How to get the widevine pssh from init mp4 JUST using python script?

[ancientbanana]

Hi! As the title suggests, I was wondering if there was a way to get the PSSH from init mp4 file without using mp4dump.exe/axiomtools website just using pure python and bento4 sdk somehow. I tried reading their code but I could not understand an easy way to use mp4_dump.py or any other python scripts directly to get init.mp4's KID for the video, which in turn can be used with get_pssh.py file to convert KID into pssh.

Edit: Solution: https://forum.videohelp.com/threads/405001-How-to-get-the-widevine-pssh-from-init-mp4-...pt#post2650501

[ancientbanana]

In case someone here is still wondering, for windows this is how I figured to do it:

Download the zip file, replace the mp4dump.exe if needed, make changes on do.py file as necessary (if at all needed), place the init mp4 files into init folder, and then run the do.py file. As far as I remember, you won't need to install any dependencies for running the python script. Please let me know if you have any questions.

[codehound]

Please let me know if you have any questions.

I have one: why are you trying to reinvent the wheel ?

This is more steps than just running mp4dump - which is one of your dependencies, lol.

[ancientbanana]

Please let me know if you have any questions.

I have one: why are you trying to reinvent the wheel ?

This is more steps than just running mp4dump - which is one of your dependencies, lol.

Currently, while using mp4dump, for certain init mp4s, we just get default KID out of it. And once, we do get the KID, we need to remove the spaces and then use the axiomtools website so as to generate the PSSH corresponding to it. And so, that may be a hassle for some of us who are trying to automate most of the things, and therefore I wanted to have an easier solution for guys like me who like "one touch" solutions to their problems
However, if you do have an idea that's simpler than this, please do let me know, I would definitely love to learn something new!
Sample init which does not have pssh data already:

[zackmark29]

Please let me know if you have any questions.

I have one: why are you trying to reinvent the wheel ?

This is more steps than just running mp4dump - which is one of your dependencies, lol.

Currently, while using mp4dump, for certain init mp4s, we just get default KID out of it. And once, we do get the KID, we need to remove the spaces and then use the axiomtools website so as to generate the PSSH corresponding to it. And so, that may be a hassle for some of us who are trying to automate most of the things, and therefore I wanted to have an easier solution for guys like me who like "one touch" solutions to their problems
However, if you do have an idea that's simpler than this, please do let me know, I would definitely love to learn something new!
Sample init which does not have pssh data already:

Since you tried it first, here's the simplest way without using mp4dump

def find_wv_pssh_offset(raw: bytes) -> str:
    print('Searching pssh offset')
    offset = raw.rfind(b'pssh')
    return raw[offset - 4:offset - 4 + raw[offset - 1]]


def to_pssh(content: bytes) -> str:
    wv_offset = find_wv_pssh_offset(content)
    return base64.b64encode(wv_offset).decode()


def from_file(file_path: str) -> str:
    print('Extracting PSSH from init file:', file_path)
    return to_pssh(Path(file_path).read_bytes())

You might wondering why it has separated functions?
Well you can use that to_pssh for like downloading the init url and reading the response content.

So same as "from_file" function

def from_init_url(init_url: str) -> str:
   # your code here
   # and then when returning, use to_pssh(response.content)

Good luck

[ancientbanana]

Please let me know if you have any questions.

I have one: why are you trying to reinvent the wheel ?

This is more steps than just running mp4dump - which is one of your dependencies, lol.

Currently, while using mp4dump, for certain init mp4s, we just get default KID out of it. And once, we do get the KID, we need to remove the spaces and then use the axiomtools website so as to generate the PSSH corresponding to it. And so, that may be a hassle for some of us who are trying to automate most of the things, and therefore I wanted to have an easier solution for guys like me who like "one touch" solutions to their problems
However, if you do have an idea that's simpler than this, please do let me know, I would definitely love to learn something new!
Sample init which does not have pssh data already:

Since you tried it first, here's the simples way without using mp4dump

def find_wv_pssh_offset(raw: bytes) -> str:
    print('Searching pssh offset')
    offset = raw.rfind(b'pssh')
    return raw[offset - 4:offset - 4 + raw[offset - 1]]


def to_pssh(content: bytes) -> str:
    wv_offset = find_wv_pssh_offset(content)
    return base64.b64encode(wv_offset).decode()


def from_file(file_path: str) -> str:
    print('Extracting PSSH from init file:', file_path)
    return to_pssh(Path(file_path).read_bytes())

You might wondering why it has separated functions?
Well you can use that to_pssh for like downloading the init url and reading the response content.

So same as "from_file" function

def from_init_url(init_url: str) -> str:
   # your code here
   # and then when returning, use to_pssh(response.content)

Good luck

Thank you so much! It works perfectly awesome, I really appreciate your help with this! Have a good day!

[Sadomasochist]

Init to pssh is smaller than eme logger pssh. Only eme logger pssh gives the key. So why init to pssh gives smaller pssh.

https://cdn.class101.net/videos/447d9e6d-3dae-4d3b-b7e1-c1a3f541168e/cmaf/video/avc1/1/init.mp4

Pssh from init to pssh :- AAAAIHBzc2gAAAAAKXAf5DzHSjSMW66Qx0OaRwAAAAA=

Pssh from eme logger:- AAAAbHBzc2gAAAAA7e+LqXnWSs6jyCfc1R0h7QAAAEwSEDQGzo/zboUj5s5B5NpAMb8aDGlua2FlbnR3b3JrcyIkNDQ3ZDllNmQtM 2RhZS00ZDNiLWI3ZTEtYzFhM2Y1NDExNjhlSPPGiZsG

[ElCap]

Init to pssh is smaller than eme logger pssh. Only eme logger pssh gives the key. So why init to pssh gives smaller pssh.

https://cdn.class101.net/videos/447d9e6d-3dae-4d3b-b7e1-c1a3f541168e/cmaf/video/avc1/1/init.mp4

Pssh from init to pssh :- AAAAIHBzc2gAAAAAKXAf5DzHSjSMW66Qx0OaRwAAAAA=

Pssh from eme logger:- AAAAbHBzc2gAAAAA7e+LqXnWSs6jyCfc1R0h7QAAAEwSEDQGzo/zboUj5s5B5NpAMb8aDGlua2FlbnR3b3JrcyIkNDQ3ZDllNmQtM 2RhZS00ZDNiLWI3ZTEtYzFhM2Y1NDExNjhlSPPGiZsG

put them both into https://tools.axinom.com/decoders/PsshBox and see for yourself

the longer one is a widevine pssh and it contains both kid and content id in it - correct.

the short one is not a widevine pssh and doesnt look complete either.
after looking at the init.mp4 you pasted, it contains 3 different psshs - widevine, playready and fairplay and the script seems to be incorrectly grabbing the fairplay pssh data.

the correct widevine pssh can be grabbed from that init.mp4 file

[Sadomasochist]

I need a script which shows all 3 scripts, do you have one? I am writing a script which will decrypt and download, if I can get the proper pssh. It would be helpful!

[sk8ordi3]

thanks zackmark29,
changeing a little to getting the all pssh's from init and filter with characters limit 70 to 190
and write just filtered in pssh.txt file

import base64
from pathlib import Path

def find_wv_pssh_offsets(raw: bytes) -> list:
    print('Searching pssh offsets')
    offsets = []
    offset = 0
    while True:
        offset = raw.find(b'pssh', offset)
        if offset == -1:
            break
        size = int.from_bytes(raw[offset-4:offset], byteorder='big')
        pssh_offset = offset - 4
        offsets.append(raw[pssh_offset:pssh_offset+size])
        offset += size
    return offsets

def to_pssh(content: bytes) -> list:
    wv_offsets = find_wv_pssh_offsets(content)
    return [base64.b64encode(wv_offset).decode() for wv_offset in wv_offsets]

def from_file(file_path: str) -> list:
    print('Extracting PSSHs from init file:', file_path)
    return to_pssh(Path(file_path).read_bytes())

pssh_list = from_file('init.mp4')
target_pssh = None
for pssh in pssh_list:
    if 70 < len(pssh) < 190:
        target_pssh = pssh
print(f'\n{target_pssh}\n')
print(target_pssh, file=open("pssh.txt", "w"))

[Attachment 70739 - Click to enlarge]

[Sadomasochist]

Thanks for this! but init.mp4 should be in the folder for it to work! It is good enough! I will find a way for it to extract init.mp4 from the site and find the pssh.

[ampersand]

I usually download a veeeery small test file using yt-dlps --test option and then use this on it:

def construct_pssh():
    mymp4dump = subprocess.Popen(
        [const.mp4dumpexe, '--verbosity', '3', '--format', 'json', 'temp/test.mp4'],
        stdout=subprocess.PIPE, stderr=subprocess.DEVNULL)
    dump = (mymp4dump.stdout.read())
    json_info = json.loads(dump.decode("utf-8"))

    for temp in json_info:
        if (temp['name'] == 'moov'):
            for child in temp['children']:
                if ('system_id' in child and child['system_id'] == '[ed ef 8b a9 79 d6 4a ce a3 c8 27 dc d5 1d 21 ed]'):
                    pssh_data = child['data']
                    pssh_data_size_dec = child['data_size']
                    break
            break

    pssh_data = pssh_data.replace('[', '').replace(']', '').replace(' ', '')
    pssh_size_dec = pssh_data_size_dec + 32
    pssh_data_size_hex = (str(hex(pssh_data_size_dec))).replace('0x', '')
    pssh_size_hex = (str(hex(pssh_size_dec))).replace('0x', '')

    pssh_hex = f'{pssh_size_hex:0>8}7073736800000000edef8ba979d64acea3c827dcd51d21ed{pssh_data_size_hex:0>8}{pssh_data}'

    return (base64.b64encode(bytes.fromhex(pssh_hex))).decode('utf-8')

[marioeivissa]

In case someone here is still wondering, for windows this is how I figured to do it:

Download the zip file, replace the mp4dump.exe if needed, make changes on do.py file as necessary (if at all needed), place the init mp4 files into init folder, and then run the do.py file. As far as I remember, you won't need to install any dependencies for running the python script. Please let me know if you have any questions.

its not working for me, ill do with the mp4 you put here, but is not working, any idea? shoul i change the go.py?how?

[marioeivissa]

thanks zackmark29,
changeing a little to getting the all pssh's from init and filter with characters limit 70 to 190
and write just filtered in pssh.txt file

import base64
from pathlib import Path

def find_wv_pssh_offsets(raw: bytes) -> list:
    print('Searching pssh offsets')
    offsets = []
    offset = 0
    while True:
        offset = raw.find(b'pssh', offset)
        if offset == -1:
            break
        size = int.from_bytes(raw[offset-4:offset], byteorder='big')
        pssh_offset = offset - 4
        offsets.append(raw[pssh_offset:pssh_offset+size])
        offset += size
    return offsets

def to_pssh(content: bytes) -> list:
    wv_offsets = find_wv_pssh_offsets(content)
    return [base64.b64encode(wv_offset).decode() for wv_offset in wv_offsets]

def from_file(file_path: str) -> list:
    print('Extracting PSSHs from init file:', file_path)
    return to_pssh(Path(file_path).read_bytes())

pssh_list = from_file('init.mp4')
target_pssh = None
for pssh in pssh_list:
    if 70 < len(pssh) < 190:
        target_pssh = pssh
print(f'\n{target_pssh}\n')
print(target_pssh, file=open("pssh.txt", "w"))

[Attachment 70739 - Click to enlarge]

should i have mp4dump.exe inside the folder?

[2nHxWW6GkN1l916N3ayz8HQoi]

thanks zackmark29,
changeing a little to getting the all pssh's from init and filter with characters limit 70 to 190
and write just filtered in pssh.txt file

...snipped...

Nice script. Can it be adapted for playready pssh?