On 2/16/07, <b class="gmail_sendername">jason maxwell</b> <<a href="mailto:decepticon@gmail.com">decepticon@gmail.com</a>> wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
If one single frame (30 FPS) is interlaced into 2 separate fields for<br>broadcast at 60 FPS, then each pair of fields would be from the same<br>point in time, and therefore should be able to be reconstructed w/o<br>artifacts.
</blockquote><div><br>Yes. If the source material was progressive 30fps then pairs of fields would be from the same frame. Of course you still have to figure out which two fields are from each frame. If the source material is progressive 24fps (
i.e. film) then it will have been interlaced using a telecine (aka 3:2 pulldown) process. This results in 2 frames being converted to 5 fields in a particular order. If everything is flagged right it is possible to recreate the original frames.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">However, If I understand Stephen's post correctly, the real issue is<br>that interlaced material is not always generated in such a simple
<br>manner, and that many times each field does represent a new moment in<br>time.</blockquote><div><br>Right. A lot video is shot interlaced or is shot progressive at more than 30fps. In these cases each field contains data from a distinct point in time so creating a frame by weaving the lines from two fields will not look great if there is rapid movement in the scene.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> This is sometimes further complicated by the need for pulldown<br>methods to compensate for different framerate standards. Correct? This
<br>is stupid. Oh well, what am I gonna do about it?<br></blockquote></div><br>It's pretty common these days for people to say interlacing is "stupid" and wonder why we have a system that introduces these complications. The reason is that interlaced video generally looks better that progressive video at a given bandwidth. Let's look at 540p60, 1080i60, and 1080p30. These all have the same bandwidth (or at least would have if they had the same horizontal resolution). Consider a fairly static scene. In this case the extra framerate of 540p60 is not needed, but the extra resolution is. So 540p60 looks the worst. 1080i60 and 1080p30 will look very similar. Now consider a scene with lots of movement. In this case framerate is most important - 1080p30 will look more jerky than the others. But the interlaced mode shows 540 lines of new data every 1/60 of a second, so it looks about as good as 540p60. Interlacing gives you the best of both worlds.
<br><br>Steve<br><br>