Commit 5f67ceef authored by multrus's avatar multrus
Browse files

first implementation

parent d936bbd3

examples/TEMPLATE.yml

+2 −2
Original line number Diff line number Diff line
@@ -64,8 +64,8 @@ input: 
# preprocessing:

    ### Target format used in rendering from input format; default = null (no rendering)

    # fmt: "7_1_4"

    ### Flag for application of 50Hz high-pass filter; default = false

    # hp50: true

    ### Define mask (HP50 or 20KBP) for input signal filtering; default = None

    # mask: None

    ### Target sampling rate in Hz for resampling; default = null (no resampling)

    # fs: 16000

    ### Target loudness in LKFS; default = null (no loudness change applied)

ivas_processing_scripts/audiotools/__init__.py

+4 −4
Original line number Diff line number Diff line
@@ -79,10 +79,10 @@ def add_processing_args(group, input=True): 
        default=None,

    )

    group.add_argument(

        f"-{ps}hp",

        f"--{p}_hp50",

        help="Apply 50 Hz high-pass filtering (default = %(default)s)",

        action="store_true",

        f"-{ps}mk",

        f"--{p}_mask",

        help="Apply filtering with mask (default = %(default)s)",

        default=None,

    )

    group.add_argument(

        f"-{ps}w",

ivas_processing_scripts/audiotools/constants.py

+1 −0
Original line number Diff line number Diff line
@@ -692,6 +692,7 @@ DELAY_COMPENSATION_FOR_FILTERING = { 
        "down": 145,

    },

    "MSIN": 92,

    "20KBP": 200,

    "LP1p5": 322,

    "LP35": 232,

    "LP7": 117,

ivas_processing_scripts/audiotools/convert/__init__.py

+17 −9
Original line number Diff line number Diff line
@@ -44,6 +44,7 @@ from ivas_processing_scripts.audiotools.wrappers.bs1770 import loudness_norm 
from ivas_processing_scripts.audiotools.wrappers.esdru import esdru

from ivas_processing_scripts.audiotools.wrappers.filter import (

    hp50filter_itu,

    kbp20filter_itu,

    lpfilter_itu,

    resample_itu,

)
@@ -133,7 +134,7 @@ def convert( 
    in_delay: Optional[float] = None,

    in_fs: Optional[int] = None,

    in_cutoff: Optional[int] = None,

    in_hp50: Optional[bool] = None,

    in_mask: Optional[str] = None,

    in_window: Optional[list] = None,

    in_loudness: Optional[float] = None,

    in_loudness_fmt: Optional[str] = None,
@@ -142,7 +143,7 @@ def convert( 
    out_delay: Optional[float] = None,

    out_fs: Optional[int] = None,

    out_cutoff: Optional[int] = None,

    out_hp50: Optional[bool] = None,

    out_mask: Optional[str] = None,

    out_window: Optional[list] = None,

    out_loudness: Optional[float] = None,

    out_loudness_fmt: Optional[str] = None,
@@ -162,7 +163,7 @@ def convert( 
        delay=in_delay,

        fs=in_fs,

        fc=in_cutoff,

        hp50=in_hp50,

        mask=in_mask,

        window=in_window,

        loudness=in_loudness,

        loudness_fmt=in_loudness_fmt,
@@ -180,7 +181,7 @@ def convert( 
        delay=out_delay,

        fs=out_fs,

        fc=out_cutoff,

        hp50=out_hp50,

        mask=out_mask,

        window=out_window,

        loudness=out_loudness,

        loudness_fmt=out_loudness_fmt,
@@ -198,7 +199,7 @@ def process_audio( 
    delay: Optional[float] = None,

    fs: Optional[int] = None,

    fc: Optional[int] = None,

    hp50: Optional[bool] = False,

    mask: Optional[str] = None,

    window: Optional[float] = None,

    loudness: Optional[float] = None,

    loudness_fmt: Optional[str] = None,
@@ -232,11 +233,18 @@ def process_audio( 
            logger.debug(f"Windowing audio with {window} ms Hann window")

        x.audio = audioarray.window(x.audio, x.fs, window)



    """high-pass (50 Hz) filtering"""

    if hp50:

    """mask filtering"""

    if mask is not None:

        if logger:

            logger.debug("Applying 50 Hz high-pass filter using ITU STL filter")

            logger.debug("Applying mask filter using ITU STL filter")

        if mask == "HP50":

            x.audio = hp50filter_itu(x)

#        elif mask == "20KBP":

#            x.audio = kbp20filter_itu(x)

#        else:

#            raise ValueError(

#                "Invalid mask filter defined"

#            )



    """resampling"""

    if x.fs != fs:

ivas_processing_scripts/audiotools/wrappers/filter.py

+52 −0
Original line number Diff line number Diff line
@@ -291,6 +291,58 @@ def hp50filter_itu( 
    return y





def kbp20filter_itu(

    x: Audio,

) -> np.ndarray:

    """

    20Hz to 20kHz bandpass filter for multi-channel audio array



    Parameters

    ----------

    x: Audio

        Input audio



    Returns

    -------

    y: np.ndarray

        Output high-pass filtered array

    """



    # set filter type and check if sampling rate is supported

    old_fs = None

    tmp = copy(x)

    if x.fs == 48000:

        flt_type = "20KBP"

    else:

        # resample if samplingrate is not supported

        warn(

            f"Filter type 20KBP only supported for 48kHz samlingrate, not for {x.fs}Hz -> resampling"

        )

        flt_type = "20KBP"

        old_fs = x.fs

        tmp.audio = resample_itu(tmp, 48000)

        tmp.fs = 48000



    # don't apply high-pass filtering to LFE channel

    if isinstance(x, ChannelBasedAudio):

        skip_channel = x.lfe_index

    else:

        skip_channel = None



    # apply filter

    y = filter_itu(tmp, flt_type=flt_type, skip_channel=skip_channel)



    # delay compensation

    y = delay_compensation(y, flt_type=flt_type, fs=tmp.fs)



    # reverse resampling

    if old_fs:

        tmp.audio = y

        y = resample_itu(tmp, old_fs)



    return y





def resample_itu(

    x: Audio,

    fs_new: int,