Any "picture quality" difference between 24p or 30p?

[Hayes Roberts]

This may have been covered somewhere, so I apologize if so. I have done some comparison tests of footage ( same in-camera settings; Panamatch and TC3 1/60 and 1/48 ) shot at 24p and 30p. It seems, to my eye, that 30p footage looks cleaner and slightly sharper. Is this an accurate assesment? Assuming material is shot for direct to video, and there is very little discernable difference in "look", when viewing on typical lcd's, plasmas, etc(24p vs 30P) what would be the logical choice when shooting to capture best image quality and most color information?

[Tim Dashwood]

There is no technical image difference except for the default shutter speeds for the frame rates - and hence "temporal" resolution. If you used the same setting and only changed the frame rate, then the shutter speed should be the only additional variable.

Since 1/60th is a shorter moment in time than 1/48th, the motion blurs will be slightly shorter, and moving objects could be percieved as "sharper." Try using a high shutter speed like 1/100th on both frame rates to confirm your observation on a frame-by-frame basis.

[William Hohauser]

The answer is yes. 30p at 1/60 is going to look sharper as there is more motion information present when compared to 24p at 1/48. Other qualities such as color information and image quality are exactly the same if assessed on a frame by frame basis. The pick-up elements in the camera and the recording system are exactly the same between the two frame rates. However the extra motion of 30p increases the percetion of detail and lifelikeness. 60p is very lifelike. This a perceptual quality that is hard to quantify, just go by the rule that the slower frame rates become more "dreamy" and faster look more "real". 24p is an accepted balance of dreamy and real.

Here's an experiment for you to try; shoot 30p at 1/30th shutter. See how much this simulates a film look to your eye. I shot a music video this way and the client loved the way it looked. I was happy that I was able to capture and edit without going thru a time-comsuming transcoding process.

[Stephan Ahonen]

30p is trying to shove more frames into the same bandwidth as 24p, so *theoretically* it should look slightly worse. In practice you've probably got to break out the magnifying glass and compare freeze frames to really tell.

[Tim Dashwood]

Originally Posted by Stephan Ahonen

30p is trying to shove more frames into the same bandwidth as 24p, so *theoretically* it should look slightly worse. In practice you've probably got to break out the magnifying glass and compare freeze frames to really tell.

JVC's ProHD format apparently doesn't work that way.
I originally theorized the same thing and mapped this all out on a chart with the different pulldown patterns within 60P to see if more or less data was dedicated to each of the flagged frames within the 60P, but in the end it made no difference between 24P or 30P. The data per flagged frame is exactly the same in 24P and 30P, and both still map the flagged frames to 6 GOP. It seems there would just be one additional GOP per second in 30P - which doesn't necessarily mean that less data is allocated per GOP.
This didn't seem to make sense to me at first from an engineering standpoint, but I suppose this method was chosen for the sake of DTV compatibility.
Actually, it all makes my head spin a little.

If you are interested, here is that original discussion from over a year ago with input from Barry Green, Steve Mullen and David Newman.
http://www.dvinfo.net/conf/showthread.php?t=54985