Data cabling

[JoeESP9]

Gordon Rankin of Wavelength Audio received a patent for asynchronous data transfer via a USB connection. Asynchronous can be and usually is better, as the transmission is re-clocked by the receiving device. All the better USB DAC's operate asynchronously.

 

[cpt_paranoia]

Gordon Rankin of Wavelength Audio received a patent for asynchronous data transfer via a USB

You wouldn't have a patent # or reference for that, by any chance? I'd like to see if there's anything actually novel in it, rather than it simply being an application of a long-used method of data conversion.

 

[JoeESP9]

If you really want to know check the Wavelength Audio website. I don't think an application of a long used method is patent worthy.

You can also check the link below.

https://www.audiostream.com/content/qa-wavelength-audios-gordon-rankin

 

[cpt_paranoia]

You can also check the link below.

Thanks. It looks like it was around 2002.

Unless there's something unusual about the Streamlength protocol, then software radio systems have been using data pull mechanisms for decades. The modulator uses a DAC, creating the modulation, with a set of cascaded FIFO buffers to deal with the hardware to software interface. Messaging between the hardware FIFO control logic and the software cause blocks of samples to be sent to the FIFO. Provided the software responds to these requests in a timely manner, the FIFO does not overflow or underflow, and all is well, and the DAC is timing master, and the source is the slave. The communications mechanism by which the request message is passed to the source is a mere detail.

I can only think that the SPDIF output from CD players dragged people down the wrong path (source clock master). But, internally to the CD player, the DAC is the clock master, and the drive is a slave (slaved to the DAC clock by the servo loops)...

ps. This is not a criticism of Gordon Rankin; it's a criticism of the patent system. He just reinvented the correct solution.

 

[wyn palmer]

Thanks. It looks like it was around 2002.

Unless there's something unusual about the Streamlength protocol, then software radio systems have been using data pull mechanisms for decades. The modulator uses a DAC, creating the modulation, with a set of cascaded FIFO buffers to deal with the hardware to software interface. Messaging between the hardware FIFO control logic and the software cause blocks of samples to be sent to the FIFO. Provided the software responds to these requests in a timely manner, the FIFO does not overflow or underflow, and all is well, and the DAC is timing master, and the source is the slave. The communications mechanism by which the request message is passed to the source is a mere detail.

I can only think that the SPDIF output from CD players dragged people down the wrong path (source clock master). But, internally to the CD player, the DAC is the clock master, and the drive is a slave (slaved to the DAC clock by the servo loops)...

ps. This is not a criticism of Gordon Rankin; it's a criticism of the patent system. He just reinvented the correct solution.

Isn't the Audio USB interface described exactly in the Silabs isochronous white paper that I referenced earlier.
I believe, and please correct me if I'm wrong, that this is the de facto standard that Audio USB interfaces use and is, in essence the asynchronous mode that is used to support real-time audio data transfers.
And yes, it is similar to that employed in handset/base-station data transfers since the days of GSM.

 

[wyn palmer]

Isn't the Audio USB interface described exactly in the Silabs isochronous white paper that I referenced earlier.
I believe, and please correct me if I'm wrong, that this is the de facto standard that Audio USB interfaces use and is, in essence the asynchronous mode that is used to support real-time audio data transfers.

http://www.farnell.com/datasheets/1738852.pdf
This is a full USB audio class 2.0 design guide in which all of the requirements are defined, including the need for the reference/retiming clock. It's a pretty complex system.