While searching the forum on converting 1080i to DVD, I've noticed that some people do some crazy insane AVISynth scripts to resize and keep the video interlaced. No offense, but I'm just wondering why do people do this.
For me at least, when I convert to DVD, I usually set my encoder to ignore one field and then have it resize the other field to 480, which I then author as a 480p DVD. That way, I'm going from 540 lines (one 1080 interlaced field) and then dropping the resolution to 480 progressive lines.
Since one 1080i field still has more lines than one progressive DVD frame, shouldn't my method be fine?
+ Reply to Thread
Results 1 to 6 of 6
If the 1080i source is truly interlaced (shot on video), by tossing out half the fields, you've cut the motion fluidity in half. If the 1080i source is from a movie (shot on film), by tossing out half the fields, rather than doing a full IVTC, you've made a video where every 5th frame is a duplicate frame, making it play jerky. To answer your question, your method most certainly is not fine.
Drop field deinterlacing will create moire artifacts:
thanks for the replies.
then I have on question: when I do an 1080i to 480i conversion, I usually end up with some weird combing effect or jagged edges on sharp corners in the video. Is this because I have the wrong field selected? It happens in MainConcept even when I select the other field...
1080i is interlaced, so it stands to reason that you'll find interlacing. However, you can't just resize down to 480i. That destroys the field structure, and is the reason for those convoluted AviSynth scripts you mentioned in your first post.
And if it's a movie, then most likely it can be IVTC'd to 23.976fps, resized normally, and encoded for 23.976fps progressive with pulldown applied to output 29.97fps for NTSC DVD.