Brian, as I wrote in my email, I am not talking about asynchronous DAC which are, of course, the best solution for audio. I am talking about ADC. Cheap or not, asynchronous mode in record is not clock master. Adaptive is. I would love to get a high-quality chip that works but this is not the point. Xmos does not work either. Thanks,
JC
Le 19 août 2016 à 22:32, Brian Willoughby < email@hidden> a écrit :
Note that Windows still does not support High Speed USB Audio 2 Class Devices, while Mac OS X supports all of them with a class compliant driver. On Windows, each vendor must write a custom driver and users must install this driver before the device will work.
Given the choice between a system that supports cheap, low-quality audio out of the box, versus a system that supports high-quality audio with many channels without installing drivers, I'd prefer the latter.
I'm not sure why Adaptive IN is not supported. The USB Specification does require that Adaptive IN be paired with isochronous synch OUT data. Do these CMedia and TI chips implement reception of the rate feedback?
Perhaps there is an incompatibility between CoreAudio's pull model and adaptive input. I haven't studied the challenges involved at all.
Brian
On Aug 19, 2016, at 12:29 PM, Jean-Charles Rousset < email@hidden> wrote: The technical note TN2274 about USB Audio on the Mac stipulates that Adaptive mode is not supported for input stream (recording), only Synchronous and Asynchronous.
It is actually quite surprising as tons of consumer products integrate CMedia or TI chips (usb audio converter) which use Adaptive mode for record.
I tested a lot of them (cheap or not), also developed PCBs and hardware with different chips and, indeed, audio buffers are quickly corrupted probably due to a clock drift.
What is exactly the limitation to support it on Apple products ?
Note : I never had any problem with DAC (playback). Only ADC (record). Note2 : it works on Windows & Android.
|