me stating so is that from a signal processing perspective, I don’t think that a phantom is doing anything substantially different from an expert with SAM turned on, and the latency in this case is imperceptible.
I believe that the latency in the phantom is coming from the fact that it is designed to be able to have multiple ones synchronized over the network, and if you want to be able to do it over a standard IP network, which is what a pair of phantoms + dialog are doing, you need bufffering to compensate packet delay variation, and therefore you have a latency in the 10s if not in the hundreds of milli-seconds.
Now I suspect that this buffering is occurring within the phantoms, even when a single phantom is used, because it is probably simpler to always have the same path, and that would be needed in the case of two phantoms.
If this is true, it should be possible for Devialet to have a special firmware bypassing this logic that enables networked phantoms and kind of having a direct path between the optical input and the DSP section. Such a phantom would have a similar latency as an expert.
I suspect that this is what they did in the context of that dongle.
My highly speculative two cent,
Jean-Marie
I don’t believe that the latency is coming from the DSP treatment. What makes me stating so is that from a signal processing perspective, I don’t think that a phantom is doing anything substantially different from an expert with SAM turned on, and the latency in this case is imperceptible.
I believe that the latency in the phantom is coming from the fact that it is designed to be able to have multiple ones synchronized over the network, and if you want to be able to do it over a standard IP network, which is what a pair of phantoms + dialog are doing, you need bufffering to compensate packet delay variation, and therefore you have a latency in the 10s if not in the hundreds of milli-seconds.
Now I suspect that this buffering is occurring within the phantoms, even when a single phantom is used, because it is probably simpler to always have the same path, and that would be needed in the case of two phantoms.
If this is true, it should be possible for Devialet to have a special firmware bypassing this logic that enables networked phantoms and kind of having a direct path between the optical input and the DSP section. Such a phantom would have a similar latency as an expert.
I suspect that this is what they did in the context of that dongle.
My highly speculative two cent,
Jean-Marie
MacBook Air M2 -> RAAT/Air -> WiFi -> PLC -> Ethernet -> Devialet 220pro with Core Infinity (upgraded from 120) -> AperturA Armonia
France
France
