Commit 38d553d6 authored by PLAINSI's avatar PLAINSI
Browse files

Merge branch '47-p-800-add-hoa2-item-generation-to-scripts' into 'main' 

Resolve "(P-800) Add HOA2 item generation to scripts"

See merge request !90
parents 9f73b183 1a740cb3

item_gen_configs/HOA2_CONFIG.yml

0 → 100644
+61 −0
Original line number Diff line number Diff line 
---

################################################

# General configuration

################################################



### Output format

format: "HOA2"



### Output sampling rate in Hz needed for headerless audio files; default = 48000

fs: 48000



### IR sampling rate in Hz needed for headerless audio files; default = 48000

IR_fs: 48000



### Any relative paths will be interpreted relative to the working directory the script is called from!

### Usage of absolute paths is recommended.

### Do not use file names with dots "." in them! This is not supported, use "_" instead

### For Windows users: please use double back slash '\\' in paths and add '.exe' to executable definitions



### Input path to mono files

input_path: "./items_mono"



### Input path to stereo impulse response files, default = './ivas_processing_scripts/generation/IR'

IR_path: "./IRs"



### Output path for generated test items and metadata files

output_path: "./items_HOA2"



### (Optional) Output path for binauralized versions of the generated HOA2 items

# binaural_path: "./items_HOA2_bin"



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

loudness: -26



### Pre-amble and Post-amble length in seconds (default = 0.0)

preamble: 0.5

postamble: 1.0



### Flag for adding low-level random background noise (amplitude +-4) instead of silence; default = false (silence)

add_low_level_random_noise: true





################################################

### Scene description

################################################



### Each scene must start with the sceneN tag

### Specify the mono source filename (the program will search for it in the input_path folder)

### Specify the stereo IR source filename (the program will search for it in the input_path_IR folder)

### Specify the overlap length in seconds for each input source (negative value creates a gap)

### Note 1: use [val1, val2, ...] for multiple sources in a scene

### Note 2: use the "start:step:stop" notation for moving sources, where step will be applied in 20ms frames



scenes:

    a1: 

        name: "out.wav"

        description: ""

        source: ["fa1.wav", "ma1.wav"]

        IR: ["IR_HOA2_env1/FreefieldFloor_TalkPos1_EigenHoA2_SinSweep_9chn.wav", "IR_HOA2_env1/FreefieldFloor_TalkPos2_EigenHoA2_SinSweep_9chn.wav"]

        overlap: -0.2

item_gen_configs/P800-4.yml

+3 −0
Original line number Diff line number Diff line
@@ -26,6 +26,9 @@ IR_path: "./IRs" 
### Output path for generated test items and metadata files

output_path: "./items_FOA"



### (Optional) Output path for binauralized versions of the generated FOA items

# binaural_path: "./items_FOA_bin"



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

loudness: -26

item_gen_configs/P800-5.yml

+3 −0
Original line number Diff line number Diff line
@@ -26,6 +26,9 @@ IR_path: "./IRs" 
### Output path for generated test items and metadata files

output_path: "./items_FOA"



### (Optional) Output path for binauralized versions of the generated FOA items

# binaural_path: "./items_FOA_bin"



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

loudness: -26

ivas_processing_scripts/audiotools/wrappers/reverb.py

+65 −1
Original line number Diff line number Diff line
@@ -238,7 +238,7 @@ def reverb_foa( 
        H = fft(foa_IR.audio, axis=0)

        align = 1.0 / np.max(np.abs(H))



    # convolve mono input with left and right IR

    # convolve mono input with FOA IR

    y_w = reverb(input, IR_w, align=align)

    y_x = reverb(input, IR_x, align=align)

    y_y = reverb(input, IR_y, align=align)
@@ -251,3 +251,67 @@ def reverb_foa( 
    y.audio = np.column_stack([y_w.audio, y_x.audio, y_y.audio, y_z.audio])



    return y





def reverb_hoa2(

    input: Audio,

    hoa2_IR: Audio,

    align: Optional[float] = None,

) -> Audio:

    """

    Wrapper for the ITU-T reverb binary to convolve mono audio signal with an HOA2 impulse response



    Parameters

    ----------

    input: Audio

        Input audio signal

    IR: Audio

        Impulse response

    align: float

         multiplicative factor to apply to the reverberated sound in order to align its energy level with the second file



    Returns

    -------

    output: Audio

        Convolved audio signal with HOA2 IR

    """



    # convert to float32

    hoa2_IR.audio = np.float32(hoa2_IR.audio)



    numchannels = 9  # HOA2 by definition



    # calculate the scaling (multiplicative) factor such that the maximum gain of the IR filter across all frequencies is 0dB

    if align is None:

        H = fft(hoa2_IR.audio, axis=0)

        align = 1.0 / np.max(np.abs(H))



    IR = copy(hoa2_IR)

    IR.name = "MONO"

    IR.num_channels = 1

    ych = []

    for i in range(numchannels):

        # separate IR into each channel

        IR.audio = np.reshape(hoa2_IR.audio[:, i], (-1, 1))

        # convolve mono input with channel IR

        ych.append(reverb(input, IR, align=align))



    # combine into hoa2 output

    y = copy(input)

    y.name = "HOA2"

    y.num_channels = numchannels

    y.audio = np.column_stack(

        [

            ych[0].audio,

            ych[1].audio,

            ych[2].audio,

            ych[3].audio,

            ych[4].audio,

            ych[5].audio,

            ych[6].audio,

            ych[7].audio,

            ych[8].audio,

        ]

    )



    return y

ivas_processing_scripts/generation/__init__.py

+4 −0
Original line number Diff line number Diff line
@@ -43,6 +43,7 @@ from ivas_processing_scripts.constants import ( 
from ivas_processing_scripts.generation import (

    config,

    process_foa_items,

    process_hoa2_items,

    process_ism_items,

    process_stereo_items,

)
@@ -96,6 +97,9 @@ def main(args): 
    elif cfg.format == "FOA":

        # generate FOA items according to scene description

        process_foa_items.generate_foa_items(cfg, logger)

    elif cfg.format == "HOA2":

        # generate HOA2 items according to scene description

        process_hoa2_items.generate_hoa2_items(cfg, logger)



    # copy configuration to output directory

    with open(cfg.output_path.joinpath(f"{cfg.format}.yml"), "w") as f: