YouTube on VHS

 


Hello readers, and viewers too, I suppose. I present to you my little school project for media channels & platforms. It's a partially animated 3-minute video on a somewhat fanciful topic. Simply put, if all videos on YouTube were put on video cassette tape, how big of a pile of tapes would you get?

Having dabbled in 3d modelling and video editing for a few years now, the main challenge for me in this task wasn't so much the technical aspects of producing the video, but moreso the problem of design and how to present information briefly, clearly and in an engaging manner.

Looking back, there are many tweaks and little improvements I'd like to make to the video, but I no longer have the energy for it. I'm still quite pleased with the end result, given the time constraints.

I'll give a rundown of the various steps of this production below, starting with the data:

1. The Data

My primary source for the data I needed to calculate an estimate of YouTube's total amount of video content comes from a Statista article that collects many YouTube metrics into one place. Among these is the rate of upload per minute from YouTube's launch to February 2020.

Figure 1: Upload rate graph

I tabulated these numbers into an excel spreadsheet and managed to calculate an estimate for the current total amount of video (in hours). This is all I need. Just as speed multiplied by time gives total distance covered, upload rate multiplied by time will give the total amount of video.

However, several assumptions and simplifications are made in order to perform the calculation:

  1. Since no specific dates are included, it is assumed that the data is collected from the start of each month. 
  2.  It is assumed that the upload rate is equal to the measured rate for the entirety of each time period.
  3. Third, after February 2020 it is assumed that the upload rate remains at 500 hours per minute, since no new data is available.
  4. Some content is removed over time, some is reuploaded. Without any relevant statistics on this, I assume that the impact of this activity is negligible for this estimation.


Table 1: Excel tabulations
 

Thankfully Excel can easily count the number of days between two dates. Looking at Table 1, for every interval

Uploaded hours = upload rate * interval * 24 * 60

Total uploaded hours is naturally all the uploaded hours added together. Since I chose to use a tape-length of 3 hours, dividing by 3 gives the number of tapes: 

589 860 000

The other significant numerical data for my project are the physical properties of VHS tape:

Dimensions: 18.8 X 10.2 X 2.5 cm
Weight: 230 g

 2. The Design

Armed with the numbers from my estimate, I could now start developing my scenes for the video. I didn't really do any sketching, but I had three distinct scenes in my head. I went straight into Blender and started compiling stuff together. 

First, I wanted a simple close-up of a single tape rotating. I found some good scans of vhs tape covers, mapped them onto a box and spun it around, next! 
Second scene. I wanted a conveyor belt spewing tapes at the appropriate speed (500 hours per minute, 2.7778 tapes per second). With some intense headscratching, I managed to get a rigid body simulation going that showed a row of tapes onto the floor. Good enough, moving on!

Texture-mapping in Blender

Third scene. Stacking a bajillion tapes with a human figure for scale. Uhhh, yeah... This turned into four separate scenes. The idea of the megatape emerged at this stage. I realised that trying to display several million tapes on screen would either kill my computer or just look like a smudgy mess. Stack a million tapes together and replace it with a box, problem solved. 

Folding up a megatape

I made one scene for stacking a million tapes next to a human figure, another for wrapping them up into a megatape, and two more for piling up the megatapes into a tower. The scale was dawning on me. Another point of reference was needed and the tallest building in the world quicly came to mind. I wasn't about to model the Burj Khalifa on my own. Lucky for me, several free-to-use 3d models of it were available to download.

a dude (180 cm)

Burj Khalifa (829 m)

The last scene I added was a tape falling on the floor that I used for the title sequence. All in all, I rendered 1162 frames of animation for this project. That's only 48 seconds. The rest was up to editing and padding it out with supplementary visuals.

3. The Edit

During the modelling I started imagining narrating the video and I ended up writing a script for the narration soon after. I recorded it and cut it down to three minutes. This gave me the 'skeleton' onto which I assembled the final video edit. The audio was recorded and trimmed in Audacity, then mixed with background music in Adobe Audition. The video was then edited and rendered in DaVinci Resolve with some added graphical elements made in Adobe After Effects and Photoshop.




Screenshots of various phases of editing

Since I didn't touch the audio narration after the initial cut after recording, it was very straightforward to assemble and time all of the footage around that. Freezeframing and looping got some more mileage out of my rendered footage. A few added images here and there, some stock footage and some basic transitions, and what do you know, I was finally done! 

Nothing left but to export and upload to youtube (and to write up this report, of course). My main takeaway from this project is an appreciation for the amount of elements and details that one needs to keep track of when producing even a short data visualisation video. Also, when working alone it is important to have resources and assets that reduce workload and speed things up where possible.

4. Tools/Credits/Sources

Adobe After Effects CC
Adobe Audition CC
Adobe Photoshop CC
Audacity 3.1.2
Blender 3.0
DaVinci Resolve 17
Excel
MB-Lab addon for Blender by MB-Lab Community

Background music "Synth Vibe" created by Mativve from Freesound.org
Burj Khalifa 3d model by ManySince910 - Sketchfab.com
"Quarry 3" HDRI image by Sergej Majboroda - Polyhaven.com

YouTube. (February 14, 2020). Hours of video uploaded to YouTube every minute as of February 2020 [Graph]. In Statista. Retrieved December 05, 2021, from https://www-statista-com.libproxy.tuni.fi/statistics/259477/hours-of-video-uploaded-to-youtube-every-minute/




Comments

  1. I love the data and the animation is really well done! Good job!

    ReplyDelete

Post a Comment