Recreation of Vadi’s guards from tournament footage

I’ve been going through a lot of my old tournament footage recently, in order to create GIFs of various parts of the text. In the process, I stumbled across a moment where the camera angle and my position lined up perfectly in a way that reconstructed one of the images from the guards section in Vadi.

This got me wondering: could I reconstruct the whole section using nothing but screen shots taken from tournament footage? The short answer was yes! I’m particularly pleased because at no point in any of this was I intentionally posing to create an image; all of these spontaneously appeared in tournament. While not all of the images line up quite as perfectly as the first one, I think the fact that I could produce this relatively easily helps show how tournament fencing can be a perfectly natural complement to a historical, reconstructed martial art.

The link below will take you to the resulting document. For best effect I suggest you look at it in presentation mode. If you’re interested in producing something similar, I go through the method below.

Slides of Reconstructed Guard section

Production

Making this was actually surprisingly easy. If you wanted to make your own, it doesn’t require any specialist skills or software, just plenty of footage and some patience.

The images used in the slide deck are taken from a wide range of fights; the 13 images come from 9 different matches. The hardest barrier to getting good images was having the camera angle line up. I’m lucky in that I have a fair amount of footage from different events (around 150 fights). Maybe 30 of these had any chance of producing a suitable image. Thankfully, you can rule out unsuitable footage very quickly, only watching the ones that might be suitable.

Even then, it might not result in any or many usable images. Common issues I had were:

  • Guards were used while facing the wrong direction for the shot
  • Camera angle was wrong (head on, from a high angle)
  • Some guards were more frequently used than others
  • Guards sometines used transiently for less than a second
  • Body parts obstructed by judges or off camera
  • Circling matches constantly changing the angle
  • Variable lighting and image quality
  • Minor differences in detail between idealised guards and ones in use.

I had a higher threshold for how perfectly the images needed to match the originals for guards that I use more commonly.

Once I had the images, I used the windows inbuilt image editor to crop the photo, and then used its image enhance filter and changed the contrast to make details clearer. This was particularly essential on the lower quality images. Without it, the black clothing we use in HEMA can obscure details, although some images were fine.

After that, I copied them in to PowerPoint. This has a great feature for removing background in images. You need to give it a little help in deciding what to keep and what to ditch, but it’s a very simple process. Sadly it is consistently bad at distinguishing sword blades and quillons, which is why those look a little rough.

I then resized the images so that all the fighters were roughly the same size, and positioned them to be at the same point on the page. This gives a really interesting effect as you scroll through and allows you to really see the transitions between guards in some cases.

Finally, I added the text portions from my favourite translation.

All in all, the hardest and most time consuming part is getting the images. The image prep and compilation took about an hour – and with this, consider that I’ve never done this before, so really not much time at all.

I would absolutely love to see more of these, from Vadi or other systems. If you make anything similar please let me know!