When it comes to USB, I feel there's a lot of confusion with regard to the USB-to-SPDIF stage. Especially in today's world of UAC2, I feel that there's even more confusion because of the drivers needed for Windows, and the fact that most audio companies don't have someone who can write code for asynchronous transfer.

There are people that see 'XMOS' and say, "Cool!" (I used to do that). I know there are people here that dislike XMOS/Thesycon. I have no problem with it. Thus far, I've had fewer issues with it than with the C-Media drivers, even with the absolute generic Thesycon driver.

But there are a lot of USB receiver chips out there, from C-Media in the CM6631/6632A, which seems to provide their asynchronous drivers at a nominal cost as well. At the same time, I spoke with C-Media engineers at this past Computex, and they said that there's not much interest, except from ASUS. ASUS loves C-Media; they use their products almost all their stuff (the negotiation for prices probably works in their favor).

More and more, I get the feeling that the companies use the XMOS and CM663X because they don't want to bother writing code. The code works, unlike Tenor's code; the Tenor TE8802L was utter crap. Tenor engineers for some reason decided to completely disregard USB Audio 2.0 specs when writing their implementation for the driver, and it was buggy as all hell, causing massive crashes. I've heard they've gone to the drawing board and have released newer prototypes that are better now, though.

Of course, the more important part is how the chip jives with the clocks and the I/O connectivity on the board. So, of all these receiver chips, which are the best for a reference board design that can pass the lowest jitter and latency digital signal to the DAC? Do they all work equally well, and this is all a case of easy licensing from Thesycon and C-Media for drivers, or is there actually a hardware component to it?

And what about the other companies? VIA makes a VT1731 chip that only Kingwa seems to care about, and there's a smaller upstart called SaviAudio (subsidiary of SaviTech-IC) that makes a slew of low-cost Bravo SA9023/9027/9226/9227/9228 chips. Let's not forget that Resonessence uses Cypress chips, which are seldom used by the audio establishment, but their Windows code is modified Thesycon.

Of course, there are companies like CEntrance that writes it own code (doesn't Benchmark and PS Audio use CEntrance code?), as well as of course Gordon Rankin, but few others.

What am I missing here? Do I have a fundamental misunderstanding of this topic? I can't tell.