After spending a few days learning how to use a video production tool (HitFilm), I managed to trim and cut synced videos of dad’s first batch of field test videos.
This particular video production tool, while easy to learn, caused some confusion earlier when I was trying to export the video files properly. There were so many settings for exporting videos and I just went blank (bit rate, frames per second, resolution… what?). Thankfully, the video production tool allows me to save my editing work as project files before committing to export them out as finalised videos.
To be honest, each session’s video took a few steps before making it onto the finalised cut. One of the issues I encountered was that the GoPro cameras, that I am using for data capture, outputs separated files once each video file (on the camera itself) reaches about 4 GB worth. That means I have to stitch all these separate files and combine them into a singular video file (plus fisheye correction rendering) on the video production tool.
I was rubbish with this step at first. So I blew an entire weekend trying to render all four sessions of videos from each camera – for a total of twelve videos (three cameras, four sessions) each averaging a total render time of four hours. And then, I realised I screwed up somewhere… and had to redo everything again.
Frustration, yes. But I think I learnt how to use
this video production tool on my own. =_=
Once those twelve videos were done, I had to spend almost a day each to eyeball and sync two videos from each session (there are four videos, from the perspective of the overhead camera, which are not used for data capture). This was where most of the editing and trimming work are done – once the project files were ready (averaging about two hours to trim and sync as a project file), I rendered them as high resolution video files for archival purposes. The final step is to queue them into another tool (Handbrake) to cut the file sizes without losing too much quality.
For these four sessions, the video production process took almost took an entire week to complete (not including the foul ups earlier). I guess I would much better prepared for the next twelve sessions!
While waiting for the next sessions to start again, I am currently tabulating the data captured by the session videos. Each video averages about 40 minutes in duration, but tabulating each session video takes about four to six hours (watching the videos from start till end repeatedly, as I need to collect information for about seven data points).
While assessing the data capture, I realised that there are ways to improve the data capture. However, for the sake of consistency, I would not be improving the data capture but discuss about the improvements in the final research paper itself.
Another artist has caught wind of what I am
doing… and wants to be documented as well.
More data for the data god!
A friend who has been working late for a few days in a row, managed to send me some messages late at night (probably not knowing whether I am still awake). For some reason, this friend’s messages kind of encouraged me to work harder (and late through the next day) for some reason.
When the crates come, couriers and curators
are really dynamic people to work with.
Probably won’t have much update on this until the field test session start again (they are based on different conditions) when dad is back from overseas.
In the meantime, ボンジュール鈴木「羊曜日に猫ごっこして」