isomer.nula.cc

Suddenly

Video. Three short sequences of camera shots from the film Battleship Potemkin are reduced to their individual frames, and each becomes a part of this filecast. The frames of I. are directly altered, while II. and III. are re-ordered or interleaved according to mathematical patterns chosen at whim.

The title comes from the intertitle that immediately precedes a brief sequence in Potemkin of a young black-haired woman in a striped blouse, which opens the famous and much-celebrated Odessa steps sequence from the film. The title, in Russian, reads и вдруг (i vdrug), which means “and suddenly.”

The entire sequence of four camera shots comprises a scant 24 frames (Based on my DVD copy of the film, leaving aside the frames that are duplicated to bring the original 16 or 18-frames-per-second camera original up to the modern sound film standard of 24 fps). This is less than two seconds of screen time.

The first shot is only five frames of tossing hair with no part of the face clearly visible. The second shot is also five frames, this time, her head and torso are thrown back quickly, either in shock at the sound of rifle fire, or perhaps because she has taken a bullet. The third shot consists of eight frames, as her head lurches forward and then backward again, as emotional shock registers on her face. The final six frames cut from the previous closeup to an extreme closeup of her facial expression seen thru her disheveled hair, as she continues to lurch slightly toward the camera. It is a tiny masterpiece of rhythm, composition, and meaning, and a microcosmic representative of what makes Potemkin such a revered work.

Addendum: I was asked how these were made, and so:

I’m not using anything special to resequence frames; basically, brute force. I split the video file into stills, saving each frame as numbered png file. If I feel like it, I can retouch the individual pngs just as you would any photograph.

I use a file renaming tool to renumber the png files any way I want. In the end, the pngs will always play in temporal order by their numeric sort position, so renaming (renumbering) them is all it takes to change their playback order. I use a javascript I wrote, or a perl script that I found, to generate lists of numbers according to patterns that I choose. The file renamer takes a list as an input and renames the files in a selected directory in order, following the list. Care is required so as not to create conflicting file numbers, which might overwrite content unintentionally during the renaming process.

So, for example, by taking a directory of png files, duplicating it en masse, and then giving the first directory of pngs all even numbers, and then reversing the sort order of the other set and renumbering them odd, you can then unite the two sets and derive an interleaved video that plays forward in every other frame, while simultaneously playing backward in the others. Once you grasp this, it’s straightforward to dream up ways to create sets of sets of images, renumber them, and then unite them into one big set, with a complex interleaving pattern being the result. The final step is simply to import the final image sequence set frame by frame into a video.

I have a sort of philosophy behind doing it this way. If these manipulations were as easy as clicking a button in some interface, then one needn’t really consider too deeply why one would make such a choice. Because my method is slow and deliberate, I am considering – consciously or not – each step in turn, and reëvaluating the envisioned end constantly thruout the process. So my end result must be different from an result achieved thru “easier” (not necessarily “simpler”) means.