New low-cost DSP platform in development

[voodooless]

AES/EBU?

That would not make a difference; it's the exact same protocol, just different cables and a bit more hardware needed in front of the receiver/transmitter chips.

 

[Esprit]

I don't think it's the same protocol (but I'm no expert), starting with the different impedance.
There is a lack of devices with AES/EBU output (obviously low cost).
Good day.

 

[voodooless]

I don't think it's the same protocol (but I'm no expert), starting with the different impedance.

Impedance doesn't have anything to do with a binary wire protocol. They are the same for AES, Toslink, and coax. The difference is only the physical connections. All fall under IEC 60958.

 

[somebodyelse]

Is CEC itself or the lack of it is the cause of these compatibility problems? I found an implementation of it for Arduino, so doing the same on the Duo board shouldn't be too hard/expensive. https://github.com/floe/CEC

I have limited knowledge having looked into it a bit out of interest, but never having tried implementing it. A white paper from Lattice about eARC says that one of the problems eARC was intended to fix was interoperability with ARC. IIRC it put this down to vendor-specific variations in CEC behaviour. From memory of a quick read of an HDMI spec version found on archive.org the CEC part is mandatory in ARC both for capability discovery and for flow control (start/stop of audio stream etc.) I got the impression that if the discovery fails, or there isn't a start of stream, the devices are supposed to not work. I wouldn't be surprised if some implementations are a bit more flexible about this though. Looking back it would have been useful to take notes, but I didn't.

Nice find on the arduino code. It's for an older HDMI version but should at least be a good base.

 

[somebodyelse]

Impedance doesn't have anything to do with a binary wire protocol. They are the same for AES, Toslink, and coax. The difference is only the physical connections. All fall under IEC 60958.

It's similar to TCP/IP remaining the same irrespective of the physical layer which could be wired, wireless, optical or even virtual. And each of those broad types has multiple variations.

 

[Esprit]

The difference is only the physical connections. All fall under IEC 60958.

From web: […] AES/EBU uses ten times the signal level of SPDIF, meaning it can tolerate more loss in long cables and if a receiver circuit has been optimised for AES/EBU it will work better as SPDIF will be at the lower limits (too much noise). […]

 

[EJ3]

We are locked and loaded for some seriously low-cost R&D!


Hoping that this will only be true for the cheapest/simplest option but it looks like adding a passable DAC for 2 more channels would only cost a dollar more.

At least a pair of SUB out DSP? (been doing 2 subs since the late 1970s) one way or another.

 

[voodooless]

From web: […] AES/EBU uses ten times the signal level of SPDIF, meaning it can tolerate more loss in long cables and if a receiver circuit has been optimised for AES/EBU it will work better as SPDIF will be at the lower limits (too much noise). […]

Yes, that's all physical...

 

[somebodyelse]

From web: […] AES/EBU uses ten times the signal level of SPDIF, meaning it can tolerate more loss in long cables and if a receiver circuit has been optimised for AES/EBU it will work better as SPDIF will be at the lower limits (too much noise). […]

That's a physical layer difference. Say I have a chip putting out an IEC 60958 compliant bit stream on a logic level pin. I could connect this output pin to one or more physical layer adapters to convert from logic level signal to AES/EBU on BNC, RCA or XLR, to toslink or to spdif coax, ARC (although that requires other stuff in parallel), Meridian's interface using network cables, or whatever proprietary thing I dream up. Even if they have the same connector they may have different voltage levels.

In the networking world they have standards to make changing the physical layer adapter easy. In the old days this was AUI with adapters for thick or thin coax, RJ45 for the predecessor of CAT5 etc. or optical fibre. Now we have SFP(+) for wired or optical adapters. So far as I know the audio world doesn't have an equivalent, although some manufacturers have proprietary interfaces to do something similar.

 

[Drengur]

I really like your enthusiasm and wish you good luck with this project. My obvious use case for such a device would be in between HDMI devices so I hope that can be implemented.

 

[TSB]

I really like your enthusiasm and wish you good luck with this project. My obvious use case for such a device would be in between HDMI devices so I hope that can be implemented.

Open source HDMI implementation: not a thing AFAIK

 

[somebodyelse]

Open source HDMI implementation: not a thing AFAIK

Not fully, but maybe close enough to implement something like this. There are SBCs available with HDMI input and output that run linux but need proprietary drivers for the HDMI input. That means you're stuck with their BSP kernel (usually android) rather than an upstream one, at least so far. I don't know how this interacts with HDCP though.

 

[voodooless]

I don't know how this interacts with HDCP though

It doesn’t. These boards all work only on non HDCP content as far as I know. Nevermind eARC support. There is also basically zero information to be found that might help to hack something yourself.

 

[bachatero]

I don't know if anybody else has looked, but you can get actual HDMI decoder chips from Analog Devices with real HDCP support for a solid $25. If one of those worked in a setup then it would be possible to have such an HDMI audio box but it would be a lot more expensive.

 

[somebodyelse]

It doesn’t. These boards all work only on non HDCP content as far as I know. Nevermind eARC support. There is also basically zero information to be found that might help to hack something yourself.

eARC in which direction? https://forum.khadas.com/t/amlogic-kernel-5-4-on-vim3l/15869/4 claims receiving eARC 7.1 LPCM from TV.

 

[bachatero]

At least a pair of SUB out DSP? (been doing 2 subs since the late 1970s) one way or another.

This could be included at near zero additional cost as the board I ordered includes an internal stereo ADC/DAC. Assuming they're not so great quality, dual sub outs is completely practical with how unimportant distortion/FR is to them.

 

[bachatero]

Progress update: It is now actually working on RISC-V and a handful of the core language features are implemented!

Next steps: Get more cases of these features working and get familiar with the cheap audio equipment I ordered (and now received) from AliExpress.

 

[bachatero]

Didn't realize this until now, but since my candidate CPU board has a USB port, you will likely be able to also plug in whatever ADC/DAC you want in addition to what's built in.

 

[bachatero]

Almost ready for testing!

 

[voodooless]

eARC in which direction? https://forum.khadas.com/t/amlogic-kernel-5-4-on-vim3l/15869/4 claims receiving eARC 7.1 LPCM from TV.

Ah, someone managed to get it working. That’s cool! So there is some hope after all!

I don't know if anybody else has looked, but you can get actual HDMI decoder chips from Analog Devices with real HDCP support for a solid $25. If one of those worked in a setup then it would be possible to have such an HDMI audio box but it would be a lot more expensive.

The chips are not the problem with HDCP, getting a decryption key is.