dai (digital audio interface)

George,

USB audio devices typically derive their clock from the PC. Either synchronous to SOF, our implicitly via the playback data being streamed. A digital audio interface as a slave to an external clock implies that you are providing a clock to the host PC. Assuming you are talking about a full-duplex device (playback and recording) this is not trivial.

Microsoft did not have proper support for this before Millennium/ XP. Apple did not support this properly before MacOSX 10.4.2. But even if these operating systems are available for a few years now: This is a continuous cause of trouble. We have gone through all of this a few years ago- educating Microsoft and Apple how to implement the USB spec properly. You might remember our hardware:

There used to be a couple of chips from Philips Semiconductors implementing I2S to USB or Analog to USB. In case you can still find them (I believe they are EOL), you might add a SPDIF receiver and be done. But look closely at the spec: Check if this is really working

With the EMI 2|6 we used a Cypress EZusb and implemented the complete audio data path in a Xilinx FPGA. I am not sure, if your application can afford this. If cost is not the issue, this will lead to a very flexible solution.

Hope this helps, best regards

Felix

George,

USB audio devices typically derive their clock from the PC. Either synchronous to SOF, our implicitly via the playback data being streamed. A digital audio interface as a slave to an external clock implies that you are providing a clock to the host PC. Assuming you are talking about a full-duplex device (playback and recording) this is not trivial.

Microsoft did not have proper support for this before Millennium/ XP. Apple did not support this properly before MacOSX 10.4.2. But even if these operating systems are available for a few years now: This is a continuous cause of trouble. We have gone through all of this a few years ago- educating Microsoft and Apple how to implement the USB spec properly. You might remember our hardware:

There used to be a couple of chips from Philips Semiconductors implementing I2S to USB or Analog to USB. In case you can still find them (I believe they are EOL), you might add a SPDIF receiver and be done. But look closely at the spec: Check if this is really working

With the EMI 2|6 we used a Cypress EZusb and implemented the complete audio data path in a Xilinx FPGA. I am not sure, if your application can afford this. If cost is not the issue, this will lead to a very flexible solution.

Hope this helps, best regards

Felix

George,

USB audio devices typically derive their clock from the PC. Either synchronous to SOF, our implicitly via the playback data being streamed. A digital audio interface as a slave to an external clock implies that you are providing a clock to the host PC. Assuming you are talking about a full-duplex device (playback and recording) this is not trivial.

Microsoft did not have proper support for this before Millennium/ XP. Apple did not support this properly before MacOSX 10.4.2. But even if these operating systems are available for a few years now: This is a continuous cause of trouble. We have gone through all of this a few years ago- educating Microsoft and Apple how to implement the USB spec properly. You might remember our hardware:

There used to be a couple of chips from Philips Semiconductors implementing I2S to USB or Analog to USB. In case you can still find them (I believe they are EOL), you might add a SPDIF receiver and be done. But look closely at the spec: Check if this is really working

With the EMI 2|6 we used a Cypress EZusb and implemented the complete audio data path in a Xilinx FPGA. I am not sure, if your application can afford this. If cost is not the issue, this will lead to a very flexible solution.

Hope this helps, best regards

Felix