Codec Reader for HTML Video Tags, gets the correct codecs parameter for AV1 and H.264 videos.

codec-reader.py

import json
import mimetypes
import subprocess
import sys

# Codec Reader for HTML Video Tags by Steve-Tech
# Usage: python3 codec-reader.py [-d] file_name
# Requires ffmpeg and ffprobe to be installed.
#
# Supported Codecs:
#   Video:
#       AV1
#       H.264
#   Audio:
#       AAC
#

debug = False


def main():
    global debug
    if len(sys.argv) > 1:
        debug = "-d" in sys.argv
        file_name = sys.argv[-1]
    else:
        file_name = input("File name: ")
    print(get_type(file_name))


def get_codecs(file_name: str) -> tuple[str]:
    """
    Returns a tuple of codecs found in the file.
    Requires ffprobe to be installed.
    """
    output = subprocess.check_output(
        ["ffprobe", "-of", "json", "-show_streams", file_name],
        stderr=subprocess.DEVNULL,
    )
    codecs = tuple(stream["codec_name"] for stream in json.loads(output)["streams"])

    if debug:
        print("Found codecs:", codecs)

    return codecs


def get_ffmpeg_headers(file_name: str) -> bytes:
    """
    Returns the headers from ffmpeg.
    Requires ffmpeg to be installed.
    """
    return subprocess.run(
        [
            "ffmpeg",
            "-i",
            file_name,
            "-c:v",
            "copy",
            "-bsf:v",
            "trace_headers",
            "-f",
            "null",
            "/dev/null",
        ],
        stdin=subprocess.DEVNULL,
        stderr=subprocess.PIPE,
    ).stderr


def get_type(file_name: str) -> str:
    """
    Returns the mime type of the file including codecs.
    """
    mime = mimetypes.guess_type(file_name)[0]
    codecs = get_codecs(file_name)

    headers = get_ffmpeg_headers(file_name)

    type_codecs = []
    for codec in codecs:
        match codec:
            case "av1":
                type_codecs.append(get_type_av1(headers))
            case "h264":
                type_codecs.append(get_type_h264(headers))
            case "aac":
                type_codecs.append(get_type_aac(headers))
            case _:
                print(f"Unknown codec: {codec}")

    return f"{mime}; codecs={','.join(type_codecs)}"


def str_chk(s, r: str | None = None) -> str:
    """
    Returns a string if it is not None, otherwise returns the replacement or raises an error.
    """
    if s is not None:
        return str(s)
    elif r is not None:
        return r
    else:
        raise ValueError("Missing value")


def read_ffmpeg(headers: bytes, item: bytes) -> int | None:
    """
    Returns the value of an item from the headers in ffmpeg.
    None if the item is not found.
    """
    index = headers.find(item)
    if index == -1:
        return None

    stop = headers.index(b"\n", index)
    start = headers.rindex(b" ", index, stop)

    if debug:
        print(headers[index:stop].decode())

    return int(headers[start + 1 : stop])


# -------- Video Codecs --------


def get_type_av1(headers: bytes) -> str:
    # Based on this: https://jakearchibald.com/2022/html-codecs-parameter-for-av1/
    # av01.P.LLT.DD.M.CCC.cp.tc.mc.F

    codec = ["av01"]

    # P
    codec.append(str_chk(read_ffmpeg(headers, b"seq_profile")))

    # LLT
    codec.append(
        str_chk(read_ffmpeg(headers, b"seq_level_idx")).rjust(2, "0")
        + ("H" if read_ffmpeg(headers, b"seq_tier") else "M")
    )

    # DD
    if read_ffmpeg(headers, b"high_bitdepth") == 1:
        codec.append("12" if read_ffmpeg(headers, b"twelve_bit") == 1 else "10")
    else:
        codec.append("08")

    # M
    codec.append("1" if read_ffmpeg(headers, b"mono_chrome") == 1 else "0")

    # CCC
    if codec[4] == "1":  # M = 1
        codec.append("111")
    elif codec[1] == "0":  # P = 0
        codec.append("110")
    elif codec[1] == "1":  # P = 1
        codec.append("000")
    elif codec[3] != "12":  # DD != 12
        codec.append("100")
    else:
        chroma_x = read_ffmpeg(headers, b"subsampling_x")
        chroma_y = read_ffmpeg(headers, b"subsampling_y")
        codec.append(f"{chroma_x if chroma_x else 0}{chroma_y if chroma_y else 0}0")

    # cp, tc, mc
    codec.append(str_chk(read_ffmpeg(headers, b"color_primaries"), "01").rjust(2, "0"))
    codec.append(
        str_chk(read_ffmpeg(headers, b"transfer_characteristics"), "01").rjust(2, "0")
    )
    codec.append(
        str_chk(read_ffmpeg(headers, b"matrix_coefficients"), "01").rjust(2, "0")
    )

    # F
    codec.append(str_chk(read_ffmpeg(headers, b"color_range"), "0"))
    return ".".join(codec)


def get_type_h264(headers: bytes) -> str:
    # Based on this: https://blog.pearce.org.nz/2013/11/what-does-h264avc1-codecs-parameters.html with knowledge from AV1
    # avc1.PPCCLL

    codec = ["avc1"]

    # PP
    codec.append("{:0>2x}".format(read_ffmpeg(headers, b"profile_idc")))

    # CC
    constraints = 0
    for i in range(6):
        constraints |= (
            read_ffmpeg(headers, str.encode(f"constraint_set{i}_flag")) or 0
        ) << (8 - i)
    codec.append("{:0>2x}".format(constraints))

    # LL
    codec.append("{:0>2x}".format(read_ffmpeg(headers, b"level_idc")))

    return ".".join(codec)


# -------- Audio Codecs --------


def get_type_aac(headers: bytes) -> str:
    # Source: https://developer.mozilla.org/en-US/docs/Web/Media/Formats/codecs_parameter
    # mp4a.oo[.A]
    # oo = 40 Audio Object
    # A  = 2 AAC LC (Profiles: Main, Scalable, HQ, LD v2, AAC, HE-AAC, HE-AAC v2)
    return "mp4a.40.2"


# -------- Main --------
if __name__ == "__main__":
    main()

I really appreciate you creating this! I forked a version that has h265 hevc code. Maybe not as well tested as yours, but seems to work. https://gist.github.com/greg-randall/798fde2376f8dd632ffffb8a7aa4bdb8

@greg-randall Oh wow thanks! You can copy my av1 & h264 code too, I pretty much consider most of my gists public domain, and I don't really intend on maintaining most of them.