> On an ordinary Windows laptop, Saldana used the command line to invoke a utility that created a compressed file from my original video file.
> the compressed file size was about 4 KB.
> On the same Windows laptop, Saldana invoked another utility that uncompressed the 4 KB file. Uncompressing took about 10 seconds. The uncompressed file was identical to the original file.
Occam's razor: the "compressed" version was a link (shortcut, symlink, hyperlink, whatever) to the original.
> Despite maintaining a healthy skepticism, I have two reasons for believing the files were identical. First, Saldana played the uncompressed video on the Windows laptop. The uncompressed video looked exactly like the original video. That was nice, but it didn't prove anything. I wouldn't be able to identify an unfaithful video copy if it bit me on the nose.
> The second proof that the original and compressed files were identical was much more compelling. Saldana put the uncompressed file on my own USB drive. Later that day, on my own computer, I ran the uncompressed and compressed files through a bit-by-bit comparison program. Sure enough, the two files contain the exact same sequence of 0s and 1s.
> It should be noted that for security reasons, Saldana didn't copy the 4 KB compressed file to my USB drive, so I don't know what bit sequence was inside the 99.78% compressed file.
All of this is consistent with the file being a link.
A more thorough test would have performed the compression and decompression on two different, non-networked machines, so that the only information available to the decompressor is the 4K file.
After that, we need to start removing layers which can be used to obscure the results, eg. performing a more low-level analysis of the compressed version, using machines which the authors have no access to (eg. to avoid a hacked OS/driver which reports false file sizes)
This is a pretty serious scam, and not the first that revolves around video compression. Barry Burd ( is that his real name?) is just a clueless tech journalist without the background or skepticism to realize how many times this scam has been launched. Some examples.. Madison Priest and his "magic box". David Kim Stanley and Pixelon. Adam Clark and the Tolly Group. Or Raystream: https://news.ycombinator.com/item?id=3211630
> the compressed file size was about 4 KB.
> On the same Windows laptop, Saldana invoked another utility that uncompressed the 4 KB file. Uncompressing took about 10 seconds. The uncompressed file was identical to the original file.
Occam's razor: the "compressed" version was a link (shortcut, symlink, hyperlink, whatever) to the original.
> Despite maintaining a healthy skepticism, I have two reasons for believing the files were identical. First, Saldana played the uncompressed video on the Windows laptop. The uncompressed video looked exactly like the original video. That was nice, but it didn't prove anything. I wouldn't be able to identify an unfaithful video copy if it bit me on the nose.
> The second proof that the original and compressed files were identical was much more compelling. Saldana put the uncompressed file on my own USB drive. Later that day, on my own computer, I ran the uncompressed and compressed files through a bit-by-bit comparison program. Sure enough, the two files contain the exact same sequence of 0s and 1s.
> It should be noted that for security reasons, Saldana didn't copy the 4 KB compressed file to my USB drive, so I don't know what bit sequence was inside the 99.78% compressed file.
All of this is consistent with the file being a link.
A more thorough test would have performed the compression and decompression on two different, non-networked machines, so that the only information available to the decompressor is the 4K file.
After that, we need to start removing layers which can be used to obscure the results, eg. performing a more low-level analysis of the compressed version, using machines which the authors have no access to (eg. to avoid a hacked OS/driver which reports false file sizes)