Commit ff88325e authored by Anika Treffehn's avatar Anika Treffehn
Browse files

adapted ISM metadata to 8 columns and enabled pass through to encoder

parent 5d9dd9d6

ivas_processing_scripts/audiotools/audio.py

+14 −10
Original line number Diff line number Diff line
@@ -46,6 +46,8 @@ from ivas_processing_scripts.audiotools.constants import ( 
    METADATA_ASSISTED_SPATIAL_AUDIO_FORMATS,

    OBJECT_BASED_AUDIO_FORMATS,

    SCENE_BASED_AUDIO_FORMATS,

    NUMBER_COLUMNS_ISM_METADATA,

    DEFAULT_ISM_METADATA

)



from .EFAP import wrap_angles
@@ -295,6 +297,7 @@ class ObjectBasedAudio(Audio): 
        return obj



    def init_metadata(self):

        # check if number of metadata files matches format

        if self.audio.shape[1] != len(self.metadata_files):

            raise ValueError(

                f"Mismatch between number of channels in file [{self.audio.shape[1]}], and metadata [{len(self.metadata_files)}]"
@@ -305,17 +308,18 @@ class ObjectBasedAudio(Audio): 
            pos = np.genfromtxt(f, delimiter=",")



            # check if metadata has right number of columns

            if pos.shape[1] < 5:

                raise ValueError("Metadata incomplete. Columns are missing.")

            elif pos.shape[1] > 5:

                if pos.shape[1] <= 8:

                    # TODO: FIXME

                    pos = pos[:, :5]

                else:

            num_columns = pos.shape[1]

            if num_columns < 2:

                raise ValueError("Metadata incomplete. Columns are missing. Azimuth and elevation are mandatory.")

            elif num_columns > NUMBER_COLUMNS_ISM_METADATA:

                raise ValueError(

                        "Too many columns in metadata (possibly old version with frame index used)"

                    "Too many columns in metadata"

                )



            # pad metadata to max number of columns

            if num_columns < NUMBER_COLUMNS_ISM_METADATA:

                pos = np.hstack([pos, np.array(pos.shape[0] * [DEFAULT_ISM_METADATA[num_columns:]])])



            # check if metadata is longer than file -> cut off

            num_frames = int(

                np.ceil(self.audio.shape[0] / (self.fs * IVAS_FRAME_LEN_MS / 1000))

ivas_processing_scripts/audiotools/constants.py

+4 −0
Original line number Diff line number Diff line
@@ -703,3 +703,7 @@ DELAY_COMPENSATION_FOR_FILTERING = { 
    "HP50_32KHZ": 559,

    "HP50_48KHZ": 839,

}



DEFAULT_ISM_METADATA = [0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0]

FORMAT_ISM_METADATA_CSV = ["%+07.2f", "%+06.2f", "%05.2f", "%06.2f", "%04.2f", "%+07.2f", "%+06.2f", "%1.0f"]

NUMBER_COLUMNS_ISM_METADATA = len(DEFAULT_ISM_METADATA)

ivas_processing_scripts/audiotools/metadata.py

+4 −8
Original line number Diff line number Diff line
@@ -40,7 +40,7 @@ from ivas_processing_scripts.audiotools import audio 
from ivas_processing_scripts.audiotools.audio import fromtype

from ivas_processing_scripts.audiotools.audioarray import trim

from ivas_processing_scripts.audiotools.audiofile import read

from ivas_processing_scripts.audiotools.constants import IVAS_FRAME_LEN_MS

from ivas_processing_scripts.audiotools.constants import IVAS_FRAME_LEN_MS, FORMAT_ISM_METADATA_CSV, NUMBER_COLUMNS_ISM_METADATA





class Metadata:
@@ -204,13 +204,9 @@ def write_ISM_metadata_in_file( 
        with open(csv_file, "w", newline="") as file:

            writer = csv.writer(file)

            for k in range(number_frames):

                row_list = [

                    "%+07.2f" % np.round(metadata[i][k, 0], 2),

                    "%+06.2f" % np.round(metadata[i][k, 1], 2),

                    "01.00",

                    "000.00",

                    "1.00",

                ]

                row_list = []

                for p in range(NUMBER_COLUMNS_ISM_METADATA):

                    row_list.append(FORMAT_ISM_METADATA_CSV[p] % np.round(metadata[i][k, p], 2))

                writer.writerow(row_list)



    return file_names