Two questions:
- What is the monitor switch that you’re using?
- Does the video from the stimulus PC go into the switch to both screens or are you using separate DVI/HDMI ports for each monitor?
Two things to try:
- Use DIN3 with the LCD display
- Unplug the second monitor to see if that improves your timing.
For the first “try” one - just to verify that you’re using only DIN3 in Net Station instead of allowing 255 or other combinations.
I’ve had good timing with experiments up to an hour. Admittedly haven’t tested longer than that.
Hello Peter,
thank you for your suggestions.
I tried the one , unplugged one monitor but nothing changed.
Unfortunately, I was not able to use DIN3, it did not work.
I have tried running the timing test with a default experiment aslo in E PRIME and i got the same results so I believe its not a problem of the software possible. I have read that the decreasing offsets and variability might be a problem of clock drifting of the stimulus PC and the acquisition one.I attached an image.
In the e prime guide there are instructions to calculate the clock scaling factor for the two pcs and put it the script of the experiment to adjust for it. I am attaching the file that there is in e prime. If I calculate the scaling factor using my program in psychopy can I added somehow in the experiment so it can adjust for it? I don’t know if it makes sense, but the NTP uses the amplifier clock and the netstation too but possile the machines have different frequencies ?
Thank you in advance for all the important help. At this point I do not know what else to think. Also becasue the DIN1 works well in the beginning but the drift advances with the task.
eprime file for clock scaling factor
CLOCK drifting idea
EPRIME
PSYCHOPY
Hi Anastasia-
Some relief that you see the same problematic behavior in E-PRIME that you do in PsychoPy. I would at this point bet pretty decent money that the problem is with the video splitter! I suggest trying to run a direct display cable from the stimulus PC to whichever monitor is easier and redoing the timing test. What happens when you try to use DIN3? Do you see the light on the timing tester flicker? Perhaps you need to change the DIN mode in Net Station to DIN3.
The next likely problem is that the display itself could be functioning in a less than desirable resolution. Can you tell me what the resolution is and what model monitor it is? Is the windows computer running Windows 10 or 11? Any anti-virus or other monitoring software running?
My recollection is that the old clock scale function isn’t relevant on 400-series amplifiers. I haven’t touched it since we go tour first 400 series amplifier 10 years ago and as I mentioned good timing here. I would focus on taking out the video splitter, using DIN3 and running the timing test again (with both E-PRIME and PsychoPy).
Happy to have a video chat or to rope in the EGI support team on this as they’re likely able to fix at least the E-PRIME timing and once that works PsychoPy should “just work”. I would recommend emailing EGI support and also Ryan (dot) Priefer (at) magstimegi (dot) com. Copying support@magstimegi.com is likely a good idea as well.
-Peter
Hello Peter,
So unfortunately even after taking out the splitter and using a cable connecting directly the monitor to the stimulus pc the problem was still there.
But i will try again. For the DIN3, for some reason it is always lit on the AV device, when I choose DIN3 on the output tranasporter and then on the Net station there is no DIN marker at all. Only if I change it to DIN1 and choose DIN1 on net station it works.
The monitor is windows 10 and there is other software running on the same time. I do not recall the rest of the details but I checked for luminance and tried with 2 different monitors.
For the clock it was my only possible idea after testing the direct communication without the splitter.
A colleague suggested using also the lab streaming viewer to sychronize all the computers.
It could be that its all fault of the DIN3 that it is not working but DIN1 works properly until a point like 15 minutes of experiment and then starts to have negative offsets that are also very variable. This is why I thought it would be a clock drift problem.
Thank you for the suggestion I will contact the EGI support for any insight on the AV device and e eprime. It would be great to have a video call and show you the details of the lab if you re available of course
Thank you
Hmmm that’s interesting. Can you post screenshots of your Net Station setup with the photocell hooked up? It should show DIN3 and some lights in the program display. Also photos of how you have the photocell hooked into the amplifier would be helpful.
What’s the screen resolution of the monitor? Refresh rate? Are these monitors you bought from EGI?
For PsychoPy you can try to move the eci_client.resync() call to the “End Routine” to see if that helps.
Happy to chat on video. First can you send me your experiment so I can try to debug it here? Email Peter (dot) Molfese (at) NIH (dot) GOV.
-Peter
hello Peter,
Sorry for my late reply. So basically the DIN3 does not work but the other groups of the lab are doing timing tests with DIN1 and they get good results usign E PRIME but a different computer!
The resolution of the monitor is 1920x1080 and 60Hz. I don’t know if they are the monitors form EGI.
I tried havung the resynch every 100 trials thats were the drfting is visible but it do not change anything basiccally in the offsets I will try also in the end of the routine.
Thank you immensely I am emailing you the task.
My best
ana
Hi, I’m curious if you all narrowed down the issue here with timings when using a Psychopy-EGI setup? Thanks.
Hello, unfortunately i did not, so I am using psychtoolbox in matlab for the task in the lab.
However, I have some colleagues in another lab that actually have good results with psychopy (using both a serial port or a parallel port) so it may be our system here that is
not compatible.
