Blur Busters Forums

Hello,

For the first time since TestUFO launched, I've modified the "default test" at http://www.testufo.com

-- It now displays the starry background by default.
-- Starry background is dimmed by 25% as it was overly bright earlier.
-- It now automatically displays 4 bands on 240Hz monitors (240fps, 120fps, 60fps, 30fps)

Here's the [/testufo] embed for forum posts on Blur Busters Forums

It would be helpful to be able to switch to a mode with whole divider FPS amounts.

144Hz: 144, 72, 48, 36.
165Hz: 82.5, 55, 41.25, 33

And so on for 180Hz and 240Hz. Basically every frame rate > 30. So this can be a dynamic calculation (find all dividers, until it's less than 30) without having to hard-code refresh rates.

RealNC wrote:It would be helpful to be able to switch to a mode with whole divider FPS amounts.

144Hz: 144, 72, 48, 36.
165Hz: 82.5, 55, 41.25, 33

And so on for 180Hz and 240Hz. Basically every frame rate > 30. So this can be a dynamic calculation (find all dividers, until it's less than 30) without having to hard-code refresh rates.

Mathematically, what you say, the number of rows is as simple as just doing this: Math.ceil(refreshrate / 30) ....

That's assuming if I did, instead of halving framerate (/2 /4 /8 /16) for every row, it could go like this (/2 /3 /4 /5)
In this case, 120/1, 120/2, 120/3, 120/4, 120/5, etc.
(which would result in 120, 60, 40, 30, 24 for 120Hz refresh when TestUFO configured to 5 rows)

I thought long and hard about doing the /2 /3 /4 /5 method, but I need to keep TestUFO simple & backwards compatible for now. Many people have posted URLs to TestUFO throughout the Internet that automatically configures the TestUFO test, and I didn't want to break those existing links creating existing framerates for specific refresh rates. You can manually change the number of rows.

That said, an additional setting to influence divisors, may be warranted. I have to think about how to do it, without complicating the "TestUFO Home Page" or breaking backwards compatibility with existing TestUFO "preconfigured" links.

It would be interesting to ask a bunch of people which fps looks smooth and at which point does it begin to stutter, because i suspect that the usual 24fps wisdom is pretty much a factoid.

lexlazootin wrote:It would be interesting to ask a bunch of people which fps looks smooth and at which point does it begin to stutter, because i suspect that the usual 24fps wisdom is pretty much a factoid.

24fps is where a slideshow starts to look like motion. The number where you can't percieve a difference in smoothness will be a lot higher. Though the exact number will depend heavily on the speed of the animation, and on how sharp the moving edges are. Kind of hard to test with a double blind trial, because of the other artifacts involved with framerate, refresh rate, and persistence. The number for a 100% persistence display will probably be a lot lower than a low persistence display, due to motion blur. A side by side comparison will also likely be more noticeable than an A then B comparison.

I've heard the "24fps is where a slideshow starts to look like motion." but I've never been able to understand that in practice. When ever i test this, anywhere near 24fps is WAY to low it always seems to be around the 50fps+ mark where "slideshowing" starts to disappear and true motion takes effect.

Obviously this is when i am eye tracking something in the room and not just looking at a fixed point on my monitor. I've noticed this mostly playing HL, where a lot of low fps tricks are used. But the same thing happens on these tests, low persistence or not. Around 50fps+~ for me is the point where it starts to look like motion. I would love to hear someone elses take on this.

The 24FPS thing was introduced for film, not animation in general. The motion blur captured by film is what allows 24FPS too look like motion. Without it, it just doesn't work. Same for video, btw, not just film.

Artificial motion blur is difficult to do. For a motion blur algorithm to produce good motion blur, it would need to render frames at a multiple of 24FPS, then convert back to 24FPS and use the excess frames as the source for the motion blur generation. The whole thing is of course an exercise in futility for video games (if you can render at a multiple of 24FPS, there's no point in showing 24FPS.) It's only useful for making videos of video games. Freeman's Mind was produced this way (I believe the recording was at 240FPS, then converted to 30FPS with the excess frames being used to generate high quality motion blur.)

Without motion blur, 24FPS will always look like a slideshow.