Quote:
Originally Posted by Lesnikus
On the previous videos, Onyx had a huge time delay of 300 milliseconds, which is much more than that of Dasung Pro.
|
Where did you get the data for the Dasung? We searched and found nothing.
Quote:
Originally Posted by Lesnikus
Playing games with a delay of 300 milliseconds is simply impossible.
|
Not really, it depends on the type. And I picked games, after the joke from Fgdas, to create an empirical scale, like "diamond to talc". So, as you have read in the other thread, "Cuphead: impossible; Braid: doable; Machinarium: no problem".
Quote:
Originally Posted by Lesnikus
It's not about the frame rate.
|
Of course, who supposed that? But actually, that's not entirely true, in some situations you may miss a two or four extra frames to be wedged in the period (a brief check on New Vegas
).
Quote:
Originally Posted by Lesnikus
You can show simultaneously Onyx Boox Pro and LCD display (in mirror-mode), so we can estimate the time delay? I want to see how much Onyx lags behind the LCD display.
|
If you want proof, no problem, and I can post a zipped mp4 here - but why don't you just ask directly for the data.
For what VNC and TV are concerned, I counted and I can tell you that the lags are sometimes around 0.33s, or 0.5s, or 0.67 s (~10f, ~15f, ~20f on 30f/s), depending - and this using the VNC embedded in VMWare, since as you know x11vnc has a structural delay on its own because it is not integrated to the system to know that some parts of the screen have changed - it has to check.
Oddly enough, the delay is inconsistent - sometimes the shorter, sometimes the longer (move a window, move it again, close the window etc.).
For what HDMI is concerned, I did not even try. And one of the reasons is, it will give you a theoretical value that you cannot fully match with practice. Really, you already know that the delay is not negligible. At that point, to determine that 200ms is good and 400ms is not would be arbitrary. You have to try.
To what I experienced, the HDMI speed would be good for normal desktop use, not to test reflexes. To my experience, it seemed speedier than remote desktop. But do not forget that the shipped HDMI solution, the 'Monitor' software, currently forces you to A2 and has massively unoptimized dot state conservation - it is at least one revision far from competing with remote desktop alternatives.
EDIT: maybe the SDK has the access to the HDMI port, are you in to try developing a custom solution ; ) ?
('Cause I am not sure we can contribute code for the original 'Monitor' application.)