I don't read ASR anymore, being misled couple times. Has it been tested on a number of devices or on a selected machine? I think a single implementation was under test which doesn't bring conclusive result about majority devices on the market, only a specific one. And now a question related to the specific device. Was I2S interface galvanically isolated? S/PDIF normally is isolated, I2S not common. ASR gear is rigged by ground loops as a Total DAC test shows. It is known that ASR is introducing ground loops to the device under test.
In contrary to the ASR results I read user feedback all around and it shows in majority support for I2S as a better sounding solution. Better than S/PDIF and also better than USB. Why?
Two reasons:
1. USB port is typically not isolated. After market isolators do not do a job properly. Only USB converters to I2S received good opinion, but now we are talking about I2S. Bottom of line is that if device has USB port not galvanically isolated, it is not so good as I2S.
2. USB has a facility to eliminate jitter. In the asynchronous transfer mode there is no need for clock recovery. Problem (a) it is not used. Problem (b) there is no advantage of asynchronous mode if device still use clock recovery. It is why users do not report any advantage of the asynchronous transfer mode.
Re
b). Breaking news is that Audio GD R-7, R-7HE 2020 models and a new DI-20 digital interface device use galvanic isolator and USB asynchronous mode already, there is no clock recovery in this mode. There will be an essential upgrade for the older and current R-1, R-8, R-8HE, R-27 and R-28 models which include a similar upgrade. Details on the A-GD website.
Rea). R7 (2020) and DI-20 were already for couple months and users didn't figure it out that they have to configure their source accordingly. I found that some applications do not activate asynchronous mode by default. In Foobar it is easy, select WASAPI Event device, but in Roon it is hidden in the Advanced menu.
Soekris DAC-2541 comes with USB isolator. I have exchanged messages with the developer asking about clock recovery in USB asynchronous mode. His first response was very standard that he does FIFO plus PLL reclocking on all ports, including USB (not answering a question), but the second response was not very clear. Whatever it is, I take it that there is a consensus that it should be done without a complicated clock recovery. Even if not implemented now, it will be done soon.
