"Scale to Frame Size" versus "Fit to Frame / Fill Frame"

[KazuT]

I have a project where all the shots have "Scale to Frame Size" selected. The footage is all kinds of dimensions, from 640x360 to 6144x3240. 

In order for resize and reposition keyframes to translate to our color vendor correctly, don't I need to undo all the "Scale to Frame Size" in the sequence and manually resize and reposition everything? Or is this unnecessary?

I undid "Scale to Frame Size", then selected the clips on V1 and selected "Fill Frame" (as a lot of it is 1.9:1, that seems to give me a better head start). However, for some reason I don't understand, it seems to have deleted all my keyframes (although the initial repositioning was maintained). Any idea why this would delete my keyframes?

[R Neil Haugen · Accepted solution]

First, changing a scale method requires complete re-computation of the image. Hence keyframes for changing are no longer applicable.

Second ... there is a massive difference between scale to and set to framesize.

Scale to framesize takes the image pixels of the original image, and computes the "correct'" rasterization of that image to the new framesize. And ...  any further change on that image comes from the rasterized new image pixels. NOT the original image pixels.

So if you "scale to" to a smaller frame, and then want to 'zoom into' the image, you are not dealing with original image size, but the smaller image of the use on that sequence. You may not want to do that ...

Set to framesize takes the original image, and does the same operation as if you simply adjusted image size in the Effects Control Panel ... and will always work from the original pixels. 

From experience dealing with colorists, working in Resolve, "scale to" can be problematic in pass-through workflows in XML or EDL processes. "Set to" is generally preferred.

[KazuT]

Thanks @R Neil Haugen.

So, as far as being able to make a good turnover to color, I'm not wasting my time by un-"Scaling to Frame Size" and using "Fit to Frame / Fill Frame" instead? I want to make sure I'm not wasting my time, as this is taking several hours to do.

[R Neil Haugen]

Scaling can be an issue in turn-over workflows.

For most colorists, Set To tends to pass through, although not 100%. Scale to is problematic most of the time.

So the colorists I know all prefer Set to to have been used in Premiere ... but ... need a flattened exported low-res H.264 file of the entire sequence so they can lay that in an upper track, and check all cut points and can check for any scaling changes done. Time ramps and such.

Conforming a project is generally a bit of a pain.

[KazuT]

Thanks @R Neil Haugen. Yeah, I'm trying to make it one bit less of a pain for the online editor/colorist by spending the hours correcting this rather than leaving it up to them. I'm having to calculate digital pushins since the archival footage is various different dimesions; but I'd rather I did that (as an Assistant Editor) rather than leave it for the online editor to have to do.

[R Neil Haugen]

That stuff is always a right pain in the tushie ... although some things make it worse.

Such as the colorist is also slated to create the deliverables ... and ... the editor or an assistant love speed ramps, and has put a ton of them in a project as they look cool. Speed ramps ... never ever ... make it through the conform. So the colorist or an assistant has to first find every one of the things, then from watching that H.264, try to match (really, re-create) the speed changes, in Resolve or Baselight, whichever they are working in.

Which always gives a great opportunity for someone to say "that isn't specifically what we delivered to you ... ". Last thing a colorist or finisher wants to hear, of course.

Or keyframed "punch-ins" ... another joy.

Yea, there's plenty of misery to share around here ...

[KazuT]

I'm running into a lot of keyframed punch ins where the keyframes are outside the IN and OUT points of the shots. Fun 🙂

[R Neil Haugen]

Fun ... not fun ... sigh.