cuda_cdev is the convolution/big operatrions device and coda_dev is the one for the rest.
So it is usually good to assign cuda_cdev the GPU that has more RAM (if thereās a difference).
Great machine! Enjoy fiddling around the DSD 512 filters too! Can also run basic DSD1024 settings without DAC correction! For DSD256 you have all possible combinations! Suggest to use the new DAC profiles settings to make it easier!
I set up the May DAC, and DSD128 works with the powerless computer I currently have.
I assume the two green bars to the upper left show the computerās load.
When changing the fs to DSD256, the bars went crazy and produced no sound.
So, Iāll stick to DSD128 for a while. (Iāll try PCM later.)
Which filters/modulators would you suggest?
Upper green bar is input buffer that is filled with source content, in this case pushed to by the HDSPe AIO. Lower green bar is output buffer that is filled by HQPlayer processing and where the DAC āpulls fromā.
Thank you for explaining what the bars indicate.
Which filters would be good for DSD128?
The DAC arrived a month sooner than expected, and I havenāt done any research.
Iāve found the sound doesnāt come out simultaneously from the following two devices ā the May DAC (USB input) and Yamahaās CX-A5200 (coaxial input). I use Holo RED as a streamer.
If I want to use the USB input on the May DAC, I need to select on the HQPe Configuration page:
naa-xxxxxxxx: Holo Audio UAC2.0 Gen2.1 Enhanc: USB Audio
But there wonāt be any sound produced through the Yamaha AV receiver.
After trying several configurations, I figured out that selecting:
naa-xxxxxxxx: Holo Audio UAC2.0 Gen2 - RED: USB Audio
and changing the Mayās input to I2S, then sound comes out from both devices.
I use the above ātwo device outputsā during work, because I use two rooms: the main listening room for drafting and my study for typing. The Yamaha receiver sends music data to the study by using the Zone function.
So, in short, I wanted to say the Mayās I2S input can be useful in some cases.
You can now also achieve this with the new combo backend feature. But your current solution will certainly ensure better syncing since thereās only one audio device and one clock, and output format is anyway limited by the S/PDIF link to the receiverā¦
In some other cases using for example two NAAs would allow DSD outputs in multiple rooms, etc.
Iām listening to ASDM7EC-fast 512+ and havenāt heard better transients in my system. Cymbals especially. I can do 1x sinc-Lh with DAC correction or sinc-MGa with DAC correction off. I prefer sinc-MGa but am concerned at the long term health effects of listening to an uncorrected DAC.