Commit 8cfdaa57 authored by janssontoftg's avatar janssontoftg
Browse files

Correction of preamble for P800-1 and P800-2. 

Moved 20 ms block alignment after addition of source signals, to ensure the overlap period is maintained as specified.
parent ce1d51a8

item_gen_configs/P800-1.yml

+1 −1
Original line number Diff line number Diff line
@@ -30,7 +30,7 @@ output_path: "experiments/selection/P800-1/proc_input" 
loudness: -26



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

preamble: 1.0

preamble: 0.5

postamble: 1.0



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

item_gen_configs/P800-2.yml

+1 −1
Original line number Diff line number Diff line
@@ -30,7 +30,7 @@ output_path: "experiments/selection/P800-2/proc_input" 
loudness: -26



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

preamble: 1.0

preamble: 0.5

postamble: 1.0



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

ivas_processing_scripts/generation/process_stereo_items.py

+9 −9
Original line number Diff line number Diff line
@@ -151,15 +151,6 @@ def generate_stereo_scene( 
            pre = np.zeros((N_delay, x.audio.shape[1]))

            x.audio = np.concatenate([pre, x.audio])



        # pad with zeros to ensure that the signal length is a multiple of 20ms

        N_frame = x.fs / 50

        if len(x.audio) % N_frame != 0:

            N_pad = int(N_frame - len(x.audio) % N_frame)



            # insert all-zero preamble

            pre = np.zeros((N_pad, x.audio.shape[1]))

            x.audio = np.concatenate([pre, x.audio])



        # add source signal to the array of source signals

        y.fs = x.fs

        if y.audio is None:
@@ -211,6 +202,15 @@ def generate_stereo_scene( 
        post = np.zeros((N_post, y.audio.shape[1]))

        y.audio = np.concatenate([y.audio, post])



    # pad with zeros to ensure that the signal length is a multiple of 20ms

    N_frame = y.fs / 50

    if y.audio.shape[0] % N_frame != 0:

        N_pad = int(N_frame - y.audio.shape[0] % N_frame)



        # insert all-zero postamble

        post = np.zeros((N_pad, y.audio.shape[1]))

        y.audio = np.concatenate([y.audio, post])



    # add random noise

    if cfg.add_low_level_random_noise:

        # create uniformly distributed noise between -4 and 4