Commit 4651439f authored by Vladimir Malenovsky's avatar Vladimir Malenovsky
Browse files

fix OMASA item generation

parent 0375de2a

ivas_processing_scripts/audiotools/convert/osba.py

+12 −6
Original line number Diff line number Diff line
@@ -129,14 +129,20 @@ def convert_osba( 
            )



        # only render SBA part to MASA

        out_sba = audio.fromtype(out.name[4:])

        out_sba.metadata_file = out.metadata_files[-1]

        render_sba_to_masa(sba, out_sba)

        out_masa = audio.fromtype(out.name[4:])

        if out.metadata_files[-1].endswith(".met"):

            # if MASA metadata filename is given, copy it

            out_masa.metadata_file = copy(out.metadata_files[-1])

        render_sba_to_masa(sba, out_masa)



        # combine ISM audio with MASA audio into OMASA audio

        out.audio = np.concatenate((oba.audio, out_masa.audio), axis=1)



        # ism audio is passed through

        ism_audio = osba.audio[:, : osba.num_ism_channels]

        # combine ISM metadata with MASA metadata filenames

        out.metadata_files = oba.metadata_files + [out_masa.metadata_file]



        out.audio = np.concatenate((ism_audio, out_sba.audio), axis=1)

        # copy ISM object positions

        out.object_pos = copy(oba.object_pos)



    # OSBA -> OSBA

    elif isinstance(out, audio.OSBAAudio):

ivas_processing_scripts/audiotools/convert/scenebased.py

+11 −0
Original line number Diff line number Diff line
@@ -197,6 +197,17 @@ def render_sba_to_masa( 
    masa_out: audio.MetadataAssistedSpatialAudio,

) -> None:

    num_tcs = masa_out.num_channels



    # convert to HOA2 if input is HOA3

    if sba_in.name.endswith("HOA3"):

        warn(

            "MASA conversion only supports up to 2nd order ambisonics! Converting to HOA2 format."

        )

        sba_hoa2 = audio.fromtype("HOA2")

        sba_hoa2.fs = sba_in.fs

        render_sba_to_sba(sba_in, sba_hoa2)

        sba_in = sba_hoa2



    masa = masaAnalyzer(sba_in, num_tcs, masa_out.dirs, masa_out.metadata_file)

    masa_out.audio = masa.audio

ivas_processing_scripts/generation/generate_ismN_items.py

+4 −4
Original line number Diff line number Diff line
@@ -288,12 +288,12 @@ def generate_ismN_scene( 
            x.fs = cfg.fs



        # adjust the level of the audio source file (need to convert to MONO first)

        if np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        elif level is None:

        if level is None:

            # do not change the level of the audio source signal

            logger.info("-- Level of the audio source signal is not changed")

        elif np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        else:

            x_temp = audio.ChannelBasedAudio(

                "MONO"

ivas_processing_scripts/generation/generate_masa_items.py

+4 −4
Original line number Diff line number Diff line
@@ -316,12 +316,12 @@ def generate_MASA_scene( 
            x = reverb_hoa3(x, IR, mode=None)



        # adjust the level of the FOA/HOA2/HOA3 signal

        if np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        elif level is None:

        if level is None:

            # do not change the level of the audio source signal

            logger.info("-- Level of the audio source signal is not changed")

        elif np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        else:

            x.audio, _ = loudness_norm(x, level, loudness_format="STEREO")

ivas_processing_scripts/generation/generate_mc_items.py

+4 −4
Original line number Diff line number Diff line
@@ -316,12 +316,12 @@ def generate_MC_scene( 
            x = reverb_hoa3(x, IR, mode=None)



        # adjust the level of the FOA/HOA2/HOA3 signal

        if np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        elif level is None:

        if level is None:

            # do not change the level of the audio source signal

            logger.info("-- Level of the audio source signal is not changed")

        elif np.isinf(level):

            # set all channels to zero

            x.audio = np.zeros_like(x.audio)

        else:

            x.audio, _ = loudness_norm(x, level, loudness_format="STEREO")