The channel is already controversial because of its redaction tactics, and it comes as a presidential task force about the nation's policing recommended that police wear body cameras. The body cam channel features seemingly Soviet-era-like footage, and it's already being criticized. That criticism comes from Tim Clemans, the Washington state computer programmer who is redacting or blurring the video and removing voices on behalf of the police department."I'm having to work with people who don't want this released period. There's a number of people trying to put up every obstacle possible," Clemans said. Clemans is doing it for free. The redaction surgery usually takes about one minute per minute of footage, he said. He runs it through "five lines of open source code."
The agency is redacting more from the footage than what's required under the state's public record laws, he said.
"The department does not want to post raw video on its YouTube channel. It fears a privacy controversy," he said.
The department is burning as many as 7,000 DVDs monthly to meet public demand for information. The agency has more than 1.5 million videos taking up 364 terabytes. The footage includes dash cam video, 911 responses, and "interviews with victims, witnesses and suspects."
Clemans understands that the agency can't keep up with demands. He said that public disclosure through the YouTube channel is a "middle ground" of sorts. [...]
Clemens hooked up with the Seattle Police Department after he rattled the agency last year by filing more than 30 public-information requests on every 911 emergency call in which officers responded. The demand included dash cam and body cam footage.
I expect to see this stuff used as raw material in some Pattern Recognition-style videos. Please get right on that.
Previously, previously, previously, previously, previously, previously.
You know, if that is actually a pure gaussian blur, then you can undo it by deconvolution.
You would need to do some trial and error to find the parameters he used, but if you got the right numbers then in principle you could recover the original video from the blurred version. In practice you would have some loss because of the noise introduced by quantisation and recompression after the effect was applied.
Help, help, I'm being defocused
I ran SmartDeblur against some screencaps. Couldn't do anything useful with it.
Needs that teal print from ~ Mar 29 on the DVDs. Really, people are asking for data and not asking for a particular flash memory (ZFS, UHS-1, nothing but net, etc.) or a more apt use of 5 lines (like maybe NOT handbrake and floovex for example.)
Next they'll release it but with public officials frozen at 10m/s momentum instead of a particular camera location. Only Matthew Barney is admissible!