Thanks for posting the comparison, the difference is certainly clear. AS you know, you are benefiting a lot because youve got camera pans, zoom & fast motion going on, and all of those look nicer with the higher framerate. Although all of those also reduce the efficiency of temporal compression, so people sitting relatively still with a static camera would probably not need to increase their bitrate as much if they up their framerate, as the differences between each frame are less. But like you said, theres probably less need for higher framerate in such circumstances anyway, although Im not sure if its totally true, maybe the emotions & gestures we see on peoples faces would seem nicer at the higher framerate, I havent really compared.
I did get a bit carried away today when I said there werent really any hardware reasons not to use higher framerate, for example I believe blip may use a lower framerate for all sorts of valid reasons, and mobile phones 3gpp format probably specifices a maximum framerate of 15. Cheers Steve Elbows --- In [email protected], Ron Watson <[EMAIL PROTECTED]> wrote: > > The difference in framerate was HUGE for our show. > > We do a lot of high energy, high velocity action footage. > > I thought it looked good on 15fps, but then someone here said that > moving to 29.97 wouldn't double the file size, or double the bitrate > (I think it was you Elbows...). I did and our footage was much more > appealing. > > I think we got a 25% or so file size increase, but it's more than > worth it. > > Of course, if you're sitting in front of your camera, or otherwise > not moving a lot, it shouldn't matter. > > See for yourself: > > 15FPS: > > http://blip.tv/file/get/K9disc-theArtOfK9DiscRememberTheSun514.mov > http://blip.tv/file/get/K9disc-theArtOfK9DiscRememberTheSun514.flv > > 29.97FPS: > > http://blip.tv/file/get/K9disc-RememberTheSun2997Fps973.mov > http://blip.tv/file/get/K9disc-RememberTheSun2997Fps973.flv > > I don't know if the FLV versions pick up the same quality change. I > have not checked them. > > Cheers, > > Ron Watson > > > On the Web: > http://pawsitivevybe.com > http://k9disc.com > http://k9disc.blip.tv > > > On Apr 24, 2007, at 12:57 PM, Steve Watkins wrote: > > > The effect of changing framerate isnt quite that straightforward, as > > most web formats use temporal compression. Instead of each frame being > > compressed in full, only keyframes contain the full image info. The > > frames that arent keyframes, just contain info about what has changed > > since the previous frame. This can be a highly effective technique, > > and means that how often you have keyframes will likely determine the > > necessary bitrate more than your frames per second will. > > > > This is one reason why I have always suggested people try > > experimenting with higher framerates in their vlogs, dont assume that > > it will make the compression articfacts twice as bad if you double the > > framerate, or that you need to make the bitrate twice as high to > > compensate for having twice as many frames. Nor should twice the > > framerate automatically be assumed to require twice as much CPU power, > > battery power etc to decode. > > > > Its also another example of Apples advice differing from the > > historical advice given by most in this group. Apple have never > > recommended using 15fps but thats often been the advice here. > > > > Certainly I couldnt declare 'everyone should use 25 or 29.97 or 30 > > fps' because 15fps is going to work better for some under certain > > circumstances, there is no 'right answer' although I expect higher > > framerates will become the norm eventally, as most portable devices > > can handle them ok there is no hardware barrier to this, more > > perception than anything else) > > > > Cheers > > > > Steve Elbows > > > > --- In [email protected], "Bill Cammack" <BillCammack@> > > wrote:Your only way > > > around that is to encode at a lower FPS so that you retain > > quality at > > > the expense of smooth, fluid motion, say, coming down from 30 fps or > > > 29.97 to 15fps. That way, you could get twice as much data per frame > > > because you're outputting half the number of total frames in the > > same > > > amount of time. > > > > > > > > > > [Non-text portions of this message have been removed] >
