[Justin Mrkva]

It’s 30fps, and so is the project. I did find that apparently iPhone video can actually have variable frame rates. Not sure what the details or the implications are, but I suspect that has something to do with the Optical Flow issues prior to Optimized conversion.

I’m in the middle of writing a new application for video processing, so I’m actually dealing with frame rate handling right now. I’ll try running that video through once I get that straightened out and see what actually happens with the timecode. 😀

[Justin Mrkva]

Do you mean like this?

Source: ABCDEFGHIJKLMNOP
Final: AAAAEEEEIIKKKKKP
(where you control each individual frame A/E/I/K/P)

I’m not familiar with the FCP6 Strobe effect so I’m not sure if that’s what you’re after.

There are two ways to do what you’re looking for – a simple way, and a complex way. First, Motion has a Strobe filter (Library-Filters-Time-Strobe). Create a FCPX Effect, add Strobe to the source, publish Strobe Rate and Mix, and you’re done. This unfortunately doesn’t give you precise control over the frames, but by controlling rate you can get close.

If this doesn’t give you enough control, the best option is to bring the footage into Motion and keyframe it manually. I’m not aware of a way to publish those keyframe controls to FCPX, unfortunately.

[Justin Mrkva]

For more info on clone layers, MacBreak Studio episode 116 explains a lot of the details. I’m sure you’ve seen that one, Mark. 🙂

https://www.pixelcorps.tv/macbreak_studio116

[Justin Mrkva]

It’s not a DSLR – it was shot hand-held on an iPhone 4s. 🙂

And yeah, I agree, I’ll just optimize the clips I need to. I’m probably going to be getting my RAID system back up and running, so at that point going optimized shouldn’t be a space problem.

While I’m at it, what kind of drives do you use for mobile editing? I have an OWC Mercury Elite Pro Dual Mini with twin 7200 RPM 750GB drives, RAID-0 striped, but on the Retina MacBook Pro (brand new!) the Apple FireWire 800 adapter doesn’t provide enough wattage to power it. I’ve got a USB3-eSATA adapter, but unfortunately I need a wall wart to power the drive itself since the adapter’s FW connection can’t. Any thoughts? I don’t want to spend any significant money right now; I’m waiting for OWC to update their lineup with some Thunderbolt RAID enclosures or something that I can power from the rMBP.

[Justin Mrkva]

I thought I had figured out a way to do this. I take the new clips, add a “dummy” clip at the end of the project, then export the project as XML, swap the IDs of the new and old media, then import. It does successfully swap which clip is being referenced.

However, there’s still one problem.

I have, for example, File A and File B. A is old, B is new. They look like this:

A: ABCDEFG
B: XXABCDEFG

where B has some extra blank time before the beginning. I edit the Event clip in the Timeline so that A and B match, according to FCPX. I then export XML, swap A and B by IDs, and it should be good.

Except it isn’t. When exporting to XML, FCPX writes in the start time of the clips. So suppose I have a project like this:

[ABCEFG] where I use the beginning and end of A and cut out the middle. The XML treats that as (for simplicity I’m using whole numbers and a 1-index) Clip 1 starting at 1 and Clip 2 starting at 5.

When I swap them, then import, what I get is this:

[XXACDE]

In other words, even though I cut off that first section of the clip by editing the Event clip in the timeline, FCPX’s XML export works by referencing the start point not in the Event clip, but in the original media file. I can see why it does this, but it frustrates my efforts to swap these clips properly. I could try to write a script, but really, this is something that should be done by FCPX, not through an XML hack. Furthermore, even if I did write a script, this trick would only work if the clip only differed at the beginning; if I had to cut some frames from the middle and that section happened to be within the clip I was modifying, it would bring about all sorts of problems.

So I’m still searching for a solution, but I feel that I might have gone beyond the capabilities of the software here.

[Justin Mrkva]

The problem isn’t resolution within projects; it’s that there’s already several projects using the old footage, and several hundred edits across multiple projects each using that media, coupled with transitions and keyframed effects. Replacing the footage has a lot of difficulties associated with it, because if I just delete the old clip and create a new one, I’d have to manually re-edit the new clip into everything. I need to somehow make FCPX recognize the new clip as an identical clip to the old one, just with a different resolution, which should be handled by the spatial conformation. Instead, FCPX doesn’t seem to be able to do this like I want to.

This is what I tried:

1. Bring in new footage as new clip, ignore old footage and just move it to an “archive” event that I keep around; I’d just use the new footage for new edits. BAD because I have to keep all of the old footage, taking up a lot of space. Re-editing to remove the dependence on old footage would take forever.

2. Bring in new footage, then open original clip in timeline, connect edit the new footage to the old, then disable the old footage. Missing media for the old footage shouldn’t be a problem there because the video used will be the new footage. Theoretically could work because clips would automatically reference the new media because it would just be connect edited into the existing clip that the projects were already using. BAD because FCPX still recognizes the whole clip as the resolution of the old clip, not 1080p like the new footage.

3. Bring new footage in as new clips, edit them to match the old footage. Then modify the clip reference for each edit in the projects to point to the new clip instead of the old clip. This is a problem because I can’t find any way to modify a clip reference, only an event reference.

Number 3 would work, and though it would be some manual work, it would be feasible, and eventually I could modify everything so that the clips would all be using new clips. Fairly manual, but acceptable for my situation.

The problem, as stated, is that I can’t replace existing clips. Even if I manually selected in points, then selected “Replace From Start” when dragging in to replace the clips, this loses effects, parameters, grading, retiming – all things that would take a long time to recreate.

It all comes down to this problem – batch replacing huge amounts of original footage with higher resolution footage without losing event and project references or any effects/modifications applied to the clips. At this point I think I’m going to go with the “create new clips and just keep the old footage” approach, at least until I can figure out a way to properly replace the footage.

If I can figure out how to take a clip in the timeline, select it and tell it to reference a different arbitrary clip while maintaining applied effects, or even better, to do that for all instances of a clip, that would solve my problem.

Summary, if you aren’t confused enough already (lol):

Event contains Clip A and Clip B. Several projects use parts of Clip A, I want them all to use the same parts of Clip B instead. Wherever Clip A is used in a project, it has complex effects and edits that are difficult to duplicate.

[Justin Mrkva]

As mentioned in this thread, I managed to solve this for good; you can set a 3D Group to use Layer Order. That ensures the overlay appears properly and doesn’t get cut off by the floor. 🙂

[Justin Mrkva]

Aha. That makes sense, although I wish rasterized 3D objects could cast shadows. I mean, why not? The thing about Flatten is that it causes a loss of 3D perspective, since it treats the layer as a projection from a camera positioned directly in front of it. Something like Flatten With Perspective would be a great option, essentially telling it to Flatten the layer, but using the active camera.

So I thought to use a normal mask, since the image mask apparently is what’s causing the rasterization. However, here’s the problem with a “normal” mask – that text layer is really a placeholder for a 3D group that I have in my real project. AFAIK (and I tried) you can’t apply a bezier mask to a 3D group, only an image mask. If I convert to 2D I lose the effect. That said, I’m not even sure if a bezier mask would cause rasterization; it might, given the 3D perspective involved.

However, I figured the whole thing out, INCLUDING the layering issue I had with the overlay cutting off on the floor (mentioned in another thread).

Here’s the layer structure I ended up using:

Logo Group
—-Flat Logo (clone layer)
—-Replicator (for 3D effect)
——–Flat Logo (replicator cell)
—-Flat Logo
——–Image Mask
——–PSD file
——–Misc. layers

First of all, I can’t apply Flatten to the Logo Group because it destroys the 3D effect generated by the Replicator. Because of the way they’re stacked, though, I could apply the Image Mask to the Flat Logo group and flatten only that layer. Then, the Replicator is set to use Flat Logo, and the Clone Layer is also set to use that group as well. Shadows are cast only from the clone layer, since casting multiple shadows from all the replicator cells doesn’t look very different and slows down the render… a lot. 😀 The end result is essentially the same visually as applying the Image Mask to the Logo Group, but allows me to Flatten the Flat Logo group, eliminating the rasterization that comes from the Image Mask.

Secondly, in using the Flatten option, I discovered another option – Layer Order. In another question, I had a floor and a 3D positioned layer that was getting cut off by the floor but that I didn’t want cut off; I wanted it to appear on top, rather than obeying layer positioning and clipping. Layer Order does just that. Incidentally, the same Image Mask bug that was causing shadows to fail was also causing that layer to appear on top before, something I was exploiting to get it to appear how I wanted, so finding the Layer Order option really saves me here as well.

Also, yes, 3 shadows is a lot, I hadn’t softened them yet. I can’t reveal much about the project at this time, but I can give you a small screenshot that shows what the shadows look like in the final comp:

Thanks for the help! 🙂

[Justin Mrkva]

The latest version comes with new Macs, otherwise you’d have to get it from the Mac App Store… you might have to go back to your install discs, but I’m not sure what version you’d have.

If you can’t use iMovie, the next thing I’d do is try the FXFactory plugin, then do it yourself as a last resort. Also, look up the MacBreak Studio podcast, there’s one where they do a 3D globe, you could probably base it off of that to some extent.

[Justin Mrkva]

I would use a text behavior, but in the final version we have a PSD logo file that’ll be replacing the actual text, as well as a replicator and some other effects. I’d be interested if there’s a way to do a gradient wipe some other way that’s compatible with full groups, but this works pretty well.