AI "Enhanced" Walk From OCAD to University and King to Wellesley

I've been playing around with AI image and video enhancement tools - shooting footage beyond what is reasonable to fix, then setting the sliders on the various software tools too high, then enjoying the blurry smudgy wobbly weirdness.

I put my 14-30mm/4 on my Z6 and set up a timelapse sequence, shooting one shot every half second. The lens was stopped down and the ISO limited to 100 to increase motion blur.

Starting from OCAD, I walked along Dundas to University, Down university to King, then Eastbound to Yonge, then I walked up Yonge to Wellesley Station. I managed to get about 5500 shots on a single battery!

The hyperlapse sequence was rendered out as a 720x720 24p video, then I ran it through Topaz Video AI, with 6X frame interpolation, 2x upscale, 100% stabilization (3 passes), and some heavy sharpening. A saturation adjustment was made right before the final render to add some punch, since the last AI walk video I posted was so dichromatic.

The audio is from the walk, but it has been squeezed to fit the length of the video. This was my second opportunity to use my fancy new Rode Wireless Pro system for something, and I wish I brought the windscreen because it was an awful day for capturing audio. The 32-bit float audio allowed me to capture everything from quiet footsteps to obnoxiously loud sirens and wind ruffle without any clipping, and I was able to apply a lot of dynamics compression so you can hear it all in the final mix.

http://kyleclements.com ... https://www.youtube.com/watch?v=B6om8c3wDtk