import videos and place based on length

[g4bs]

I work with my team in the area of video automation / video personalization. We create videos for events that have 1000+ guests and each of them will get a personalized video to go home with. That's usually no problem when changing text sources to place names, venues and so on.

Lately we planned to include videos created on location and render them into the project. That's where the problems really started:

• we were faced with different lengths of on-location clips
• these clips needed to be placed in the middle of the final video

We did a workaround that splitted up the whole thing: Rendering the first part with personalized content (names, etc.) - no problem. We then rendered the on-location part with some overlays. Finally we had the last part, again with regular text changes (names, etc.). Now we needed to stitch them together. That's where we really could use a good idea.

Finally my question:

Is it possible to stitch videos based on their length? Best would be a possibility to tell AE to place videos one after another and link different audio tracks on the depending video part. Because of the various lengths the audio wouldn't link anymore in the last part if the middle part is too long.

I hope, I could describe what's on my mind and maybe there are some pros out there that know a simple solution!

Hardware/Software we use:

• Usually we render the videos on multiple virtual machines (based on the number of clips needed for an event).
• We always use the latest version of AE / Media Encoder.

[Mylenium]

Sure, that's what scripting is for and various solutions already exist. Start by checking AEScripts.com.

Mylenium

[g4bs]

Will do 😉

[Arie Stavchansky]

Hi There,

Templater for After Effects includes a suite of techniques for handling variation of footage asset duration. It's called "Time Sculpting" and can be thought of as intelligent ripple editing.  One layer can change duration at the time of re-versioning and all sibling layers, including nested pre-comp layers, will adjust accordingly. So you can have voice-over audio clips predictably shift around depending on a layer's changing duration.

There are a set of rules that the end user configures to ensure layers respect the duration of the changing footage source. There are some tutorial videos that demonstrate the techniques for accomplishing this in a playlist on Dataclay's YouTube channel.

While this isn't a custom scripting solution, it does offer it as a packaged solution, and it may help your post workflow.

Hope this helps!

[g4bs]

We used to have Templater for long time now. 🙂 Great tool but we tool it to it's limits. Especially by using it on hundrets of virtual machines at the same time. But the Time Sculpting is new to me - I'll have a look on that! Thanks for that hint!

[Arie Stavchansky]

You don't need to run Templater on hundreds of virtual machines.  You can have Templater replicate the master AEP, instead of rendering out of AE, and pass that replicate file onto a machine that has aerender installed and run a script that executes the aerender command against the newly replicated project file.

[g4bs]

You can't imagine how many times we tried to solve that problem. And you just did it in one single sentence. You're  absolutely right - that's a workaround that could eliminate a lot of problems.

I will take that to my team and discuss it.

For future projects we surely will use Template once in a while. But I think we need to try a Scripting workaround aswell. 🙂

Thanks for the fast reply!

[Rick Gerard]

After Effects is not a video editing app. I can see using AE to create some of the inserts, but I would think that Premiere Pro would be better suited to editing the final product. PPro always renders faster than AE anyway. If all you are doing is inserting personalized shots in the middle of an already edited event video the solution may be as simple as a rendering package that combines videos by just stitching them together. 

If you have a folder with 1000 personal inserts and a beginning and end section already rendered scripting can assemble the parts, but I'm sure there are other ways to do that don't involve AE or PPro.

[g4bs]

Thanks for the reply!
We use AE because our inserts are text based (names, locations, dates, numbers, …) and images (JPGs that are moved in 3D space over the Corner Pin effect). In Premiere that's way more limited.

At the moment we use AE to render all the different parts of a video and then we programmed a tool to put them together and add the sound to it. In my eyes best would be to sticht all together in AE. That way I have the full control over it until the end and we could erase the step of our developers stitching something together.

I think Scripting in AE is the way to go. I need to have a closer look on that.

What do you mean with «there are other ways that don't involve AE or PPro»? I'm curious if I'm just too deep in Adobe tools and maybe I should think outside the box… 🙂

[Mathias Moehl]

Not 100% sure what your problem is. IN particular what you mean with "Because of the various lengths the audio wouldn't link anymore in the last part if the middle part is too long."

Do you mean depending on the lengh of the middle clip you need to pick a different audio file (like you have various songs as background music with different length and need to pick the one with more or less the right lenght)?

In general, with scripting you can create a new composion and insert your three parts as layers (at whatever time you want). But this  requires you do to write some custom scripts yourself (not sure if you currently use your own script or use solutions like Tempalter, Text2Spreadsheet or CompsFromSpreadsheet).

[g4bs]

Hi Mathias, thanks for your reply!

Imagine having three prerendered videos that are getting personalized in AE with images and text at the day of an event. Between these three videos we add another two clips that are getting recorded on the event. We now have five videos that need to be stitched together:

PreRender > EVENT CLIP > PreRender > EVENT CLIP > PreRender

The PreRenders always have the same length while the Event Clips change (sometimes 10s, sometimes 30s, …). The tricky part is to align the clips on a timeline and set the audio files in a way that if we have a speaker in all the prerendered Videos, his voice matches the movement of his mouth. So we need to have at least 3 audio files that always start together with the PreRender Videos and fade out after the following Event Clip has ended.

At the moment we use a tool we programmed ourselves to stitch together all these files. But in my eyes that's quite complicated. As I have seen in all the replies Scripting should do the job. I need to get a closer look on that.

Or do you have a better idea? 🙂

Thanks in advance!