It gives a good example of why some people think near IPD is what we’re supposed to use. Really, it’s not but it does result in a more focused image. Now, some might ask. Why did Pimax make it so that the IPD is not measured from the center of the lenses? Well there are a couple possibilities. One would be that they wanted to advertise a larger IPD range. Two would be that they needed to do this in order to reduce distortions. Or, maybe a bit of both. Anyway, some people really don’t like this because it results in the image being slightly out of focus. So naturally people will just reduce the hardware IPD until the lenses are in focus.
However, if I understand it correctly this causes an issue. Pimax’s software assumes that your eyes aren’t seeing directly through the sweet spot. It thinks your eyes are looking more through the edge of the lenses. This is a problem, because Pimax moves the image to match with the IPD they display. However since your eyes aren’t where Pimax think’s they are, they put the image in the wrong spot. This results in your eyes looking somewhat inward, sort of like when you’re looking at your nose. Obviously this is very uncomfortable and results in eyestrain. My hypothesis is that by setting the IPD offset as the difference it takes to get to the center of the lenses (plus a bit more since we are looking at a different location on the lens), you should be able to fix this and get a normal focused image without eyestrain.
Here’s an example. Let’s say your IPD is 65 but you need to reduce the hardware IPD from 65 to 61 in order to see through the sweet spot of the lens. That’s a reduction of your normal IPD by 4. So, you should increase the IPD offset by +4(plus another few mm to center the view). Now unfortunately I can’t test this right now, because my computer is not finished. But if anyone else wants to try this, please let me know if it works for you. I may be wrong here, but if I’m understanding how Pimax’s IPD works then this should fix the problem in this specific scenario. You’re essentially moving the center of your view over to the sweet spot of the lens via the IPD offset.
(Edit: Doesn’t work In theory it should, but in practice it doesn’t. Weird )
I understand where you are coming from, but I think this is merely a coping mechanism for already wrong fit (moving small focused area more to center). First thing I would always do is to set natural IPD, remove the face cowling and move the headset around relative to eyes to see where is the best position and work from there. I am out of luck as the position with best picture is without the cowling, pressing the lens as close to eyes as physically possible.
If your issue is opposite, it is always easier to add some more padding to increase the distance to get to optimal distance from lens and get wider sharp and distortion-less area.
This is definitely a reasonable idea, which I would suggest to anyone, who suffers from the lens center offset (https://community.openmr.ai/t/interesting-fix-to-the-pimax-eyestrain-problem/25815/10). I would also like to point out that the post you quoted (and which seems to be quoted recently) was written when there was no software IPD offset option in Pitool, so there is no word about it there.
There are just few “details” to keep in mind, before advertising it as a general solution.
Moving the lens into a wrong (smaller) IPD not only changes the virtual camera position (which can be compensated by software IPD offset), but also changes the perceived lens distortion. I am not sure that Pimax recalculates the distortion function relative to the software IPD offset. If not, you are going to suffer not correctly “undistorted” image (i.e. some distortion artefacts, which are not compensated, because you look through the lens differently from what Pimax used in the distortion model).
You may need to move the lenses in a different IPD position, even if you do not need to look through theirs optical centers, to compensate for the eye relief difference. Pimax can calculate the undistortion function only for one eye position (model position). If your eyes are closer to the lenses, in order to preserve the lens axes meeting the eyes at the right spot, you need to move the lenses closer. Correspondingly, when your eyes are further from the lenses (than Pimax original model), you need to set the hardware IPD even larger than your real IPD. In this case, you are simply compensating for difference in the eye relief and should use the software offset in the same manner as you suggest.
The second point also means that people who needs to compensate for the smaller eye relief have an advantage, because they get a bit better (less blurry) image because of the more favorable lens position.
On the other hand, people who have the eyes further from the lenses (than the model) would need to increase the IPD even more, and while they also could compensate it by software IPD offset, they will suffer even worse image (optically).
I will add that when I did it by -4/+4 it felt uncomfortable after time - eye strain and hint of cross-eye feel. Perhaps compensating so much by SW IPD does not work well, or we are getting too far from the correct viewing spot on the lense. At the end I settled on some compromise around -2/+2.
So experiment with several settings and see what is best [for you].
Also my Pitool does not remember (save) SW IPD setting. Each day I have to enter it anew. So if it feels wrong next day check in Pitool if it was not just reset to 0.
Same. I don’t know why it doesn’t work like you’d think it would. I’ve found that my normal IPD seems good. Maybe -0.5 vertical offset because my right eye is higher than my left.
I’ve tried this and it always felt awful honestly, hardware IPD around my actual IPD has been best. With the lens sweet spot being a little off center for each eye, it’s frustrating to get used to, but I think this is how the hmd is designed.
While playing with this I’d also recommend being careful referencing what pimax claims is the current IPD. I’m not sure if they’re using an encoder - but it is definitely not accurate (you can move it without software coming up saying you did). To ‘calibrate’ move the slider all the way to the max, then back to minimum. Should read ~59 at min, ~70 at max. Last time I took the hmd to the minimum IPD it claimed i was at ‘62 IPD’ which is obviously not the device min.
In theory it should, but in practice it doesn’t. Weird 
)
I’ve found that my normal IPD seems good. Maybe -0.5 vertical offset because my right eye is higher than my left.