Commit 2a972eab authored by Lauros Pajunen's avatar Lauros Pajunen
Browse files

PI tests for audio focus

parent 116bacde

apps/decoder.c

+16 −6
Original line number Diff line number Diff line
@@ -328,15 +328,20 @@ static void IVAS_RTP_LogPiData( FILE *f_piDataOut, PIDATA_TS *piData, uint32_t n 
            break;

            case IVAS_PI_AUDIO_FOCUS_INDICATION:

            {

                fprintf( f_piDataOut, "{\n" );

                fprintf( f_piDataOut, "{" );

                if ( cur->data.focusIndication.availDirection )

                {

                    fprintf( f_piDataOut, "\t\t\t\"w\": %f,\n\t\t\t\"x\": %f,\n\t\t\t\"y\": %f,\n\t\t\t\"z\": %f \n\t\t\t",

                    fprintf( f_piDataOut, "\n\t\t\t\"direction\": {\n" );

                    fprintf( f_piDataOut, "\t\t\t\t\t\t\"w\": %f,\n\t\t\t\t\t\t\"x\": %f,\n\t\t\t\t\t\t\"y\": %f,\n\t\t\t\t\t\t\"z\": %f \n\t\t\t}",

                             cur->data.focusIndication.direction.w, cur->data.focusIndication.direction.x, cur->data.focusIndication.direction.y, cur->data.focusIndication.direction.z );

                    if ( cur->data.focusIndication.availLevel )

                    {

                        fprintf( f_piDataOut, "," );

                    }

                }

                if ( cur->data.focusIndication.availLevel )

                {

                    fprintf( f_piDataOut, "\t\t\t\"level\": %d", cur->data.focusIndication.flvl );

                    fprintf( f_piDataOut, "\n\t\t\t\"level\": %d", cur->data.focusIndication.flvl );

                }

                fprintf( f_piDataOut, "\n\t\t}" );

            }
@@ -355,15 +360,20 @@ static void IVAS_RTP_LogPiData( FILE *f_piDataOut, PIDATA_TS *piData, uint32_t n 
            case IVAS_PI_RESERVED13:

            case IVAS_PI_AUDIO_FOCUS_REQUEST:

            {

                fprintf( f_piDataOut, "{\n" );

                fprintf( f_piDataOut, "{" );

                if ( cur->data.focusRequest.availDirection )

                {

                    fprintf( f_piDataOut, "\t\t\t\"w\": %f,\n\t\t\t\"x\": %f,\n\t\t\t\"y\": %f,\n\t\t\t\"z\": %f \n\t\t\t",

                    fprintf( f_piDataOut, "\n\t\t\t\"direction\": {\n" );

                    fprintf( f_piDataOut, "\t\t\t\t\t\t\"w\": %f,\n\t\t\t\t\t\t\"x\": %f,\n\t\t\t\t\t\t\"y\": %f,\n\t\t\t\t\t\t\"z\": %f \n\t\t\t}",

                             cur->data.focusRequest.direction.w, cur->data.focusRequest.direction.x, cur->data.focusRequest.direction.y, cur->data.focusRequest.direction.z );

                    if ( cur->data.focusRequest.availLevel )

                    {

                        fprintf( f_piDataOut, "," );

                    }

                }

                if ( cur->data.focusRequest.availLevel )

                {

                    fprintf( f_piDataOut, "\t\t\t\"level\": %d", cur->data.focusRequest.flvl );

                    fprintf( f_piDataOut, "\n\t\t\t\"level\": %d", cur->data.focusRequest.flvl );

                }

                fprintf( f_piDataOut, "\n\t\t}" );

            }

tests/rtp/ivasrtp.py

+31 −11
Original line number Diff line number Diff line
@@ -42,7 +42,7 @@ import json 
import base64

import argparse

from pathlib import Path

from typing import cast

from typing import cast, Optional



NO_REQ="NO_REQ"
@@ -225,7 +225,6 @@ class AUDIO_FOCUS_LEVEL(int, Enum): 
    AUDIO_FOCUS_LEVEL_DEFAULT = 14

    AUDIO_FOCUS_LEVEL_NO_PREFERENCE = 15





@dataclass

class RTPHDR:

    version: int = 2
@@ -363,11 +362,8 @@ class ACOUSTIC_ENVIRONMENT: 


@dataclass

class AUDIO_FOCUS:

    w: float = 0.0

    x: float = 0.0

    y: float = 0.0

    z: float = 0.0

    level: AUDIO_FOCUS_LEVEL = AUDIO_FOCUS_LEVEL.AUDIO_FOCUS_LEVEL_DEFAULT

    direction: Optional[ORIENTATION] = None

    level: Optional[AUDIO_FOCUS_LEVEL] = None



@dataclass

class PIDATA:
@@ -599,6 +595,30 @@ def packAcousticEnv(bitstrm: BitStream, data: any): 
                absCoeff = mapNearestIndex(absorptionCoeffValues, aenv.abscoeff[n])

                bitstrm.append(f'uint:2={absCoeff}')



def unpackAudioFocus(bitstrm: ConstBitStream, piSize: int) -> AUDIO_FOCUS:

    assert piSize == 1 or piSize == 8 or piSize == 9, "Incorrect PI Data Size for AUDIO_FOCUS"

    direction = None

    level = None

    if piSize == 1:

        level = bitstrm.read(4).uint

        _ = bitstrm.read(4)

    else:

        direction = unpackOrientation(bitstrm, 8)

        if piSize == 9:

            level = bitstrm.read(4).uint

            _ = bitstrm.read(4)



    return AUDIO_FOCUS(direction=direction, level=level)



def packAudioFocus(bitstrm: BitStream, data: any):

    assert type(data) == AUDIO_FOCUS, "Diegetic type PI Data expects a data of type AUDIO_FOCUS"

    auFocus = cast(AUDIO_FOCUS, data)

    if auFocus.direction is not None:

        packOrientations(bitstrm, [auFocus.direction])

    if auFocus.level is not None:

        bitstrm.append(f'uint:4={auFocus.level}')

        bitstrm.append(f'uint:4=0')





PIDataUnpacker = [

    unpackOrientation, # SCENE_ORIENTATION,
@@ -615,13 +635,13 @@ PIDataUnpacker = [ 
    unpackUnsupported, # ISM_DIRECTIVITY

    unpackDiegetic,    # DIEGETIC_TYPE

    unpackUnsupported, # RESERVED13

    unpackAudioFocusCommon, # AUDIO_FOCUS_INDICATION

    unpackAudioFocus,  # AUDIO_FOCUS_INDICATION

    unpackUnsupported, # RESERVED15

    unpackOrientation, # PLAYBACK_DEVICE_ORIENTATION

    unpackOrientation, # HEAD_ORIENTATION

    unpackPosition,    # LISTENER_POSITION

    unpackDAS,         # DYNAMIC_AUDIO_SUPPRESSION

    unpackAudioFocusCommon, # AUDIO_FOCUS_REQUEST

    unpackAudioFocus,  # AUDIO_FOCUS_REQUEST

    unpackUnsupported, # PI_LATENCY

    unpackUnsupported, # R_ISM_ID

    unpackUnsupported, # R_ISM_GAIN
@@ -650,13 +670,13 @@ PIDataPacker = [ 
    packUnsupported, # ISM_DIRECTIVITY

    packDiegetic,    # DIEGETIC_TYPE

    packUnsupported, # RESERVED13

    packUnsupported, # RESERVED14

    packAudioFocus,  # AUDIO_FOCUS_INDICATION

    packUnsupported, # RESERVED15

    packOrientation, # PLAYBACK_DEVICE_ORIENTATION

    packOrientation, # HEAD_ORIENTATION

    packPosition,    # LISTENER_POSITION

    packDAS,         # DYNAMIC_AUDIO_SUPPRESSION

    packOrientation, # AUDIO_FOCUS_DIRECTION

    packAudioFocus,  # AUDIO_FOCUS_DIRECTION

    packUnsupported, # PI_LATENCY

    packUnsupported, # R_ISM_ID

    packUnsupported, # R_ISM_GAIN

tests/rtp/test_rtp.py

+36 −15
Original line number Diff line number Diff line
@@ -102,9 +102,14 @@ def test_rtp_bitstream_evs ( 
    )





@pytest.mark.parametrize("bitrate", [24400, 80000, 512000])

@pytest.mark.parametrize("bandwidth", ["WB", "SWB", "FB"])

@pytest.mark.parametrize("format", ["STEREO", "SBA", "MC", "MASA"])

#@pytest.mark.parametrize("bitrate", [24400, 80000, 512000])

#@pytest.mark.parametrize("bandwidth", ["WB", "SWB", "FB"])

#@pytest.mark.parametrize("format", ["STEREO", "SBA", "MC", "MASA"])

#@pytest.mark.parametrize("framesPerPacket", [8])



@pytest.mark.parametrize("bitrate", [24400])

@pytest.mark.parametrize("bandwidth", ["SWB"])

@pytest.mark.parametrize("format", ["MASA"])

@pytest.mark.parametrize("framesPerPacket", [8])

def test_rtp_bitstream_ivas_nodtx (

    test_info,
@@ -164,21 +169,24 @@ def generatePiData(startTs: int, endTs: int) -> dict: 
    someDesc = lambda : AUDIO_DESCRIPTION(isSpeech=bool(random.getrandbits(1)), isMusic=bool(random.getrandbits(1)), isAmbiance=bool(random.getrandbits(1)), isEditable=bool(random.getrandbits(1)), isBinaural=bool(random.getrandbits(1)))

    someDAS = lambda : DYNAMIC_AUDIO_SUPPRESSION(preferSpeech=bool(random.getrandbits(1)), preferMusic=bool(random.getrandbits(1)), preferAmbiance=bool(random.getrandbits(1)), level=random.randint(0, 15))

    someDIG = lambda : DIEGETIC_TYPE(isDigetic=[ bool(random.getrandbits(1)) for _ in range(random.randint(1, 5)) ])

    someAFocus = lambda : AUDIO_FOCUS(w=2*random.random()-1.0, x=2*random.random()-1.0, y=2*random.random()-1.0, z=2*random.random()-1.0, level=random.randint(0, 15))

    someAuFocusDirLvl = lambda : AUDIO_FOCUS(ORIENTATION(w=2*random.random()-1.0, x=2*random.random()-1.0, y=2*random.random()-1.0, z=2*random.random()-1.0), level=AUDIO_FOCUS_LEVEL(random.randint(0, 15)))

    someAuFocusDir = lambda : AUDIO_FOCUS(ORIENTATION(w=2*random.random()-1.0, x=2*random.random()-1.0, y=2*random.random()-1.0, z=2*random.random()-1.0))

    someAuFocusLvl = lambda : AUDIO_FOCUS(level=AUDIO_FOCUS_LEVEL(random.randint(0, 15)))

    someAuFocusList = [someAuFocusDirLvl, someAuFocusDir, someAuFocusLvl]



    for ts in range(startTs, endTs, 320):

        pidata = dict()

        pidata["SCENE_ORIENTATION"] = someOrientation()

        pidata["DEVICE_ORIENTATION_COMPENSATED"] = someOrientation()

        pidata["DEVICE_ORIENTATION_UNCOMPENSATED"] = someOrientation()

        pidata["PLAYBACK_DEVICE_ORIENTATION"] = someOrientation()

        pidata["HEAD_ORIENTATION"] = someOrientation()

        pidata["AUDIO_FOCUS_REQUEST"] = someAFocus()

        pidata["LISTENER_POSITION"] = somePosition()

        pidata["DYNAMIC_AUDIO_SUPPRESSION"] = someDAS()

        pidata["AUDIO_DESCRIPTION"] = [someDesc() for n in range(random.randint(1, 5))]

        pidata["DIEGETIC_TYPE"] = someDIG()

        pidata["ACOUSTIC_ENVIRONMENT"] = ACOUSTIC_ENVIRONMENT(aeid=random.randint(0, 127))

        #pidata["SCENE_ORIENTATION"] = someOrientation()

        #pidata["DEVICE_ORIENTATION_COMPENSATED"] = someOrientation()

        #pidata["DEVICE_ORIENTATION_UNCOMPENSATED"] = someOrientation()

        #pidata["PLAYBACK_DEVICE_ORIENTATION"] = someOrientation()

        #pidata["HEAD_ORIENTATION"] = someOrientation()

        pidata["AUDIO_FOCUS_REQUEST"] = random.choice(someAuFocusList)()

        #pidata["LISTENER_POSITION"] = somePosition()

        #pidata["DYNAMIC_AUDIO_SUPPRESSION"] = someDAS()

        #pidata["AUDIO_DESCRIPTION"] = [someDesc() for n in range(random.randint(1, 5))]

        #pidata["DIEGETIC_TYPE"] = someDIG()

        #pidata["ACOUSTIC_ENVIRONMENT"] = ACOUSTIC_ENVIRONMENT(aeid=random.randint(0, 127))

        data[str(ts)] = pidata

    return data
@@ -225,6 +233,17 @@ def isEqualAcousticEnv(ref: ACOUSTIC_ENVIRONMENT, dut: ACOUSTIC_ENVIRONMENT): 
    for r, d in zip(ref.rt60, dut.rt60):

        assert r == d, f"Acoustic Env PI Data mismatch in rt60 {r} != {d}"



def isEqualAudioFocus(ref: AUDIO_FOCUS, dut: AUDIO_FOCUS):

    if ref.direction is not None or dut.direction is not None:

        assert ref.direction is not None, "Audio Focus PI Data missing direction"

        assert dut.direction is not None, "Audio Focus PI Data missing direction"

    if ref.direction is not None and dut.direction is not None:

        assert abs(ref.direction.w - dut.direction.w) < 0.0001, "Audio Focus PI Data mismatch in direction w"

        assert abs(ref.direction.x - dut.direction.x) < 0.0001, "Audio Focus PI Data mismatch in direction x"

        assert abs(ref.direction.y - dut.direction.y) < 0.0001, "Audio Focus PI Data mismatch in direction y"

        assert abs(ref.direction.z - dut.direction.z) < 0.0001, "Audio Focus PI Data mismatch in direction z"

    assert ref.level == dut.level, "Audio Focus PI Data mismatch in level"



class CSVREADER:

    def __init__(self, csvFile: Path):

        self.rIdx = 0
@@ -448,6 +467,8 @@ def run_rtp_bitstream_tests ( 
                        isEqualDiegetic(DIEGETIC_TYPE(**decoded), data)

                    elif type(generatedPIData[ts][pitype]) == ACOUSTIC_ENVIRONMENT:

                        isEqualAcousticEnv(ACOUSTIC_ENVIRONMENT(**decoded), data)

                    if type(generatedPIData[ts][pitype]) == AUDIO_FOCUS:

                        isEqualAudioFocus(AUDIO_FOCUS(**decoded), data)

                    elif type(generatedPIData[ts][pitype]) == list:

                        for r, d in zip(generatedPIData[ts][pitype], decodedPiData[ts][pitype]):

                            isEqualAD(AUDIO_DESCRIPTION(**d), r)