Hi,
After some years I am looking to replace the stimulus computers in my EEG Labs. Intend to use consumer grade equipment (not CRS/Pixx) and work within a budget. Reading through previous suggestions from Mario and others and being limited by University purchasing rules (ugh) and availability of products in my location (ugh) I have arrived at the following key specs. The key features we want for our lab are good stimulus timing (we have diode set ups etc for verification) - we don’t do experiments with colour. Plan the computer to run Ubuntu/MATLAB/PTB.
Computer:AMD Ryzen 5 3600
Graphics: AMD Radeon RX 5600 XT
Monitor(s): ASUS ROG Swift PG259QNR 25" 360Hz (see Asus ROG Swift 360Hz PG259QNR)
I/O: Labjack U3-LV USB DAQ device
Keyboard: Corsair K70 MK.2
I am really unsure about what audio card (USB?) to get that would be linux supported - for standard ERP experiments.
Anyone see any problems with these specs or have any suggestions?
The specs look good to me. I’ve switched over to use the LabJack T4 as they can run code like an Arduino which makes them more flexible, and can run off PoE ethernet to make setup more flexible, but for digital I/O that is irrelevant and they are $100 more. I just use the built-in audio but I only use audio for subject feedback so I’ve never measured audio timing accuracy etc.
You should test with a photodiode that the monitor doesn’t shift luminance during display, as many gaming monitors have dynamic contrast settings. My Aorus FI27Q-P changes black backgrounds as a stimulus comes on, and I can’t turn it off in the settings.
That monitor looks like it has great timing. I was curious about the NVidia Latency system built-in. I assume it has a built-in photodiode somewhere but the marketing spiel doesn’t give any details. The best I could find was a paper published by NVidia where they measure system latency:
Which assumes they both monitor the mouse input and a photodiode embedded somewhere in the display. It would be great if the SDK worked on Linux, so it could potentially be accessible from PTB…
Hardware config looks ok to me. The Asus monitor does support AMD FreeSync / DP adaptive sync, which may come in handy. While i don’t have first hand experience with that AMD gpu, i don’t have any reason to believe it wouldn’t work well, and i think we had now multiple PTB users with good results under Linux with that or a similar model. It is good to have a modern and fast gpu especially if you want to drive a high refresh rate monitor and/or want to make use of PTB’s fine-grained stimulus timing mode, which is depends on AMD Freesync on Linux for best results.
Standard onboard sound chips (aka Intel HDA compatible) under Linux usually perform with timing accuracy better than 1 msec, and latency in the < 15 msecs range, usually between 5 - 10 msecs, iow. excellent, as far as my testing goes. So unless you are quite unlucky with the specific model in your machine, the onboard sound will probably just fine timing-wise.
Although you are used to Labjack interface, I would just like to mention our toy, RTBox:
It is convenient to use for event markers as well as subject and photo diode signal). It should be the only interface between the stimulus computer and EEG system.