Commit 166ecacc authored by Archit Tamarapu's avatar Archit Tamarapu
Browse files

- add missing errors for delay and trim related to metadata based formats 

- simplify isinstance() checks to check against a tuple for better readability
- add missing checks against OMASA and OSBA in the above instances
parent a7f78b6b

ivas_processing_scripts/audiotools/convert/__init__.py

+19 −14
Original line number Diff line number Diff line
@@ -163,9 +163,7 @@ def convert_file( 
    # write output audio

    write(out_file, output.audio, output.fs)

    # write metadata

    if isinstance(output, audio.ObjectBasedAudio) or isinstance(

        output, audio.OSBAAudio

    ):

    if isinstance(output, (audio.ObjectBasedAudio, audio.OSBAAudio)):

        write_ISM_metadata_in_file(output.object_pos, [out_file], automatic_naming=True)

    elif isinstance(output, audio.MetadataAssistedSpatialAudio) and in_fmt == out_fmt:

        # audio objects point to same MD file, create new one with default naming for output
@@ -292,16 +290,26 @@ def process_audio( 


    """delay audio"""

    if delay is not None:

        if isinstance(x, audio.ObjectBasedAudio):

            raise ValueError("Delay not possible for ISM input")

        if isinstance(

            x,

            (

                audio.ObjectBasedAudio,

                audio.MetadataAssistedSpatialAudio,

                audio.OMASAAudio,

                audio.OSBAAudio,

            ),

        ):

            raise ValueError("Delay not possible for ISM, MASA, OMASA and OSBA inputs")

        if logger:

            logger.debug(f"Delaying audio by {delay} ms")

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



    """trim or pad audio"""

    if trim is not None:

        if isinstance(x, audio.ObjectBasedAudio):

            # metadata concatenation necessary for ISM

        if isinstance(x, (audio.MetadataAssistedSpatialAudio, audio.OMASAAudio)):

            raise ValueError("Trimming not possible for MASA and OMASA inputs")

        elif isinstance(x, (audio.ObjectBasedAudio, audio.OSBAAudio)):

            # metadata concatenation necessary for ISM/OSBA

            metadata.trim_meta(x, tuple(trim), pad_noise)

        else:

            x.audio = audioarray.trim(x.audio, x.fs, tuple(trim), pad_noise)
@@ -388,12 +396,11 @@ def format_conversion( 
    # validation

    # check for MASA/OMASA as output

    if isinstance(output, audio.MetadataAssistedSpatialAudio) and not (

        isinstance(input, audio.SceneBasedAudio)

        or isinstance(input, audio.MetadataAssistedSpatialAudio)

        isinstance(input, (audio.MetadataAssistedSpatialAudio, audio.SceneBasedAudio))

    ):

        raise NotImplementedError("Can only convert to MASA from SBA")

    if isinstance(output, audio.OMASAAudio) and not (

        isinstance(input, audio.OSBAAudio) or isinstance(input, audio.OMASAAudio)

        isinstance(input, (audio.OSBAAudio, audio.OMASAAudio))

    ):

        raise NotImplementedError("Can only convert to OMASA from OSBA")
@@ -402,10 +409,8 @@ def format_conversion( 
        raise NotImplementedError(

            "ISM is not supported as an output for rendering! Only usable as pass-through"

        )

    if isinstance(output, audio.OMASAAudio) or isinstance(output, audio.OSBAAudio):

        if not (

            isinstance(input, audio.OMASAAudio) or isinstance(input, audio.OSBAAudio)

        ):

    if isinstance(output, (audio.OMASAAudio, audio.OSBAAudio)):

        if not isinstance(input, (audio.OMASAAudio, audio.OSBAAudio)):

            raise NotImplementedError(

                "OMASA and OSBA only possible as output if input is OMASA or OSBA"

            )

ivas_processing_scripts/audiotools/wrappers/bs1770.py

+5 −9
Original line number Diff line number Diff line
@@ -89,9 +89,7 @@ def bs1770demo( 
            'A bs1770demo executable without RMS support (cmdl option "-rms") was detected. Please update the bs1770demo executable. See bin/README.md for details.'

        )



    if not isinstance(input, audio.BinauralAudio) and not isinstance(

        input, audio.ChannelBasedAudio

    ):

    if not isinstance(input, (audio.BinauralAudio, audio.ChannelBasedAudio)):

        raise NotImplementedError(

            f"{input.name} is unsupported in ITU-R BS.1770-4. Ensure loudness format was set correctly in CLI/YAML."

        )
@@ -194,14 +192,12 @@ def get_loudness( 


    if loudness_format is None:

        # for some formats rendering is necessary prior to loudness measurement

        if isinstance(input, audio.SceneBasedAudio) or isinstance(

            input, audio.MetadataAssistedSpatialAudio

        if isinstance(

            input, (audio.SceneBasedAudio, audio.MetadataAssistedSpatialAudio)

        ):

            loudness_format = "7_1_4"

        elif (

            isinstance(input, audio.ObjectBasedAudio)

            or isinstance(input, audio.OMASAAudio)

            or isinstance(input, audio.OSBAAudio)

        elif isinstance(

            input, (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio)

        ):

            loudness_format = "BINAURAL"

        elif hasattr(input, "layout_file"):

ivas_processing_scripts/processing/ivas.py

+1 −3
Original line number Diff line number Diff line
@@ -152,9 +152,7 @@ class IVAS(Processing): 
        if isinstance(self.in_fmt, audio.MetadataAssistedSpatialAudio):

            md_file = in_file.parent / (in_file.name + ".met")

            metadata_files.append(md_file)

        elif isinstance(self.in_fmt, audio.ObjectBasedAudio) or isinstance(

            self.in_fmt, audio.OSBAAudio

        ):

        elif isinstance(self.in_fmt, (audio.ObjectBasedAudio, audio.OSBAAudio)):

            metadata_files = in_meta

        elif isinstance(self.in_fmt, audio.OMASAAudio):

            metadata_files = in_meta

ivas_processing_scripts/processing/preprocessing_2.py

+2 −2
Original line number Diff line number Diff line
@@ -64,8 +64,8 @@ class Preprocessing2(Processing): 
            self.in_fmt, in_file, fs=self.in_fs, in_meta=in_meta

        )



        if isinstance(audio_object, audio.MetadataAssistedSpatialAudio) or isinstance(

            audio_object, audio.OMASAAudio

        if isinstance(

            audio_object, (audio.MetadataAssistedSpatialAudio, audio.OMASAAudio)

        ):

            if self.preamble > 0 or self.background_noise or self.repeat_signal:

                raise ValueError(

ivas_processing_scripts/processing/processing_splitting_scaling.py

+34 −9
Original line number Diff line number Diff line
@@ -60,7 +60,10 @@ class Processing_splitting_scaling(Processing): 


        # read file and metadata

        x, fs = read(in_file, nchannels=num_channels, fs=self.fs)

        if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

        if isinstance(

            audio.fromtype(self.out_fmt),

            (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

        ):

            meta_arrays = []

            for meta in in_meta:

                meta_arrays.append(np.genfromtxt(meta, delimiter=","))
@@ -72,7 +75,10 @@ class Processing_splitting_scaling(Processing): 
            in_file_noerror = Path(f"{in_file.with_suffix('')}.noerror.wav")

            out_file_noerror = Path(f"{out_file.with_suffix('')}.noerror.wav")

            x_noerror, _ = read(in_file_noerror, nchannels=num_channels, fs=self.fs)

            if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

            if isinstance(

                audio.fromtype(self.out_fmt),

                (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

            ):

                meta_arrays_noerror = []

                for meta in in_meta:

                    path_parts = str(meta).split(".")
@@ -153,7 +159,10 @@ class Processing_splitting_scaling(Processing): 
                    str(f).replace(tmp_name, out_name).replace(f".{self.name}.", ".")

                )

            out_out_files.append(f_out)

        if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

        if isinstance(

            audio.fromtype(self.out_fmt),

            (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

        ):

            out_out_meta = []

            for f in out_out_files:

                oom = []
@@ -170,7 +179,10 @@ class Processing_splitting_scaling(Processing): 
        ):

            write(of, file_s, fs)

            write(oof, file_s, fs)

            if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

            if isinstance(

                audio.fromtype(self.out_fmt),

                (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

            ):

                write_ISM_metadata_in_file(meta_s, om)

                write_ISM_metadata_in_file(meta_s, oom)

        # write noerror files in tmp folder
@@ -182,7 +194,10 @@ class Processing_splitting_scaling(Processing): 
                meta_splits_noerror,

            ):

                write(ofne, file_sne, fs)

                if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

                if isinstance(

                    audio.fromtype(self.out_fmt),

                    (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

                ):

                    write_ISM_metadata_in_file(meta_sne, omne)



    def revert_preamble_concatenation(
@@ -257,7 +272,10 @@ class Processing_splitting_scaling(Processing): 
                    for sn in split_names

                ]



            if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

            if isinstance(

                audio.fromtype(self.out_fmt),

                (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

            ):

                out_meta = []

                for of in out_files:

                    of_list = []
@@ -271,7 +289,10 @@ class Processing_splitting_scaling(Processing): 
            # set output values

            out_files = [out_file]

            file_splits = [x]

            if isinstance(audio.fromtype(self.out_fmt), audio.ObjectBasedAudio):

            if isinstance(

                audio.fromtype(self.out_fmt),

                (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio),

            ):

                meta_splits = [in_meta]

                out_meta = [

                    [
@@ -305,7 +326,9 @@ def adjust_loudness( 
    scaled_signals = []

    for f, m in zip(file_splits, meta):

        audio_object = audio.fromarray(fmt=out_fmt, x=f, fs=fs)

        if isinstance(audio_object, audio.ObjectBasedAudio):

        if isinstance(

            audio_object, (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio)

        ):

            audio_object.object_pos = m

        scaled_signal, _ = loudness_norm(

            audio_object, loudness, loudness_fmt, logger=logger
@@ -322,7 +345,9 @@ def measure_loudness(file_splits, out_fmt, fs, loudness, loudness_fmt, meta, log 
    scaling_splits = []

    for f, m in zip(file_splits, meta):

        audio_object = audio.fromarray(fmt=out_fmt, x=f, fs=fs)

        if isinstance(audio_object, audio.ObjectBasedAudio):

        if isinstance(

            audio_object, (audio.ObjectBasedAudio, audio.OMASAAudio, audio.OSBAAudio)

        ):

            audio_object.object_pos = m

        _, scale_factor = loudness_norm(

            audio_object, loudness, loudness_fmt, logger=logger