Quote:
Originally Posted by broken head  You are right. It would be essentially double roundtrip. The .35 was one direction. .89 would be the round trip total. By the way, the AES16 driver will detect a connected Aurora and report that modified latency value up to the ASIO app, so manual DC is not required.
The Aurora definitely uses some of the faster converters. AD latency at a 2X rate is 9 samples, DA is 4.6 samples. 3 samples are required at any rate for servicing the FPGA. |
Quote:
Originally Posted by David Hoatson Round-trip latency for the AES16e is the same as the AES16, which is 4 samples plus the buffer size. |
So... using the 32 buffer:
A/D: 9 samples
In buffer: 32 samples
Servicing the FPGA: 3 samples
Out buffer: 32 samples
D/A: 4.6 samples
That's 80.6 samples roundtrip, unless I forgot something (BTW, what's "2X rate"?)
80.6 samples is 1.82 milliseconds @ 44.1, or 0.84 milliseconds @ 96 kHz, which is terrific.
Does the driver support recording @ 96 kHz - using the 32 buffer - without getting artifacts? That's not possible on a RME Fireface 800...
If these numbers are correct, the total roundtrip using the 64 buffer would be 144.6 samples, which is 3.28 ms @ 44.1 or 1.5 milliseconds at 96 kHz. Also impressive.