With my Sony cameras (HC1 and FX7), which film at 60i, I can set the shutter speed to 1/30 and as a result get pairs of fields recorded at the same moment in time (I believe this is done by buffering the sensor output for the 1/30 duration, then outputting two fields-worth of information to tape).

The thing that has always puzzled me is that when I pull that footage into Premiere, it still looks interlaced, because instead of pulling in the paired fields (AA then BB then CC and so on) it insteads pulls the footage in as AB then BC then CD and so on, which obviously looks like cr*p.

Using avisynth and virtualdub, I manually separate the fields, cut off the first and last field, then weave the remainder back together to get nice progressive BB CC DD footage.

My questions are:
1) why the heck does the NLE load the footage this way???
2) Can anyone suggest an NLE setting, or avisynth script, or whatever, that will automatically ignore that first field (i.e., can I speed up the the rather tedious manual trimming process I go through presently)?

Hey Graham........

Before any conclusions are drawn, the only true way to know if you have true progressive frames, are if you do a field by field comparison.
It's been a while for me, but if you do a SeparateFields() in AVISynth, load into Vdub, and move keyframe by keyframe, and see what does the footage does?
Does it step forward, step back?
Or does it actually create two true identical frames??

I was always of the school of thought that shutter speeds in the Digital world was more of an exposure length onto the sensors, rather than a physical action...

If you separate fields on the raw footage, the temporal sequence is ABBCCDDEEFF.. etc

The two members of each pair are identical time-wise, but one is an upper field and one is a lower field, and so (I believe) together retain the full resolution of the 1080x1440 image..

Edit...incidentally, the motion blur of each pair joins seamlessly to the next, giving the effect of a 360-degreee shutter.

And yes, the "shutter speed" is how long the sensor is read - if the shutter speed is slower than the 60i field rate, the data is accumulated in a buffer, and then sent out to the appropriate number of fields. For example, at 1/15 the buffer contents would be sent to 4 fields.

I remember an old post, (in the dark cobwebs of my brain), in where the only true way to interperate progressive footage was by analyzing frame by frame via SeparateFields()..
That being mentioned, when you create a SeparateFields(), the footage becomes 720x240, for each successive frame. Of course, if the BB CC DD frames are jumping up and down on the script, you've definately got field based interlacing...

Not trying to drag the topic (since my knowledge is limited), but assuming you're correct with your progressive footage, perhaps it's simply a Project Settings in Premiere that needs to be tweaked?