Unauthoritative Pronouncements

My post for Six Colors yesterday is a video demo of Clean Up and other photo retouching tools for iOS. It’s not an embarrassing video, even though it isn’t exactly how it turned out.

I don’t like to do anything at all with video, which is ironic considering I’ve been a visual effects artist for film, TV, and commercials for 19 years. That’s different because someone else has to set up the shot, set up the camera, manage the files, record the audio, conform, cut, color, and export. The little part I do for VFX is near the end of the process, and as critical as I can be of choices other people have made when shooting a thing, I can be even harder on myself.

If I’m so bad at it, and so unhappy with the results, why go through all the fuss? It’s typically not worth it, but words can only convey so much.

I previously made a little video demo for this post on Shake, and that was easier because I knew I wasn’t going to be on screen very much and that it would be mostly voice over with screen recordings, to supplement what I was writing. I personally don’t like videos that are exclusively video with voice over because it feels impersonal in some weird way. I’d like to see who’s talking at some point.

When I recorded that Shake video I could show that I wasn’t trying too hard by using my Logitech C920e webcam —the webcammiest of webcams— to get the little snippet of me heaving the box of Shake 4.1 software aloft. See, look at me, I’m not a try-hard making video about compositing software! I’m hip, daddy-o!

This time, however, I couldn’t quite get away with that. I was going to be demoing software on a 9:16 iPhone screen, which meant I had three options:

• Make a portrait 9:16 video
• Make a landscape 16:9 video with a portrait 9:16 video centered in the middle over a generic background
• Make 9:16 video where I was on part of the screen taking up the 16:9 landscape portion.

I chose the last option, but that meant I had to be on screen the whole time. I didn’t have a lot of options for other footage to cut away to. I know some people on YouTube will cut to the iPhone in their hands for B-roll coverage, but I don’t have another suitable camera to use for that, or place to overhead mount one. I have a lot of cameras, but only one suitable for recording video, my Sony a6400. That’s barely passable by modern standards, and in many ways my iPhone 16 Pro is a better video camera —but that was the device I was working on so it wasn’t an option.

Knowing that I had one angle, I figured I would frame it with the vertical video on the screen left-ish third of the frame. My animated hand gestures are visible flicking to the left every now and then but I thought it looked weird if I crammed the screen recording all the way to the left edge of the frame.

I had my script, but I couldn’t read from my script because it looked like a hostage video where Clean Up was holding me for ransom. Instead, I tried to get the gist of what I wrote, even if it wasn’t exact. That also meant I had to do many takes. I really didn’t want to do a lot of jump cuts, but since there was no coverage, and there were occasions where technical issues came up (camera overheating, screen recording just randomly stopping) jump cuts would be unavoidable.

The audio was recorded with my Shure SM5B through my Elgato Wave XLR to my Mac. The mic was on an arm I tried to keep low in frame. The Sony recorded audio on it’s mic, but that was what I used to align my higher quality audio to.

My camera was on a Joby GorillaPod I’ve had for years balanced on top of the box my Elgato Wave XLR came in (yes, I keep the boxes my electronics come in, sue me). The lens on the camera was my Sigma 19-50 F2.8.

The a6400 can’t record any higher than 8-bit video. This put me at a disadvantage when it comes to flexibility in exposure and dynamic range.

Fine, whatever, but what I forgot to do was turn off auto white balance and auto exposure. Rookie mistake! The a6400 kept jerking the exposure around not only from the bounce light coming in through the windows, but from the whites of my eye or teeth being aimed at camera. I didn’t want to reshoot the whole thing, so I did what anyone else would do and tried to use DaVinci Resolve’s color stabilize.

Unfortunately, that’s locked to DaVinci Resolve Studio which is $299 and out of the budget of $0 so I went back into Premiere Pro and hand keyed the exposure changes every few seconds. There is lingering light fluctuation, of course, but the big swings in exposure correction are gone. Bespoke, hand-crafted, color-correction isn’t going to fix that.

I didn’t have enough room to work on my Mac’s hard drive so I did it on my external spinning disk that I mostly use for archiving old podcast recordings. I was worried that the project would be a stuttering mess to work on, but Premiere smartly cached enough material that I would playback at speed.

Oh, I forgot to mention the part where I recorded wrote this Tuesday, and recorded it Wednesday. Then I tried to make that edit work Thursday and Friday, before I scrapped it and re-recorded everything Sunday morning, then scrapped that, and recorded again Sunday afternoon. Yes, I recorded it all of these times and forgot the auto white balance and auto exposure every time I did it. I’m not very smart.

That last recording session Sunday is what makes up the final video. None of the earlier takes were used, but because I performed the actions, and said the words, so many times there are instances where the words I use seem to indicate that I had done something before (because I had) and the two times that I say, “last example”. Which, if you’re counting, is not how the word “last” works.

I have a greater appreciation of the “talking head” videos that YouTubers put together regularly. I know that if you do something repeatedly things get easier (except turning off AWB/AE) but speaking into camera and then having to review how you spoke to camera is draining.

Writing about these examples of using the software really doesn’t work as well as video, so I think that this was necessary to go into the detail I wanted to, but I’m not envisioning a new career of making videos instead of writing things. Text is significantly more forgiving.