In https://github.com/apfelaudio/tiliqua/pull/38, I modified a whole bunch of things to add support for the -25K LUT SC variant. This included modifying the PLL configuration to use 2 PLLs instead of 3 (reduced accuracy on audio Fs, was 12.288 / 48.0kHz, now 12.50 / 48.8kHz)
I have a followup task to either A) use a resampler between the IO and the USB streams, or B) rewrite/remove all the old eurorack-pmod gateware and migrate it to Amaranth so the I2S logic doesn't depend on a precise system clock (probably use oversampling).
As an alternative, we could just remove video support from the USB audio bitstream, as no other example really requires these precise audio sample rates.
In https://github.com/apfelaudio/tiliqua/pull/38, I modified a whole bunch of things to add support for the -25K LUT SC variant. This included modifying the PLL configuration to use 2 PLLs instead of 3 (reduced accuracy on audio Fs, was 12.288 / 48.0kHz, now 12.50 / 48.8kHz)
I have a followup task to either A) use a resampler between the IO and the USB streams, or B) rewrite/remove all the old eurorack-pmod gateware and migrate it to Amaranth so the I2S logic doesn't depend on a precise system clock (probably use oversampling).
As an alternative, we could just remove video support from the USB audio bitstream, as no other example really requires these precise audio sample rates.