Does 5:5 pulldown for 24p on 120Hz TVs cause a lot of motion blur?

[sldvd]

I've read that 120Hz LCD TVs can do a 5:5 pulldown on 24p material so that it plays each frame 5 times, resulting in output that looks like 24 frames per second. This is supposed to be better than 3:2 pulldown (or I guess for a 120Hz TV it would be 6:4 pulldown), because 3:2 pulldown results in a judder effect.

But if it looks like it's 24fps (or 24Hz), then won't it have some serious motion blur? Even with 60Hz LCD TVs, there was some motion blur and that's why they made 120Hz TVs that do frame interpolation to reduce the motion blur. 24Hz is much less than 60Hz so I would imagine a 5:5 pulldown on 24p material would result in even greater motion blur - is that true?

[jagabo]

Look at a photograph 5 times. Do you see more motion blur than if you looked at it one time? 5:5 frame repeats cause no motion blur. If there's motion blur in the film source you'll get motion blur on-screen. And LCD's have an inherent motion blur from the relatively low switching time (and this is one reason why they've gone to 120 and 240 Hz). Motion interpolation often causes blur and other artifacts when the motions aren't properly interpolated.

[sldvd]

Look at a photograph 5 times. Do you see more motion blur than if you looked at it one time? 5:5 frame repeats cause no motion blur.

Right, a 120Hz TV that does 5:5 pulldown would output frames every 8.3ms, so that the 5 repeated frames are on the screen for a total of 41.7ms, the same amount of time between frame changes for a hypothetical 24Hz TV. I was just thinking that the amount of perceived blur from the sample and hold in that scenario would be greater than in the case of a 60Hz TV, where there's a frame change every 16.7ms. And then a 120Hz TV that does interpolation would have less perceived blur than a 60Hz TV since the frame is changing every 8.3ms.

[jagabo]

Right, a 120Hz TV that does 5:5 pulldown would output frames every 8.3ms, so that the 5 repeated frames are on the screen for a total of 41.7ms, the same amount of time between frame changes for a hypothetical 24Hz TV. I was just thinking that the amount of perceived blur from the sample and hold in that scenario would be greater than in the case of a 60Hz TV, where there's a frame change every 16.7ms.

No, on a 60 Hz TV the film frames are visible for 33 ms or 50 ms (ie, 2/60 or 3/60 second), alternating between the two. Hence the judder. Between each of those 5 frame repeats (or 2 or 3 in the case of 3:2 repeats) the picture is unchanged (aside from the previously mentioned switching lag). Using 120 Hz refresh lets them optimize the LCD switching for less blur, not more. Ie, the LCD cells are optimized to switch in 1/120 of a second rather than at 1/60 of a second. So the switching lag is shorter and you see less motion blur.

[sldvd]

No, on a 60 Hz TV the film frames are visible for 33 ms or 50 ms (ie, 2/60 or 3/60 second), alternating between the two. Hence the judder. Between each of those 5 frame repeats (or 2 or 3 in the case of 3:2 repeats) the picture is unchanged (aside from the previously mentioned switching lag).

OK, I get it now - with the 3:2 pulldown on a 60Hz TV, the frames are visible for an average of 41.7ms, which is the same as a 120Hz TV using 5:5 pulldown, so there really isn't any difference in sample and hold blur between those.

Using 120 Hz refresh lets them optimize the LCD switching for less blur, not more. Ie, the LCD cells are optimized to switch in 1/120 of a second rather than at 1/60 of a second. So the switching lag is shorter and you see less motion blur.

Are you talking about the response time of the LCD (the time it takes for pixels to completely change their color once the TV tells them to change), or something else?

[jagabo]

Using 120 Hz refresh lets them optimize the LCD switching for less blur, not more. Ie, the LCD cells are optimized to switch in 1/120 of a second rather than at 1/60 of a second. So the switching lag is shorter and you see less motion blur.

Are you talking about the response time of the LCD (the time it takes for pixels to completely change their color once the TV tells them to change), or something else?

Yes, the response time. Note that the response time values reported by manufacturers have little meaning. They report the fastest value they can find and use rather loose specifications. If they need to report a faster response time than a competitor to justify a hither price they'll just change the test requirements until they find something that gives a better number. The time it takes to change from black to white is different than the time it takes to change from white to black. And the time it takes to change from 25 percent to 75 percent is different than the time it takes to change from 10 to 90 percent. Etc.