Data driven personalized videos – comps of varying length

[rolandw6993648]

Hi;

I’m in a bit of a fix creating personalized videos using AERender.

To simplify my question and the problem, assume I’m creating personalized videos that explain each line item of an invoice to the customer; each line item is a comp.

For each line of data, assume:

• there could be 1-10 line-items per invoice
• the voice-over is in the language choice of the customer and some line items take longer to explain than others

Here’s where I’m currently at with this;

• I have a project file with 10 comps, all of equal length. Each comp has a MP3 voice-over placeholder.
• For each line of data:
  • The application calls AWS for translation (if required)
  • The application calls AWS Poly for text-to-speech MP3
  • The application launches AERender to create a video from the project file.

This all works fine, but is not satisfactory because the length of the MP3 voice-overs vary based on the complexity of line-item and the language.

Question:

-------------

Here’s where I need to go:

When AERender launches the project, I need the length of each comp to be based on that of its voice-over; then, I need the in-point of each comp to be based on the out-point of the previous comp. I believe  this can be done in a script, but don’t know how to execute a script automatically in a project when it is launched with AERender.

Thanks in advance for your help.

Roland

[Mylenium · Accepted solution]

You can't and that is that. Scripts only work in the full AE app, not AErender and it's logical - AErender really just does what the name implies. Whatever scripts you fancy need to open up a full instance of AE, replace the sources, adjust the comp durations and then save the project. Your only other alterantive would be setting render ranges via AErender's command line options, but you'd have to do the math where each comp begins and ends yourself.

Mylenium

[rolandw6993648]

Thanks Mylenium;

It looks like I'll have to use a script for the composition and render; my 'wrapper' application will manage the data intake , massage and error check it, send status updates to client,  translate (if required), synthesize voice, deal with the distribution of the rendered file once AE has created it and handle the billing.

BTW,  I meant to use ‘layer’ and  not ‘comp’ in the long intro to my dilemma – but doesn’t really make a difference I suppose.

[Arie Stavchansky]

Hi Roland,

With aerender it's not possible as Mylenium stated.  You would need to manipulate the timeline itself which requires the use the authoring tool, not just the rendering engine.  

I have to chime into this thread because the problem you are dealing with is one that I tought about for a considerable amount of time and eventually implemented it into Templater which is an extension for After Effects.  At Dataclay, we architected a suite of techniques, called Time Sculpting, that allows you to achieve this very thing.  Take a look at our videos on Time Cropping rules and Time Shifting rules to get a better sense of what this patented algorithm does for personalized videos.

Thanks!

[rolandw6993648]

Thanks Arie;

Yes, I checked out Templater and DL’ed the demo just a few days ago. It is a very comprehensive package, but the only part I need is dealing with the dynamic length of the audio and subsequent adjustment of layer lengths.

As I mentioned in my reply to Mylenium, I’ll start off with a script.  The script work-around will suffice until I become familiar with AE’s SDK and program what I need.