Maximum practical frames per second? Options

[HTS]

http://www.hydrogenaudio.org/forums/index....st&p=135606

For average human eyes, what is the maximum number of FPS which above that the human eye will no longer be able to detect skips? (i.e. the "motion" definition given)

I don't want to resurrect that thread.

This post has been edited by HTS: Dec 22 2011, 09:32

[knutinh]

http://www.hydrogenaudio.org/forums/index....st&p=135606

For average human eyes, what is the maximum number of FPS which above that the human eye will no longer be able to detect skips? (i.e. the "motion" definition given)

I don't want to resurrect that thread.

If you have content consisting of 1 white frame with a duration of 1/1000 second followed by 999 black frames each with a duration of 1/1000 second, you would probably notice it if the white frame was dropped.

So, since video does not play nicely with sampling theory it is hard to find the kind of content-independent hard limit that you seem to search for.

-k

[smok3]

short one:

Seeing framewise is simply not the way how the eye\brain system works. It works with a continuous flow of light\information.

longer one:
http://www.100fps.com/how_many_frames_can_humans_see.htm

thinkering one:
so if the actual tech of displaying one frame after another is the thing, then i keep thinking about motion blur and what happens with it if you shoot some motion at 1/50 or 1/500 of the Second. (or how annoying would be to watch 500 fps footage for some time, due to almost no motion blur)

practical one:
I did actually edit some 50p projects and they do look damn nice, also the file-sizes are not going to blow your editing suite. In the (near) future where editing might be practical with some complex Delta-compression codecs (like some sort of AVC stream), it might get practical to shoot certain events (sports maybe) at 200 fps (again, how to control motion blur?).

some facts:
right now the hobbit is being shoot at 48 fps (i think).

the question:
how to completely separate motion blur from fps? is that something thay may be possible with 3d cameras (one shooting at higher speed maybe?)

how to do a (free) test:
blender http://wiki.blender.org/index.php/Doc:Theo...phy/Motion_Blur

This post has been edited by smok3: Dec 27 2011, 20:12

[knutinh]

thinkering one:
so if the actual tech of displaying one frame after another is the thing, then i keep thinking about motion blur and what happens with it if you shoot some motion at 1/50 or 1/500 of the Second. (or how annoying would be to watch 500 fps footage for some time, due to almost no motion blur)

I think in terms of Nyquist sampling (this is an audio website, isnt it? .-) )

Whether you sample at 10fps or 500fps, you can have temporal aliasing (the reverse wagon wheel). Motion blur (and display "sample & hold" characteristics) tends to work as lowpass filters that reduce temporal aliasing/imaging. The cinema crowd have found that "180 degrees shutter" tends to work well. I.e. choose an integration time that it 50% of the frame period. 500fps @1/1000s shutter should look fine. 500fps using a hard-to-realize asymmetric gaussian-like integrator might work even better (?).

500fps video at sports events is going to have some hefty lighting/noise tradeoffs.

-k

[smok3]

500fps @1/1000s shutter should look fine.

i have some doubts that 180 degree shutter formula will actually work for such speedy fps, but it just might...
Again, blender seems to be a nice tool to test that, since it allows separated motion blur pass and whatnot.

This post has been edited by smok3: Dec 28 2011, 09:42