I am wondering whether I should export my movie as progressive or lower-field first from PPro.

I've searched and googled, but haven't found an answer to my question, so I was hoping someone here might help...

Here's the way I work:
First I capture the DV-footage in Premiere, and cut the movie. Then I export clip by clip to MSDV avi-files with "Recompress" off, to avoid any recompression.

I usually run those clips in avisynth, virtualdub, and after effects to do various postprossesing. I always start by deinterlacing the footage in avisynth, and from then I work with the clips as uncompressed (!) to avoid recompression from all the various programs I use.

The final step is to import the uncompressed clips back to Premiere, place them on a new track above the original clips.

At this point I'm done and want to export everything to a DV clip, and here's the real question; at "rendering options" should I use "progressive" or "lower field first"? The footage is as I stated earlier on already deinterlaced.

Any feedback is welcome!

You should choose progressive, meaning that the footage is already progressive. Lower field first would mean deinterlacing already deinterlaced footage. Same goes for Upper field first.

Carl:

Can you help a newb in this area out.

I have heard a lot of references to Virtual Dub and Avisynth, and for my edification, how does it benefit you to use those tools in pre-editing handling ?

Why not edit everything in native DV format (in my case NTSC interlace DV), then run your final film through what ever post processing you want, ie., deinterlacing, converting to progressive or 24p ?

And doesn 't detinterlacing actually mean you are converting your footage to progressive ?

You've got it right, deinterlacing makes the frame progressive. The reason for me to use Avisynth is the fact that some plugins actually produce superior progressive footage compared to many commercial plugins. I normally go dv all the way, but after deinterlacing in avisynth I stailize and speed up the footage in After Effects. Therefore, good deinterlacing is crucial and because of the many steps I go for uncompressed video. It takes A LOT of disk space...

The interlacing method I use in avisynth is amazing, and not hard to use at all, even though documentation for avisynth is "by hackers for hackers" = non-understandable for normal video editors. Avisynth acts like a frameserver. What you simply do is to create a text file and save it as script.avs or whatever you like, mine looks like this:

DirectShowSource("d:\fix.avi", fps=25)
interp = SeparateFields().SelectEven().EEDI2()
TDeint(edeint=interp)

The first row specifies the input file and where it is located, and the framerate of the file. The second and third row uses two plug-ins that have to be downloaded and placed in the avisynth pluginfolder.
http://www.missouri.edu/~kes25c/EEDI2v092.zip
http://www.missouri.edu/~kes25c/TDeintv1RC7.zip

Then you open the .avs file in VirtualDub, just like you normally would open an .avi file, and voila, there's the deinterlaced footage. Then save it as a new .avi.

The only reason for me to use uncompressed is because of all the importing and rendering in virtualdub and aftereffects. The dv codec really destroys the video quality, especially red parts. This is due to the chroma sampling in dv.

Carl:

Thanks for the information. That gave me a better idea in a short explanation of what people are doing with scripts and Avisynth. In the early days of PC I did some rudimentary programming in Basic and Dbase programs, using similar text files, so I understand what is being done with scripts. I will look a bit closer at Avisynth and VirtualDub to see what I can pick up. Thanks again !!

Returning to my question, theoretically, would there be any "problems" with the fields when importing uncompressed progressive footage to the timeline (the project settings are standard dv - lower field first)? I mean, fields or not shouldn't matter because the two fields in the progressive footage together make up a progressive frame. Or am I confused? :)

From what I understand, the editor in Premier reads and uses progressive and interlaced videos and edits them while they coexist. Then the rendering options allows you to choose if you want to interlace, deinterlace, which field to use or to stick with full scan. You should not have problems using progressive video in Premier. Just make sure that if you want your videos to stay progressive when you export, that the tricker is set to progressive and not even field or odd fields.

I wonder how MSDV format actually stores the information, because progressive footage in a lower-field-first project doesn't have to be re-rendered. So I wonder if it's just a some kind of field-order byte purely intended for playback that is changed in the MSDV header when going from lower-first to pårogressive or vice versa.