Skip to content

FFTHideStrinWAV

stegobox.codec.FFTHideStrinWAV

Bases: BaseCodec

This module implements simple FFT steganography hide string in audio (wav).

Referenced from shalom06/Audio-Stego.

Source code in stegobox/codec/fft_hide_str_in_wav.py 
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
class FFTHideStrinWAV(BaseCodec):
    """
    This module implements simple FFT steganography hide string in audio (wav).

    Referenced from [shalom06/Audio-Stego](https://github.com/shalom06/Audio-Stego).
    """

    def __init__(self) -> None:
        super().__init__()

    def encode(self, carrier: np.ndarray, payload: str) -> tuple[np.ndarray, int]:
        """
        Encoder requires the carrier audio in format WAV and the payload in format
        string.

        Args:
            carrier (np.ndarray): Carrier audio in format WAV. Read with
                `stegobox.io.audio.wavfile_read()`.
            payload (str): Payload (secret message) to be encoded.

        Returns:
            np.ndarray: the frame information of the audio with the payload embedded.
            int: the rate of the audio with the payload embedded.
        """
        # step 1 divide into chunks
        carrier = carrier.copy()
        payload_length = 8 * len(payload)
        blocklength = int(2 * 2 ** np.ceil(np.log2(2 * payload_length)))
        blocknumber = int(np.ceil(carrier.shape[0] / blocklength))

        # checks shape to change data to 1 axis
        if len(carrier.shape) == 1:
            carrier.resize(blocknumber * blocklength, refcheck=False)
            carrier = carrier[np.newaxis]
        else:
            carrier.resize(
                (blocknumber * blocklength, carrier.shape[1]), refcheck=False
            )
            carrier = carrier.T

        blocks = carrier[0].reshape((blocknumber, blocklength))
        # Calculate DFT using fft
        blocks = np.fft.fft(blocks)

        # calculate magnitudes
        magnitudes = np.abs(blocks)

        # create phase matrix
        phases = np.angle(blocks)

        # get phase differences
        phasediffs = np.diff(phases, axis=0)

        # conert message to encode into binary
        textinbinary = np.array(
            [[int(y) for y in format(ord(x), "08b")] for x in payload]
        )
        textinbinary = textinbinary.reshape(-1)

        # Convert txt to phase differences
        textinpi = textinbinary.copy()
        textinpi[textinpi == 0] = -1
        textinpi = textinpi * -np.pi / 2

        blockmid = blocklength // 2

        # do phase conversion
        phases[0, blockmid - payload_length : blockmid] = textinpi
        phases[0, blockmid + 1 : blockmid + 1 + payload_length] = -textinpi[::-1]

        # re compute  the ophase amtrix
        for i in range(1, len(phases)):
            phases[i] = phases[i - 1] + phasediffs[i - 1]

        # apply i-dft
        blocks = magnitudes * np.exp(1j * phases)
        blocks = np.fft.ifft(blocks).real

        # combining all block of audio again
        carrier[0] = blocks.ravel().astype(np.int16)

        return carrier.T, payload_length

    def decode(self, _):
        raise NotImplementedError("This codec does not support decoding without length")

    def decode_with_length(self, carrier: np.ndarray, payload_length: int) -> str:
        """
        Decode the secret payload from the carrier audio

        Args:
            carrier (np.ndarray): Carrier audio in format WAV. Read with
                `stegobox.io.audio.wavfile_read()`.
            payload_length (int): The length of secret message.

        Returns:
            str: The decoded payload (secret message).
        """

        blocklength = 2 * int(2 ** np.ceil(np.log2(2 * payload_length)))
        blockmid = blocklength // 2

        # get header info
        if len(carrier.shape) == 1:
            secret = carrier[:blocklength]
        else:
            secret = carrier[:blocklength, 0]

        # get the phase and convert it to binary
        secretphases = np.angle(np.fft.fft(secret))[
            blockmid - payload_length : blockmid
        ]
        secretinbinary = (secretphases < 0).astype(np.int8)

        #  convert into characters
        secretincode = secretinbinary.reshape((-1, 8)).dot(
            1 << np.arange(8 - 1, -1, -1)
        )

        # combine characters to original text
        return "".join(np.char.mod("%c", secretincode))

encode(carrier, payload)

Encoder requires the carrier audio in format WAV and the payload in format string.

Parameters:

Name Type Description Default
carrier ndarray

Carrier audio in format WAV. Read with stegobox.io.audio.wavfile_read().

 required
payload str

Payload (secret message) to be encoded.

 required

Returns:

Name Type Description
ndarray

np.ndarray: the frame information of the audio with the payload embedded.
int int

the rate of the audio with the payload embedded.
Source code in stegobox/codec/fft_hide_str_in_wav.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
def encode(self, carrier: np.ndarray, payload: str) -> tuple[np.ndarray, int]:
    """
    Encoder requires the carrier audio in format WAV and the payload in format
    string.

    Args:
        carrier (np.ndarray): Carrier audio in format WAV. Read with
            `stegobox.io.audio.wavfile_read()`.
        payload (str): Payload (secret message) to be encoded.

    Returns:
        np.ndarray: the frame information of the audio with the payload embedded.
        int: the rate of the audio with the payload embedded.
    """
    # step 1 divide into chunks
    carrier = carrier.copy()
    payload_length = 8 * len(payload)
    blocklength = int(2 * 2 ** np.ceil(np.log2(2 * payload_length)))
    blocknumber = int(np.ceil(carrier.shape[0] / blocklength))

    # checks shape to change data to 1 axis
    if len(carrier.shape) == 1:
        carrier.resize(blocknumber * blocklength, refcheck=False)
        carrier = carrier[np.newaxis]
    else:
        carrier.resize(
            (blocknumber * blocklength, carrier.shape[1]), refcheck=False
        )
        carrier = carrier.T

    blocks = carrier[0].reshape((blocknumber, blocklength))
    # Calculate DFT using fft
    blocks = np.fft.fft(blocks)

    # calculate magnitudes
    magnitudes = np.abs(blocks)

    # create phase matrix
    phases = np.angle(blocks)

    # get phase differences
    phasediffs = np.diff(phases, axis=0)

    # conert message to encode into binary
    textinbinary = np.array(
        [[int(y) for y in format(ord(x), "08b")] for x in payload]
    )
    textinbinary = textinbinary.reshape(-1)

    # Convert txt to phase differences
    textinpi = textinbinary.copy()
    textinpi[textinpi == 0] = -1
    textinpi = textinpi * -np.pi / 2

    blockmid = blocklength // 2

    # do phase conversion
    phases[0, blockmid - payload_length : blockmid] = textinpi
    phases[0, blockmid + 1 : blockmid + 1 + payload_length] = -textinpi[::-1]

    # re compute  the ophase amtrix
    for i in range(1, len(phases)):
        phases[i] = phases[i - 1] + phasediffs[i - 1]

    # apply i-dft
    blocks = magnitudes * np.exp(1j * phases)
    blocks = np.fft.ifft(blocks).real

    # combining all block of audio again
    carrier[0] = blocks.ravel().astype(np.int16)

    return carrier.T, payload_length

decode_with_length(carrier, payload_length)

Decode the secret payload from the carrier audio

Parameters:

Name Type Description Default
carrier ndarray

Carrier audio in format WAV. Read with stegobox.io.audio.wavfile_read().

 required
payload_length int

The length of secret message.

 required

Returns:

Name Type Description
str str

The decoded payload (secret message).
Source code in stegobox/codec/fft_hide_str_in_wav.py 
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
def decode_with_length(self, carrier: np.ndarray, payload_length: int) -> str:
    """
    Decode the secret payload from the carrier audio

    Args:
        carrier (np.ndarray): Carrier audio in format WAV. Read with
            `stegobox.io.audio.wavfile_read()`.
        payload_length (int): The length of secret message.

    Returns:
        str: The decoded payload (secret message).
    """

    blocklength = 2 * int(2 ** np.ceil(np.log2(2 * payload_length)))
    blockmid = blocklength // 2

    # get header info
    if len(carrier.shape) == 1:
        secret = carrier[:blocklength]
    else:
        secret = carrier[:blocklength, 0]

    # get the phase and convert it to binary
    secretphases = np.angle(np.fft.fft(secret))[
        blockmid - payload_length : blockmid
    ]
    secretinbinary = (secretphases < 0).astype(np.int8)

    #  convert into characters
    secretincode = secretinbinary.reshape((-1, 8)).dot(
        1 << np.arange(8 - 1, -1, -1)
    )

    # combine characters to original text
    return "".join(np.char.mod("%c", secretincode))