When DRM-protected video is displayed on an output device, these streams are sometimes saved and re-encoded, resulting in quality loss. However, let's say we can obtain the outputted video in a pixel-perfect way using some kind of HDCP-bypassing recording device, and call it Vout. To my understanding, we can assume Vout is the result of applying some known compression algorithm 𝔸 to the original* video, which we'll call Vorig. This means there exists some output that is a result of applying 𝔸 to some input (Vorig actually, but that is not relevant) that should be deterministically discoverable. Am I correct in stating that that means that instead of re-encoding and thus losing quality, Vout can be represented in a lossless manner (lossless relative to Vout, not necessarily lossless relative to Vorig) thus making the need for cracking Widevine CDMs and the like unnecessary, because instead the already-compressed representation of Vorig , which would be equal to Vout , can be obtained?


We can obtain the outputted video in a pixel-perfect way
I think this is possible because they have to be outputted to a screen and this data transfer can be intercepted and stored with pixel-perfect accuracy.

𝔸 is known
Generally for playback to be possible, the decompression algorithm and thus the compression algorithm needs to be known by the user's device.

*: Of course some actual original -> web compression is already applied, but that is not relevant here.

  • I’m pretty sure that your assumptions are correct but I’m not entirely sure. Commented May 15 at 9:01
  • 1
    @security_paranoid: What is the OP supposed to do with this statement?
    – Ja1024
    Commented May 15 at 19:00

1 Answer 1


The intent of HDCP is to prevent this attack. So, Vout will be an encrypted stream that can only be played on the device it was generated for. DRM tools will refuse to encrypt/send to a bypass device, because the bypass device's device key won't have been signed by the HDCP master key (assuming falsely that the HDCP master key has not been leaked). If one does manage to create a bypass device, the key(s) assigned to that device/manufacturer will be revoked. To be clear, the above ignores the fact that the HDCP master key has since been leaked.

  • You're falsely assuming, as you said, that the HDCP master key hasn't been leaked, while in fact it is and thus it will allow capture of the video output. I'm not talking about this. I'm taking about there existing a lossless way of storing this output, essentially making its quality equivalent to web-dls, instead of transcoding it lossily which is what is normally done with the recordings now.
    – stenlan
    Commented May 16 at 6:55
  • @stenlan: If you losslessly store the encrypted output, you still won't be able to do anything with it since that output is encrypted. The unencrypted output only exists inside the box that is receiving the signal, but that box is required to be tamper-resistant. All of this becomes moot if the device's private key is leaked (either via physical attack on the hardware or intentionally by the device manufacturer). However, as-designed the spec for HDCP does block such attacks, though it does so by assuming the hardware is a blackbox and that the master key won't be leaked.
    – Brian
    Commented May 16 at 15:41
  • @stenlan: If your HDCP device supports spitting out data to other devices, it will refuse to send this data (at full resolution) to other devices unless those devices also support HDCP, in which case the output will be encrypted to block eavesdropping. Here, I'm assuming the device is spec-compliant (because device keys are in theory only granted to spec-compliant devices and then revoked if the device turns out to be non-compliant).
    – Brian
    Commented May 16 at 15:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .