help="volumes of input sound (0<=v<=1)",
default=1.0,
type=str,
metavar="VOL")
(options, args) = parser.parse_args()
# argv check
if len(args) < 1:
quit()
#
npr.seed(1234)
src_volume = options.volume
output_filename = options.output_file
data = []
#
wav_filename = args[0]
print "... reading", wav_filename
wav_data = simmch.read_mch_wave(wav_filename)
wav = wav_data["wav"]
fs = wav_data["framerate"]
nch = wav_data["nchannels"]
amp = np.max(np.abs(wav))
print "[INFO] max amplitude:", amp
g = 32767.0 / amp * src_volume
print "[INFO] gain:", g
wav = wav * g
# save data
if output_filename != None:
simmch.save_mch_wave(wav, output_filename)
spec2=simmch.stft_mch(wav2,win,step)
##
##
nframe1=spec1.shape[1]
nframe2=spec2.shape[1]
nframe=min(nframe1,nframe2)
spec1_temp=spec1[:,0:nframe,min_freq_bin:max_freq_bin]
spec2_temp=spec2[:,0:nframe,min_freq_bin:max_freq_bin]
if options.noise is not None:
w=wiener_filter_eigen(spec1_temp,spec2_temp,win_size=nframe,r_step=1)
print "# filter:",w.shape
out_spec=apply_filter_eigen(spec1_temp,w)
# ISTFT
recons=simmch.istft_mch(out_spec, win, step)
simmch.save_mch_wave(recons*32767.0,"recons_eigen.wav")
quit()
#print spec1_temp.shape
#print spec2_temp.shape
#win_size=50
# spec[ch, frame, freq_bin]
# w[ch2,ch1]
w,_,_=wiener_filter_freq(spec1_temp,spec2_temp,win_size=nframe,r_step=1)
print "# filter:",w.shape
if options.tf is not None:
tf_config=read_hark_tf(options.tf)
if(len(w.shape)==3):
for i in xrange(len(w.shape)):
save_sidelobe("sidelobe_wiener%i.png"%(i+1),tf_config,w[:,:,i],sidelobe_freq_bin)
else:
save_sidelobe("sidelobe_wiener.png",tf_config,w,sidelobe_freq_bin)
default=1.0, type=str, metavar="VOL") (options, args) = parser.parse_args() # argv check if len(args)<1: quit() # npr.seed(1234) src_volume=options.volume output_filename=options.output_file data=[] # wav_filename=args[0] print "... reading", wav_filename wav_data=simmch.read_mch_wave(wav_filename) wav=wav_data["wav"] fs=wav_data["framerate"] nch=wav_data["nchannels"] amp=np.max(np.abs(wav)) print "[INFO] max amplitude:",amp g=32767.0/amp*src_volume print "[INFO] gain:",g wav=wav*g # save data if output_filename!=None: simmch.save_mch_wave(wav,output_filename)
if not src_index in tf_config["tf"]: print >>sys.stderr, "Error: tf index",src_index,"does not exist in TF file" quit() ## read wav file print "... reading", wav_filename wav_data=simmch.read_mch_wave(wav_filename) scale=32767.0 wav=wav_data["wav"]/scale fs=wav_data["framerate"] nch=wav_data["nchannels"] ## print info print "# channel num : ", nch print "# sample size : ", wav.shape print "# sampling rate : ", fs print "# sec : ", wav_data["duration"] mono_wavdata = wav[0,:] ## apply TF fftLen = 512 step = fftLen / 4 mch_wavdata=apply_tf(mono_wavdata,fftLen, step,tf_config,src_index) a=np.max(mch_wavdata) mch_wavdata=mch_wavdata/a mch_wavdata=mch_wavdata*scale*src_volume ## save data simmch.save_mch_wave(mch_wavdata,output_filename)
nsamples=options.samples length=((nsamples-(fftLen-step))-1)/step+1 elif options.frames!=None: length=options.frames nsamples=length*step+(fftLen-step) elif options.duration!=None: nsamples=options.samplingrate*options.duration length=((nsamples-(fftLen-step))-1)/step+1 elif options.template!=None: wav_filename=options.template print "... reading", wav_filename wav_data=simmch.read_mch_wave(wav_filename) nsamples=wav_data["nframes"] nch=wav_data["nchannels"] length=((nsamples-(fftLen-step))-1)/step+1 else: print >>sys.stderr,"[ERROR] unknown duration (please indicate --duration, --samples, or --frames)" quit() print "#channels:",nch print "#frames:",length print "#samples:",nsamples print "window size:",fftLen mch_wavdata=make_white_noise(nch,length,fftLen,step) g=32767.0*options.amp print "[INFO] gain:",g mch_wavdata=mch_wavdata*g simmch.save_mch_wave(mch_wavdata,output_filename,sample_width=2,framerate=options.samplingrate)
src_index = simmch.nearest_direction_index(tf_config, src_theta)
print "# mic positions :", mic_pos
print "# direction index:", src_index
if not src_index in tf_config["tf"]:
print >> sys.stderr, "Error: tf index", src_index, "does not exist in TF file"
quit()
## read wav file
print "... reading", wav_filename
wav_data = simmch.read_mch_wave(wav_filename)
scale = 32767.0
wav = wav_data["wav"] / scale
fs = wav_data["framerate"]
nch = wav_data["nchannels"]
## print info
print "# channel num : ", nch
print "# sample size : ", wav.shape
print "# sampling rate : ", fs
print "# sec : ", wav_data["duration"]
mono_wavdata = wav[0, :]
## apply TF
fftLen = 512
step = fftLen / 4
mch_wavdata = apply_tf(mono_wavdata, fftLen, step, tf_config, src_index)
a = np.max(mch_wavdata)
mch_wavdata = mch_wavdata / a
mch_wavdata = mch_wavdata * scale * src_volume
## save data
simmch.save_mch_wave(mch_wavdata, output_filename)
segment_enabled_flag=True
if segment_enabled_flag:
start_fram=m
print "... detected segments by frame-based thresholds"
sample_segs=[(seg[0]*step,seg[1]*step+length)for seg in frame_segs]
print sample_segs
print "... detected segments"
filtered_segs=[seg for seg in sample_segs if seg[1]-seg[0]>thresh_seg_size*fs]
print filtered_segs
for seg_id,s in enumerate(filtered_segs):
w=wav[:,s[0]:s[1]]
# save data
if options.output_file!=None:
o=options.output_file%seg_id
print "[save]",o
simmch.save_mch_wave(w,o)
#
plt.subplot(4, 1, 1)
disable_xy_label=False
if disable_xy_label:
plt.tick_params(labelbottom='off')
plt.tick_params(labelleft='off')
plt.xlabel("time [second]")
plt.ylabel("wav")
plt.plot(wav[ch])
#
plt.subplot(4, 1, 2)
disable_xy_label=False
if disable_xy_label:
plt.tick_params(labelbottom='off')
plt.tick_params(labelleft='off')
nframe1 = spec1.shape[1]
nframe2 = spec2.shape[1]
nframe = min(nframe1, nframe2)
spec1_temp = spec1[:, 0:nframe, min_freq_bin:max_freq_bin]
spec2_temp = spec2[:, 0:nframe, min_freq_bin:max_freq_bin]
if options.noise is not None:
w = wiener_filter_eigen(spec1_temp,
spec2_temp,
win_size=nframe,
r_step=1)
print "# filter:", w.shape
out_spec = apply_filter_eigen(spec1_temp, w)
# ISTFT
recons = simmch.istft_mch(out_spec, win, step)
simmch.save_mch_wave(recons * 32767.0, "recons_eigen.wav")
quit()
#print spec1_temp.shape
#print spec2_temp.shape
#win_size=50
# spec[ch, frame, freq_bin]
# w[ch2,ch1]
w, _, _ = wiener_filter_freq(spec1_temp,
spec2_temp,
win_size=nframe,
r_step=1)
print "# filter:", w.shape
if options.tf is not None:
tf_config = read_hark_tf(options.tf)
if (len(w.shape) == 3):
for i in xrange(len(w.shape)):
#mv_freq=np.array([mv_freq])
save_sidelobe("sidelobe_mv.png",tf_config,w_mv,sidelobe_freq_bin)
### GSC for MV beamformer
w_a,_,_=wiener_filter.wiener_filter_freq(blocked_freq,mv_freq)
y_mv=wiener_filter.apply_filter_freq(blocked_freq,w_a)
w_gsc_mv=np.zeros((nfreq_bin,nch),dtype=complex)
for freq_bin in xrange(nfreq_bin):
w_gsc_mv[freq_bin,:]=w_mv[freq_bin,:]-w_a[freq_bin,:].dot(B.dot(A[freq_bin,:,:]))
save_sidelobe("sidelobe_gsc_mv.png",tf_config,w_gsc_mv,sidelobe_freq_bin,clear_flag=False)
###
out_gsc_ds=ds_freq-y_ds
out_gsc_mv=mv_freq-y_mv
recons_out_gsc_ds=simmch.istft_mch(out_gsc_ds, win, step)
recons_out_gsc_mv=simmch.istft_mch(out_gsc_mv, win, step)
recons_ds_y=simmch.istft_mch(y_ds, win, step)
recons_mv_y=simmch.istft_mch(y_mv, win, step)
recons_b=simmch.istft_mch(blocked_freq, win, step)
recons_ds=simmch.istft_mch(ds_freq, win, step)
recons_mv=simmch.istft_mch(mv_freq, win, step)
simmch.save_mch_wave(recons_mv*32767.0,"mv.wav")
simmch.save_mch_wave(recons_ds*32767.0,"ds.wav")
simmch.save_mch_wave(recons_ds_y*32767.0,"y_ds.wav")
simmch.save_mch_wave(recons_mv_y*32767.0,"y_mv.wav")
simmch.save_mch_wave(recons_out_gsc_ds*32767.0,"gsc_ds.wav")
simmch.save_mch_wave(recons_out_gsc_mv*32767.0,"gsc_mv.wav")
simmch.save_mch_wave(recons_b*32767.0,"b.wav")
quit()
