that processing key has to be contiguous at some point
No, it does not. There is no reason to claim that, unless you assume everybody working on the code is a bad programmer. You could add any number of obfuscations to avoid keeping they key in memory in cleartext, or even in registers in cleartext.
Obfuscations can be worked around, of course, but that requires actually disassembling and reverse-engineering the algorithms, a task which is orders of magnitude harder than the current attacks. That is not to say nobody will do it, but to assume that it is a "trivial task" is extremely naïve.
It is only a "trivial process" if you assume that the software will not be tightened up. That is an extremely naïve assumption. The methods used to find the keys that have been found are trivial, but this just means that the programmers are being very, very sloppy. It is highly unlikely that the AACS LA is going to let that go on.
Step 1 is a horrible, horrible idea, because a lot of people who have something worthwhile to add will just not bother if they are first forced to jump through hoops in order to add it.
Welcome to Slashdot, where your actually informative post is buried far down on the page of comments where nobody will ever see it, because everybody's busy posting nonsense in reply to first few posts.
You seems to be quite confused on the layout of image sensors. Most cameras use one sensor per pixel, and interpolate colour information from neighbouring pixels. The only notable exception being the previously mentioned Foveon sensors, which haven't really caught on.
And I don't know what you're trying to say in the second half of the post - my point was that by saving in raw format, you skip all those steps you're describing, in order to do them later on the computer with more control over the result.
No, you're thinking one single use of the word "raw" out of many is the only official one. It is not. There is no specific definition of the word "raw" that means what you are talking about. It is sometimes used in that sense, and sometimes in other senses.
The raw format of cameras is far more "raw" than PFM or any other RGB format. Camera raw formats save the actual output from the image sensor, before applying the numerous algorithms needed to massage the data into RGB form. The point of camera raw data is not just to avoid compression, it is to do the highly complex processing needed at a later stage, and allow for finer control over the resulting output.
You know, apart from all the ads, and the excessive clicking needed to actually get to the post contents, which it shares with most modern forums, this is actually a very nice and readable layout. At least compared to phpbb and its ilk with a million useless stats printed all over the page, along with avatars and signatures and other usesless visual clutter that ends up leaving room for two or three actual sentences of user content per screen.
2), 3), and 4) are small stuff for the sort of folks that run a multi-billion dollar spam industry on botnets
Of all the idiocy in your post, this is the most hilarious. You're comparing taking over a secure dedicated communications line to hacking Windows? What do you think, this will be written in Visual Basic and run on a Windows 95 machine that the pilots also use to browse porn on?
1. Apple Quick Time; Lead to a major 150 million settlement.
Possible, but the issue was settled and it never became clear if Microsoft knowingly tried to steal code from Apple, or if that was all the work of a third party.
2. The code and idea for the embedding in MSIE. Still in the settlement as I recall.
Eolas? Patent issue, not copyright.
3. How about MSIE itself. they cut a deal to pay the mozaic group spin-off a.01 / each one sold and then embedded it (i.e. the company got SQUAT).
Entirely legitimate deal, no copyright infringement going on. A shitty deal for the other part, but they don't really have anyone but themselves to blame for that if they signed the deal.
4. Or how about the theft of the stacker's work in dos 6.2. IIRC, they had to pay something like 75 million (not chump change back in the 80's).
Patent issue, not copyright.
In sum, one possible case of copyright theft. That's hardly a "large amount" so far.
Well, I was really looking for something more clever than simple interpolation algorithms, which is why I was interested in the anisotropic filtering. But reading up on how it is used for resizing on the site, it doesn't seem all that feasible to implement on the GPU after all.
Me, I'm interested. I've got an OpenGL-based image viewer that I'm working on, and implementing a realtime smart upscaling algorithm on the GPU would be awesome. I'm not sure if it's feasible to use that algorithm for doing resizing in realtime and not just noise reduction, though. Any idea about that, before I start wasting my time on it?
Slashdot Mirror
Xee
that processing key has to be contiguous at some point
No, it does not. There is no reason to claim that, unless you assume everybody working on the code is a bad programmer. You could add any number of obfuscations to avoid keeping they key in memory in cleartext, or even in registers in cleartext.
Obfuscations can be worked around, of course, but that requires actually disassembling and reverse-engineering the algorithms, a task which is orders of magnitude harder than the current attacks. That is not to say nobody will do it, but to assume that it is a "trivial task" is extremely naïve.
It is only a "trivial process" if you assume that the software will not be tightened up. That is an extremely naïve assumption. The methods used to find the keys that have been found are trivial, but this just means that the programmers are being very, very sloppy. It is highly unlikely that the AACS LA is going to let that go on.
They are right. It hasn't been hacked, and it is still exactly as strong as it was thought to be when it was launched.
Keys have been leaked, but this was expected to happen and the system is designed to be able to deal with this.
Perhaps you missed the part where they paid NASA to do this. That's not how a scam usually works.
Step 1 is a horrible, horrible idea, because a lot of people who have something worthwhile to add will just not bother if they are first forced to jump through hoops in order to add it.
He is not "trying" to. He is simply accepting that this already happened decades ago, and that there is no reason to do anything about this.
Welcome to Slashdot, where your actually informative post is buried far down on the page of comments where nobody will ever see it, because everybody's busy posting nonsense in reply to first few posts.
Oh man, you just cured cancer! I guess it's a nobel prize for you!
You seems to be quite confused on the layout of image sensors. Most cameras use one sensor per pixel, and interpolate colour information from neighbouring pixels. The only notable exception being the previously mentioned Foveon sensors, which haven't really caught on.
And I don't know what you're trying to say in the second half of the post - my point was that by saving in raw format, you skip all those steps you're describing, in order to do them later on the computer with more control over the result.
No, you're thinking one single use of the word "raw" out of many is the only official one. It is not. There is no specific definition of the word "raw" that means what you are talking about. It is sometimes used in that sense, and sometimes in other senses.
Even though PNG did bi-level transparency just as fine as GIF files - even better, you didn't lose 1 palette entry
Incorrect. You do lose one palette entry.
The raw format of cameras is far more "raw" than PFM or any other RGB format. Camera raw formats save the actual output from the image sensor, before applying the numerous algorithms needed to massage the data into RGB form. The point of camera raw data is not just to avoid compression, it is to do the highly complex processing needed at a later stage, and allow for finer control over the resulting output.
It also means formats with only GPL'd implementations don't take off either. Like DjVu.
Oh man, they must have totally forgotten about that! Good thing you caught it in time, there's still a chance to stop the launch!
I am sorry, but the world inside your head is not actually visible to other people. We are thus unable to "see" it, as per your request.
You know, apart from all the ads, and the excessive clicking needed to actually get to the post contents, which it shares with most modern forums, this is actually a very nice and readable layout. At least compared to phpbb and its ilk with a million useless stats printed all over the page, along with avatars and signatures and other usesless visual clutter that ends up leaving room for two or three actual sentences of user content per screen.
2), 3), and 4) are small stuff for the sort of folks that run a multi-billion dollar spam industry on botnets
Of all the idiocy in your post, this is the most hilarious. You're comparing taking over a secure dedicated communications line to hacking Windows? What do you think, this will be written in Visual Basic and run on a Windows 95 machine that the pilots also use to browse porn on?
What, you're supporting software patents? On Slashdot?
1. Apple Quick Time; Lead to a major 150 million settlement.
.01 / each one sold and then embedded it (i.e. the company got SQUAT).
Possible, but the issue was settled and it never became clear if Microsoft knowingly tried to steal code from Apple, or if that was all the work of a third party.
2. The code and idea for the embedding in MSIE. Still in the settlement as I recall.
Eolas? Patent issue, not copyright.
3. How about MSIE itself. they cut a deal to pay the mozaic group spin-off a
Entirely legitimate deal, no copyright infringement going on. A shitty deal for the other part, but they don't really have anyone but themselves to blame for that if they signed the deal.
4. Or how about the theft of the stacker's work in dos 6.2. IIRC, they had to pay something like 75 million (not chump change back in the 80's).
Patent issue, not copyright.
In sum, one possible case of copyright theft. That's hardly a "large amount" so far.
Man, you MS types
That's hilarious, man. I'm a Mac developer.
What are some examples of these "large amounts of theft"?
Well, I was really looking for something more clever than simple interpolation algorithms, which is why I was interested in the anisotropic filtering. But reading up on how it is used for resizing on the site, it doesn't seem all that feasible to implement on the GPU after all.
Me, I'm interested. I've got an OpenGL-based image viewer that I'm working on, and implementing a realtime smart upscaling algorithm on the GPU would be awesome. I'm not sure if it's feasible to use that algorithm for doing resizing in realtime and not just noise reduction, though. Any idea about that, before I start wasting my time on it?
On Slashdot, actual cleverness is off-topic.
Ever thought that is a bad picture to denoise any way?
Why are you asking him? He didn't put it up on his website to promote his denoising software.