def fun2(hk,fp): start = dt.now() os.remove(filepath) rec(hk,fp) print(search.searchA(transcribe(fp),[s[:-4] for s in songs])) end = dt.now() elapsed=end-start print(elapsed.total_seconds())
def searchThenPlay():
global songnum
lap()
rec(searchHotkey,recordingfile)
print('listening took ',lap(),' seconds')
transcription = transcribe(recordingfile)
print('transcribing took ',lap(),' seconds and I think you said ',transcription)
found = search(transcription,[s[:-4] for s in songs])
print('searching took ',lap(),' seconds and found the song ',found)
index = songs.index(found+'.mp3')
if isinstance(index,int):
songnum=index-1
next()
def voice_recognize():
record.rec('test1.wav') #生成录音
time_start = time.time() #计时开始
pcm_file = wav2pcm.wav_to_pcm('test1.wav') #将wav转成pcm
# 识别本地文件
result = client.asr(get_file_content(pcm_file), 'pcm', 16000, {
'dev_pid': 1536,
})
if (result.get("err_no") == 0):
time_end = time.time() #计时结束
res_str = '识别结果：%s,耗时%.2fs' % (result.get("result")[0],
(time_end - time_start)) #截取结果
wav2pcm.deletef(pcm_file) #将此次识别的pcm文件删除
else:
res_str = "识别失败"
wav2pcm.deletef(pcm_file) #将此次识别的pcm文件删除
return res_str
def main():
face_c = cv2.CascadeClassifier(
"classifier/haarcascade_frontalface_default.xml")
recs = record.rec()
subprocess.call("reproduce_results.bat", stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
f = glob.glob("Results/*.avi")[0]
vid = cv2.VideoCapture(f)
# x, y, w, h = 312, 90, 16, 16
samples = 750
avg_r = np.zeros(samples)
avg_g = np.zeros(samples)
avg_b = np.zeros(samples)
# ready = False
# print("Waiting ...")
# print("Enter q to start")
# while not ready:
# ret, img = vid.read()
# cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 1)
# cv2.imshow('vid', img)
# if cv2.waitKey(1) & 0xFF == ord('q'):
# break
print("Starting ... ")
if not vid.isOpened():
print("not opened")
input()
for i in range(samples):
ret, img = vid.read()
blur = cv2.GaussianBlur(img, (7, 7), 0, 0)
x, y, h, w = recs[i]
face_r, face_g, face_b = blur[y:(y + h), x:(x + w), 0], blur[y:(y + h), x:(x + w), 1], blur[y:(y + h), x:(x + w), 2]
avg_r[i], avg_g[i], avg_b[i] = np.sum(face_r) / (w * h), np.sum(face_g) / (w * h), np.sum(face_b) / (w * h)
cv2.rectangle(blur, (x, y), (x + w, y + h), (255, 0, 0), 1)
printProgressBar(i, samples - 1, prefix='Progress:', suffix='Complete', length=50)
cv2.imshow('vid', blur)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
print("... Done")
vid.release()
cv2.destroyAllWindows()
plt.plot(np.arange(len(avg_r)), avg_r, color="r")
plt.plot(np.arange(len(avg_g)), avg_g, color="g")
plt.plot(np.arange(len(avg_b)), avg_b, color="b")
plt.show()
fftr = np.abs(np.fft.rfft(avg_r))
fftg = np.abs(np.fft.rfft(avg_g))
fftb = np.abs(np.fft.rfft(avg_b))
rrange = np.arange(0, len(fftr)) * 30 * 30 / len(fftr)
grange = np.arange(0, len(fftg)) * 30 * 30 / len(fftg)
brange = np.arange(0, len(fftb)) * 30 * 30 / len(fftb)
minf = np.where(rrange - 40 > 0)[0][0]
maxf = np.where(rrange > 120)[0][0]
ri = npmax(fftr[minf:maxf]) + minf
gi = npmax(fftg[minf:maxf]) + minf
bi = npmax(fftb[minf:maxf]) + minf
print("\n\nRed max freq is {} bpm\n".format(rrange[ri]))
print("Green max freq is {} bpm\n".format(rrange[gi]))
print("Blue max freq is {} bpm\n".format(rrange[bi]))
print("Average max freq is {} bpm\n".format((rrange[ri] + rrange[bi] + rrange[bi]) / 3))
plt.plot(rrange, fftr)
plt.plot(grange, fftg)
plt.plot(brange, fftb)
plt.plot(np.arange(0, len(fftr)) * 30 * 30 / len(fftr), fftr[:] + fftg[:] + fftb[:])
plt.show()
import record
import wav2pcm
import sound2word
import tuling
import speakout
record.rec("1k.wav") # 实现录音，将文件存储在1k.wav
#pcm_file = wav2pcm.wav_to_pcm("1k.wav") #将wav格式的语音转化为pcm格式
words = sound2word.asr_main("1k.wav") # 读取录音文件，通过API实现语音转写
new_words = tuling.Tuling(words) #实现与图灵机器人会话
speakout.tts_main(new_words) #将机器人回复的文字转化为语音
wav2pcm.play_mp3("test.mp3") #播放机器人的语音文件
def recordcommand():
rec().record()