Which is best for compositing HVX200 960x720 1.33 footage in After Effects?

[jesse]

Which is best for compositing HVX200 960x720 1.33 footage in After Effects? Square or 1.33 pixel aspect ratio workspace? The final output is for D1 SD but also want to preserve for future HD resolution i.e. 1280 x960 square resolution.

Here is the step by step example comparison:

1.33
1. Import 960x720 1.33 footage (leave interpret footage as is)
2. Compify clip, creating a 960x720 1.33 comp
3. Composite and work in this comp. Use Pixel Toggle aspect ratio correction to temporarily display for square viewing without compromising footage.

[Pro's: 1.33 toggled space keeps the footage completely untouched until final output. Cons: displays with slight artifacts on edges stair-casing etc. This may make it difficult to cut a matte.]

Square:
1. Import 960x720 1.33 footage (leave interpret footage as is)
2. Create a new comp 1280x720 Square.
3. Drop in footage in new square comp.

[Pro's: you can see your footage clean with no artifacts. Con's: you have compromised your footage. You will then need to stretch again to D1 ]

A composite package like Fusion apparently doesn't have any temporary display artifacting while in the PAR correction.

Does anyone have any recommendations for the optimal workflow?

[Matt Grunau]

I've always preferred to work with the video's original aspect ratio. If one of the "cons" is display, I would look into enabling Preserve Edges or something in program to get rid of them.

Or, if you want to cut a matte, use your footage in a square pixel comp, create your matte and then bring the matte and whatever else back into the 1.33 composition for use with your original footage. Use the best of both. Get the clarity, setup, and control from Square, but render from the 1.33.

[Arson]

There is no artifacting of HVX footage in AE, squishing your footage down to 960x720 instead of leaving it at 1280x720 is whats stairstepping it

I import HVX footage into Premiere and After Effects and in a year or so of shooting with the HVX I have never seen any of what you are experiencing.

Import the footage into AE project then drag the clip on the make composition icon and you will have a comp set perfectly to the correct resolution and PAR for the HVX footage.

[jesse]

Matt, I agree original aspect ratio would be best more most lossless. I don't see a preserve edges function. If there is there is no problem. If you find that hidden somewhere let me know. I believe its display a design flaw in AE. Fusion is able to display square toggle without artifacts.

Pre-comping the matte is an interesting recommend. I just tested a matte spline in a square comp and that copies perfectly into the 1.33 comp. I assume other additional elements can then be comped in square and copied to original 1.33 for final output.

Noah K mentioned in my post to the HVX200 workflow section... "I'd go 1280- better quality and much easier to view in AFX as properly unsqueezed. It will scale back to match perfectly in FCP." - Noah

The big question is there loss in vertically un-squeezing 1.33 to square and then horizontally un-squeezing (downresing) for an SD output. For SD it may not make much difference, but I am trying to simply understand all the options and why.

[jesse]

Arson, thanks for your reply. I believe the HVX200 footage shot at 720p = 960x720 1.33 PAR. That is the format the camera records in. If you look at .mfx files that have been converted or "wrapped" as quicktimes they are 1.33 par. The camera does not shoot square pixels.

You mentioned "squishing your footage down to 960x720 instead of leaving it at 1280x720 is whats stairstepping it" I am not squishing it. That is the original file's PAR 1.33. When you drop comp a clip-- it creates a comp 960x720 PAR HDV/1080/DVCPRO HD 720 (1.33) When you view it and it doesn't looked squeezed it is because toggle pixel aspect ratio is turned on.

Maybe I misunderstood you. Or if I have this wrong someone please correct me.

[jwiley20]

the artifacting drove me nuts at first in AE. you're right to create a square pixel comp and conform the PAR 1.33 footage into it. make sure you interpret your footage as 1.33 in the project palette though:thumbup:

[oneinfiniteloop]

Always work in square pixels, much easier and there is no loss in resolution/quality. I've tested this extensively for the workflow our company uses and have used over the last year. If you're going to D1, import your footage, drop it in a 1280x720 square pixel comp, work as usual, then drop that into a 720x540 D1 square pixel comp and render out D1 with .9 PAR or a square comp, whichever you need. We'll usually render out a 720x540 comp and dump into FCP and lay off to Beta.