"Cinavia"/Verance BluRay audio watermark protection

[diogen]

If he's not a spy, he is going to be busy
http://www.theregister.co.uk/2012/04..._sopa_returns/

...bill currently being considered by the US House of Representatives could make the SOPA legislation look positively idyllic.

Diogen.

[CruNcher]

aha CISPA sounds familiar like the "Vorratsdatenspeicherung/Stock Data Recording Legeslation Act of the European Union" hmm i wonder who was the mother of which

[Atak_Snajpera]

Originally Posted by Ghitulescu

How is the quality of the sound in 32kHz?

i doubt that we will notice audio degradation in movies(explosions,shoots,voices...).

also mising 16khz could have been restored by algo similar to SBR in he-aac.

[rotty]

Wiki list has changed. Whats happened to the full list of Cinavia infected disks.

http://en.wikipedia.org/wiki/Cinavia

P.S. Seems to have been rectified.

[Big Vern]

Originally Posted by MRCOCOset

(I am of Polish translated by Google)
WELCOME. I do not know if this has already been described in this thread. So excuse me if I repeat.
Inspired by that someone wrote above that the code is stored in the upper parts of the soundtrack I decided to cut the upper part of the path of CINAVIA film (for me CARRANGE with dts) from 48kHz to 32kHz (as in Divix once).

First I open MKV with DTS x64 media coder and the formation of 32kHz stereo MP3 file (2 minutes work at 3GHz 4 cores)

PS. You can also create .ac3 5.1/32khz

Second I open MKV with DTS MKVToolnix and add the path mp3 32kHz and delete dts. (MUX without compression)

Third It works! D5500 BD player does not detect CINAVIA

So......has anyone actually tried this yet, or is it just an April fool's joke?

[Atak_Snajpera]

i suspect that it was a joke. cinavia is suppose to survive even recording by cheap mono microphone. most of them can't record anything above 16khz (32khz pcm).

[Ghitulescu]

It might have some hidden truths. Did anyone really tested this claimed invulnerability against cheap mikes? Marketing claims like these have been proven false over time (do I need to remind you those surrounding the CSS?) ...
The mark is somewhere in the upper part of the spectrum, so cutting the trebles makes sense. How much is another story.

[Atak_Snajpera]

If I only had DVD/BD movie with infected soundtrack and proper BD-player with enabled Cinavia I would already be after intensive tests.

[diogen]

Originally Posted by Atak_Snajpera

cinavia is suppose to survive even recording by cheap mono microphone.

I think that's true.
It was suppose to survive home camcording with Cinavia-infected audio in the background.

Going back to the original SDMI paper by Felten & Co. rebuts the inaudible frequency argument.
On page five it shows the difference between the original and Cinavia infected soundtrack in the frequency domain.
And the figure caption reads

A short-term complex echo. Above, the frequency response between the watermarked and original music,
taken over 1/50 second, showing a sinusoidal ripple between 8 and 16 KHz....

I.e. the echo ripple (Cinavia watermark) goes below 16kHz and well into speech frequencies...

Diogen.

[Ghitulescu]

The ideas of SDMI and cinavia are striking similar, yet the implementation details might be different. AFAIK SDMI failed some 10 years ago so its "technology" is frozen at ~2004.
All the details about cinavia are in the patents I put several hundreds of posts earlier
For technical reasons it must have only one max. 2 narrow spectra where the mark is to be found, otherwise the decoding part before finding the mark will incur too much technological costs (analogue or digital filters).

[MRCOCOset]

Originally Posted by Big Vern

So......has anyone actually tried this yet, or is it just an April fool's joke?

no joke ... Everything works. Just try and convince people.

[Atak_Snajpera]

can you try with another cinavia infected movie. this might be just a lucky shot.

[MRCOCOset]

Originally Posted by Atak_Snajpera

can you try with another cinavia infected movie. this might be just a lucky shot.

Jestem Polakiem więc piszę po polsku do Ciebie Polaku.
To nie był przypadek. Kolega kupił odtwarzacz BluRay Samsung D5500 BD. Zaktualizował w nim firmware. Pościągał kupę filmów MKV 1080p. Oglądał. W czterech filmach. "RZEŹ" "The GUARD" (STRAŻNIK) , "HUGO i jego wynalazek" oraz "Jack & Jill" zaczeły mu wyskakiwać wiadome komunikaty CINAVIA. Poprosił mnie o to aby coś z tym zrobić.
Więc. To już znasz:
"Zainspirowany tym że ktoś wyżej napisał że kod jest zapisany w górnych partiach ścieżki dźwiękowej postanowiłem obciąć górne partie ścieżki film z CINAVIA ( u mnie CARRANGE z dts) z 48khz do 32khz ( jak w divix kiedyś ).

1. otwieram MKV z dts w media coder x64 i nim tworze plik mp3 32khz stereo ( 2 minuty roboty przy 4 rdzeniach 3ghz)

2. otwieram MKV z dts w mkvtoolnix i dodaję ścieżkę mp3 32khz a usuwam dts. (MUX bez kompresji)

3. Działa !!! odtwarzacz BD D5500 nie wykrywa CINAVIA

Tak postąpiłem we wszystkich czterech filmach. I działa. Filmy przeleciały po 4 razy każdy i zero komunikatu.
Jakość wiadomo delikatnie obcięte góry ale działa.
Wystarczy spróbować żadna filozofia.
I pamiętaj , że zastosowałem format ścieżki .mp3 (mpeg, mpg) a nie jak piszecie coś wyżej o formacie PCM (np. wav) bo to różnica.

GOOGLE

I am a Pole, so I am writing to you in Polish Pole.
This was no accident. A friend bought a BluRay player Samsung BD D5500. It updated the firmware. Pościągał 1080p MKV movies pile. Watched. In the four films. "Carnage," "The Guard" (GUARD), "Hugo and his invention" and "Jack & Jill" began to pop him known CINAVIA messages. He asked me for it to do something about it.
So. It already know:
"Inspired by that someone wrote above that the code is stored in the upper parts of the soundtrack I decided to cut the upper part of the path of CINAVIA film (for me CARRANGE with dts) from 48kHz to 32kHz (as in Divix once).

First I open MKV with DTS x64 media coder and the formation of 32kHz stereo MP3 file (2 minutes work at 3GHz 4 cores)

Second I open MKV with DTS MKVToolnix and add the path and delete mp3 32kHz dts. (MUX without compression)

Third It works! D5500 BD player does not detect CINAVIA

So I acted in all four films. And it works. Videos flew after 4 times each and no message.
Quality knows gently cut the top but it works.
Just try no philosophy.
I remember that I used the format of the path. Mp3 (mpeg, mpg) and not write something like the above PCM format (eg wav) for the difference.

[Atak_Snajpera]

resampling 48khz to 32khz is not the same as cutting off only upper frequencies.
you should decompress dts to wav. Then open wav in audacity. remove upper 16 khz and save to wav again. compress that file to ac3 5.1 640kbps (48Khz obviously). this method will give better quality than yours.

original


now we apply low pass filter


and here is result

[hello_hello]

Could it be the resampling that stops Cinavia from working, rather than removing high frequencies as such? Has anyone tried this?

I was going to test it. I even upgraded the Samsung Bluray player (BD-C6500) to the latest firmware (dated January), but after the update it played Carnage with DTS audio from start to finish, so I assume the firmware doesn't include Cinavia. Australia seems to be getting left out of the Cinavia craze. Still no Cinavia infected firmware from Sony either.

[Atak_Snajpera]

If we only had software player which detects cinavia... I'm very interested how many "tortures" can cinavia survive

[MRCOCOset]

Originally Posted by Atak_Snajpera

resampling 48khz to 32khz is not the same as cutting off only upper frequencies.
you should decompress dts to wav. Then open wav in audacity. remove upper 16 khz and save to wav again. compress that file to ac3 5.1 640kbps (48Khz obviously). this method will give better quality than yours.

original


now we apply low pass filter


and here is result

After passing the path through the filter is slightly better than with resampling:

Orginal:



Resampled:



Low pass:



But is this method removes CINAVIA?
And after that, more work to get the output file without CINAVIA. The user has the choice of faster using "media coder" at the expense of slightly lower performance "mountains". Or more time-consuming but qualitatively better way to "LOW PASS"
Assuming that your thesis does. Do not check because I have at hand a colleague player.

[Atak_Snajpera]

it would be nice if you could verify whether my method realy destroys cinavia or not?

[MRCOCOset]

Originally Posted by Atak_Snajpera

it would be nice if you could verify whether my method realy destroys cinavia or not?

Prepare a video of the path (low pass). And I will give to check the time.

[Atak_Snajpera]

tomorrow i will prepare few samples using different methods. for example
resampling to 32khz,40khz,44khz and then to 48khz

original to he-aac 5.1 and then to ac3