Comment by gjsman-1000 | Hacker News Reader

Comment by gjsman-1000 | Hacker News Reader

Alleged violations of patents:

- 7,830,967 "Video Camera"

- 8,174,560 "Video Camera"

- 9,245,314 "Video Camera"

- 9,436,976 "Video Camera"

- 9,521,384 "Green Average Subtraction in Image Data"

- 9,716,866 "Green Image Data Processing"

- 10,582,168 "Green Image Data Processing"

Almost all patents use the following generic description or variations thereof:

"Embodiments provide a video camera that can be configured to highly compress video data in a visually lossless manner. The camera can be configured to transform blue, red, and/or green image data in a manner that enhances the compressibility of the data. The camera can be configured to transform at least a portion of the green image data in a manner that enhances the compressibility of the data. The data can then be compressed and stored in this form. This allows a user to reconstruct the red, blue, and/or green image data to obtain the original raw data or a modified version of the original raw data that is visually lossless when demosacied. Additionally, the data can be processed in a manner in which at least some of the green image elements are demosaiced first and then the red, blue, and/or some green elements are reconstructed based on values of the demosaiced green image elements."

IANAL, but this legal description is gobbledygook describing a video camera, that can compress video, in a visually lossless manner, by rearranging RGB data to be more compressible, and then being able to obtain original data from a "demosaiced" version of the RGB data.

RED clearly doesn't like Nikon's use of GREEN.

I think this is straightforward technical language being translated into patent language. But you should always read the claims; the language is usually more precise there.

I believe the technique at issue is a way to compress the data -- using standard codecs like JPEG2000 -- without first de-bayering the image. I imagine it was pretty novel when the patent was filed, which would've been the mid '00s. (The Red One used this technique, and it launched in ~2007.)

The fundamental problem seems to be that in practice, mere novelty is enough to get a patent.

I don't think a company should be granted a 20-year monopoly just because they were the first to write an idea down. I don't see how patents on run-of-the-mill innovations further innovation in the economy as a whole.

Even if there were no such thing as patents, companies would have begun compressing individual channels prior to de-bayering, because it makes sense and is a fairly obvious thing to try. The patent system didn't spur innovation in this case. It just allowed one company to demand a rent from everyone else.

Completely agreed, for the way the patent system actually works in practice.

There is a theory that patent rights can encourage inventors to publish their invention, so that everybody benefits in the long run. This is a nice theory, except it doesn't hold up in practice (at least in the tech industry) because (1) patent language is so obfuscated that it is not a useful source of knowledge; and (2) the threat of triple damages means that no (honest) practitioner in their right mind would ever read a patent other than their own.

It is possible that in other, slower moving industries, people do get some minor value from reading expired patents. Though somehow I doubt it: in our industry, the useful inventions spread through other channels, and I suspect the same is true everywhere.

The application to images may have been novel, but this is just a trivial application of a rule that has been well known since the first uses of compression methods.

Whenever there is some data which passes through several conversion stages, there is a certain point in the chain of transformations where compression is more efficient than in the other points, so it is always necessary to choose carefully where to insert a compression transformation.

For example, if you aggregate some files into an archive file and then you encrypt the archive file, a compression step must be inserted between the concatenation of the input files and the encryption step.

Inserting the compression before concatenation or after encryption will give much worse compression ratios.

The same is true for any chain of transformations. It is always necessary to identify the point where the compression must be inserted, which for image processing in a camera happens to be before debayering.

Absolutely anyone who would receive the task to add compression to an image processing chain of algorithms would start by making tests to determine where to insert the compression.

Discovering the claim of the patent does not require any kind of creativity or any other special skill in the domain. It would have happened automatically to the first one who happened to work at this problem.

From what I understand we should we looking at the patent claims to see what the patent really covers.

Reply on news.ycombinator.com