24PA w/16:9 Adaptor in AE.

[stidle]

Okay...I've read through all of the threads I can find on how to work with widescreen 24pA footage in After Effects, but I am just a little more confused now than I was before! There seems to be a lot of conflicting information. Let's get this straight.

According to the Trish/Chris Meyers book, their preferred method is to work with the footage using square pixels in a 16:9 comp -- i.e. make the comp 864:480 w/square pixels while you are working with it...and then, when you are outputting it, "squeeze" to the original anamorphic dimensions of 720:480.

Makes sense to me, though I haven't tried it yet. Has anyone used this method?

 

[milksac]

Short answer #1, yes.

Long answer.
For D1, you can work at 960 x 540 or 864 x 486 and drop into a 720 x 486 comp for output. Both are 16:9 ratios. For the DV footage (720 x 480) I'm not quite clear on the workflow. I suppose if you started with a comp of 864 x 480 instead of 864 x 486 it would all work out. That 6 line difference is a bugger. You can crop those 6 lines, but if you're not careful about it you can throw off the field order.

Short answer #2, trust the Meyers.

 

[Matt Grunau]

For the DV footage (720 x 480) I'm not quite clear on the workflow. I suppose if you started with a comp of 864 x 480 instead of 864 x 486 it would all work out. That 6 line difference is a bugger. You can crop those 6 lines, but if you're not careful about it you can throw off the field order.

If you are working progressive, and (should be presumably) outputting progressive, how could you throw off field order?

Why work with square pixels at all? If you are working with footage that is squeezed, and will be outputting footage that is squeezed, and can use generated elements with the correct aspect ratio to make them squeezed, where is the advantage of the extra steps of additional comps?

I've been wanting to play with true 16x9 footage, and their method seems a little odd. I don't get it.

Seems to me easiest and safest way would be export your footage from your NLE as a 24 bit Quicktime file with the Animation codec (highest quality), progressive, in widescreen 23.976, and then work from that.

All are preset settings in AE, and you dont have to worry about pulldowns and whatnot. All graphic elements can be generated in Photoshop with presets for widescreen, and should work perfectly.

Is there an explanation from the Meyers as to why to take their route? Something maybe like a quality loss or possible incompatibility issues that would necessitate it? I'm genuinely curious here. I can't see the upside.

 

[milksac]

Yes, the field order is irrelevent in this case. My work is for TV so I guess my brain is wired to spew worthless thoughts about fields.

The Meyers have been working with After Effects since the Cosa days. Long before pixel aspect ratio correction ever existed. Doing layout in a 720 x 540 comp was a workaround that some dinosaurs, like myself, still use. Regarding using the workflow now, I don't know. I've heard that PAR correction can cause some distortion, can't confirm or deny this myself because I don't use PAR correction.

"I've been wanting to play with true 16x9 footage, and their method seems a little odd. I don't get it.." -mg
Define true 16:9 footage. Footage is footage. For SD DV that means 720 x 480. The only difference is the optics and/or imager squeezes more information into that space. The Meyers method is replicating what happens in the camera - squeezing a frame size of 864 x 480 into a space of 720 x 480. Does that make sense?

 

[Matt Grunau]

Define true 16:9 footage. Footage is footage. For SD DV that means 720 x 480. The only difference is the optics and/or imager squeezes more information into that space.

I meant footage that was either shot with a widescreen adapter, or one that was shot using native 16:9 chips. Not something like Squeeze mode in many of our prosumer cameras.

The Meyers method is replicating what happens in the camera - squeezing a frame size of 864 x 480 into a space of 720 x 480. Does that make sense?

This is the part I cant quite get my head around (as far as workflow). The amount of pixels doesnt change, just their size, and how they are "squeezed" into a space. If the amount of pixels doesnt change, it is their aspect ratio that changes, with respect to square pixels. It seems the Meyer's method in replicating what is done in camera is redundant. Also, it seems by switching to a space with more pixels and then going back to the "normal" may even be slightly destructive, since there would be resampling/interpolation going both to the 864 pixel space and from the 864 space back to the 720 space.

It just doesn't make sense to me to do it that way, epecially since we have had widescreen pixel aspect ratios since version 6. Unless this was a way to include graphics and or elements that were only available as square pixels, but even then, squeezing them back into a smaller space would cause distortions. A perfect circle or square wouldn't look right.

That's why I don't understand the workflow.

 

[milksac]

... squeezing them back into a smaller space would cause distortions. A perfect circle or square wouldn't look right.

The circle wouldn't look right on a 4:3 monitor. On a 16:9 monitor it would display a perfect circle.

Your point about the resampling is a valid point. I'm not sure what's going on there with the potential for degration from scaling up and down. I do know that the Meyers have been doing this for much longer than I have - I find their knowledge and experience valuable. When I can find some free time I'll conduct some tests to see how the image holds up to all the resampling.