Quote from: what-is-in-it on August 23, 2017, 08:45:48 AM

Quote from: veredictum-cto on August 23, 2017, 12:31:33 AM

Response to a couple of questions on another thread:

Quote

what-is-in-it
Newbie
*

Activity: 14

View Profile Personal Message (Offline)
Trust: 0: -0 / +0
Ignore

Re: [ANN]Veredictum.io DECENTRALISED ANTI-PIRACY & CONTENT DISTRIBUTION PLATFORM
August 22, 2017, 03:04:49 PM
#4
This could be interesting.. I have a few questions though.

According to your white paper, 70% of the subscription fee will be converted to Vertana for the use of their computational power on a monthly basis.
(... Members collectively use their unused computing power and bandwidth to search for digitally fingerprinted video based content - in the similar way SETI1 ...)
Is this client specific ? Do you limit the number of "miners" who can support a client? OR is this based on the % computational power used irrespective of the number of users?

The second aspect is related to searching & identifying content:
Page 11 of the introductory white paper talks about how the client compensation of "miner/miners". Is this a non-subscription based approach?
Reason being, the fee paid by the subscriber should ideally cover the cost of identifying the pirated copies (70%)

Technically, digital watermarks can be erased - also this technology is constantly evolving . I understand that you are using an in-house solution - is this a patented steganographic solution? The fact that this is embedded this in the blockchain only helps the miners to identify it, but doesnt talk about the steganographic technology involved..which is critical.

As far as i know, you also need device manufacturers to agree to use this.
Just curious how that would play out.

On another note, apart from the subscriber solution, is this different from Custotech? see here: https://custostech.com/

Good luck!!

1. Is this client specific?
- Not sure what you mean by this, but Veredictum is building the client. It is a web-based 'miner' that you can turn on by logging on to the Veredictum platform. It uses a C-based video transcoder that is compiled to a WASM target.

2. Do you limit the number of miners?
- Yes we will. This is because there is only a limited number of URLs that we will want to inspect to detect watermarked material. More miners past a certain point puts us in diminishing returns territory, and reduces the rewards for all miners as they're split up equally. The 'miner' will scale up based on how much computing power you want to throw at it. Initially we may only support single machine mining (i.e. log in, set some parameters, press start), but we will also look at a dedicated miner. Note that the miner is does not work on a GPU, CPU cores only.

3. Miner compensation and subscription
- Not sure what the question is here. Basically, the more fees we get from subscribers (people wanting to pay to search for their content), the more we can funnel through to the mining pool (to search more areas). In the first while however, we will need to be searching a significant enough section of the web in order to make the product useful, even though we might not have many customers (this is where the 'incentivisation pool' comes in to boost the funds sent through to the miners).

4. "Is this a patented steganographic solution?
- Yes. The decoder will be embedded in the miner so that when files are pulled down, the miner can detect the watermark. It is possible for the watermark to be erased, however this is very very difficult. If someone can erase the watermark and restore the original video, tell them to email me for a job.

5. As far as i know, you also need device manufacturers to agree to use this.
- Not so. Out of curiosity, where did you pick this up from? We have *some* hardware-enabled tech in the pipeline (relating to the distribution side of things) however the specific hardware we need is already in a great majority of devices globally, and is increasing over time.

6. is this different from Custotech?
- Very different. Custostech have a manual process for claiming rewards. Our system is automatic and is based on even a small unit of effort. With Custos, you need to actually find a file with the information buried within. With us, you get rewarded regardless.

Trust this helps. Let me know if any more questions.

Cheers
Sam

Thanks for picking this up from the other thread.
1. Sorry, that didnt come out right - when I meant client - I was referring to a paying customer.

2.That clarifies.

3. What I meant to ask is whether a paying customer (subscriber) has the search & detect part included in the subscription (because 70% of subscription fees is pay out to miners).

4. I raised the erasing of watermarks based on :

https://arstechnica.com/tech-policy/2015/01/pirates-defeating-watermarks-releasing-torrents-of-oscar-movie-screeners/
http://www.mecs-press.org/ijmecs/ijmecs-v7-n12/IJMECS-V7-N12-5.pdf (page 38- 39 of the research paper highlights the limitations of different watermarking technologies)
Audio watermarks in Videos/Cinavia: https://ps3yellowlightofdeaths.com/bypass-cinavia-protection-permanently-remove-blue-ray-protection/
There is also an IEEE paper that lists some of the gaps. sending it when I find it

The techniques involved, based on my limited understanding, will require content creators to adopt your patented solution..

5. http://digitalwatermarkingalliance.org/journal-implementation-challenges-opportunities-forensic-watermarking/

Please dont get me wrong, I think this is a great example of providing a solution with value based on the blockchain. I am just trying to find out if the USP is the subscription model (in which case, it can be copy-pasta) or if there is a robust stenographic technology backend that provides a differentiation.

Thanks again for your answers!

1. "Client"
- Ah, I see. No, miner's are not customer-specific. Miners are assigned at random to jobs in the search space.

2.That clarifies.
- Great

3. Subscription fees
Customers pay for the two different services differently:
- watermarking and registration (one-off fee per video size/length)
- search and detect (separate subscription)

4. Watermark erasure
- Right. You are correct that it's technically possible to remove or destroy watermarks within a forensically marked piece of content. I should say that it's easier to destroy the watermark, but this also typically means reducing the quality of the content to a very low (unwatchable?) level. This line of reasoning falls into the margins of what we're trying to achieve, which is a general reduction in piracy, not total, complete elimination. Our goal is 80% reduction in 10 years. A sufficiently motivated and skilled individual may be able to remove our watermarks. If this starts to occur, we can improve our tooling, but ultimately that's missing the point. This is about providing this technology to everyone, not just big film studios.

5. ?
- I'm a little lost on the last question, but yes, we have a steganographic back end which we provide to encode video with a watermark. This watermarked file is given back to the customer and the watermark is registered to a blockchain. Okay, fair enough, but how do you find where your content has gone? Phase 2 is the distributed search system that is described in the whitepaper. Okay, so I can find my content and prove that its mine so I can sue/issue-takedown/celebrate-because-it's-actually-supposed-to-be-a-viral-marketing-video, but I just don't want to get pirated in the first place! Phase 3 (will be done in parallel with full funding) tackles this by using an on-the-fly decryption mechanism. Again, there is a way to defeat this too, but it would be a video camera recording, or it would take a very, very motivated individual to rip the video and audio streams and post the (watermarked) content on the open web. But, if this happens, we know exactly where it has come from.

6. Thanks for your answers!
- No worries. Thanks for your questions

Cheers

Sam