def wav_header_unpack(data):
(
riff,
riffsize,
wave,
fmt,
fmtsize,
format,
nchannels,
samplespersecond,
datarate,
blockalign,
bitspersample,
data,
datalength,
) = struct.unpack("<4sl4s4slhhllhh4sl", data)
if riff != b"RIFF":
raise ValueError("invalid wav header")
if fmtsize != 16 or fmt != b"fmt " or data != b"data":
# fmt chuck is not first chunk, directly followed by data chuck
# It is nowhere required that they are, it is just very common
raise ValueError("cannot understand wav header")
wfx = pywintypes.WAVEFORMATEX()
wfx.wFormatTag = format
wfx.nChannels = nchannels
wfx.nSamplesPerSec = samplespersecond
wfx.nAvgBytesPerSec = datarate
wfx.nBlockAlign = blockalign
wfx.wBitsPerSample = bitspersample
return wfx, datalength
def testRecord(self):
d = ds.DirectSoundCaptureCreate(None, None)
sdesc = ds.DSCBUFFERDESC()
sdesc.dwBufferBytes = 352800 # 2 seconds
sdesc.lpwfxFormat = pywintypes.WAVEFORMATEX()
sdesc.lpwfxFormat.wFormatTag = pywintypes.WAVE_FORMAT_PCM
sdesc.lpwfxFormat.nChannels = 2
sdesc.lpwfxFormat.nSamplesPerSec = 44100
sdesc.lpwfxFormat.nAvgBytesPerSec = 176400
sdesc.lpwfxFormat.nBlockAlign = 4
sdesc.lpwfxFormat.wBitsPerSample = 16
buffer = d.CreateCaptureBuffer(sdesc)
event = win32event.CreateEvent(None, 0, 0, None)
notify = buffer.QueryInterface(ds.IID_IDirectSoundNotify)
notify.SetNotificationPositions((ds.DSBPN_OFFSETSTOP, event))
buffer.Start(0)
win32event.WaitForSingleObject(event, -1)
event.Close()
data = buffer.Update(0, 352800)
fname = os.path.join(win32api.GetTempPath(),
'test_directsound_record.wav')
f = open(fname, 'wb')
f.write(wav_header_pack(sdesc.lpwfxFormat, 352800))
f.write(data)
f.close()
def __init__(self, nchnl=1, sps=16000, bps=16, t=0.1):
dsc = directsound.DirectSoundCaptureCreate(None, None) #创建设备对象
cdesc = directsound.DSCBUFFERDESC() #创建DSCBUFFERDESC结构对象
self.bSize = int(sps * nchnl * bps / 8 * t)
cdesc.dwBufferBytes = self.bSize #缓存大小
cdesc.lpwfxFormat = pywintypes.WAVEFORMATEX() #DirectSound数据块格式
cdesc.lpwfxFormat.wFormatTag = pywintypes.WAVE_FORMAT_PCM
cdesc.lpwfxFormat.nChannels = nchnl
cdesc.lpwfxFormat.nSamplesPerSec = sps
cdesc.lpwfxFormat.nAvgBytesPerSec = int(sps * nchnl * bps / 8)
cdesc.lpwfxFormat.nBlockAlign = int(nchnl * bps / 8)
cdesc.lpwfxFormat.wBitsPerSample = bps
self.buffer = dsc.CreateCaptureBuffer(cdesc) #创建缓冲区对象
self.evt = []
for i in range(2):
self.evt.append(win32event.CreateEvent(None, 0, 0,
None)) #创建两个事件通知
Notify = self.buffer.QueryInterface(
directsound.IID_IDirectSoundNotify) #创建事件通知接口
Notify.SetNotificationPositions([
(int(self.bSize / 2) - 1, self.evt[0]),
(self.bSize - 1, self.evt[1])
]) #error
#设置两个通知位置，缓冲区每填充bSize/2个样本即发送一个通知消息
self.data = b'' #用于实时存储捕获的音频数据
self.STATUS = False #录音状态标志
self.wfx = cdesc.lpwfxFormat #存储声音格式
def play(self): # 播放文件
f = open(self.file, 'rb') # 打开文件
header = f.read(WAV_HEADER_SIZE) # 读取WAV文件头
(riff, riffsize, wave, fmt, fmtsize,
format, nchannels, samplespersecond,
datarate, blockalign, bitspersample,
data, size) =\
struct.unpack('<4sl4s4slhhllhh4sl', header) # 获取参数值
if riff != 'RIFF' or fmtsize != 16 or fmt != 'fmt ' or data != 'data': # 判断文件格式
raise 'Data Error'
wfx = pywintypes.WAVEFORMATEX() # 创建WAVEFORMATEX结构
wfx.wFormatTag = format
wfx.nChannels = nchannels
wfx.nSamplesPerSec = samplespersecond
wfx.nAvgBytesPerSec = datarate
wfx.nBlockAlign = blockalign
wfx.wBitsPerSample = bitspersample
d = directsound.DirectSoundCreate(None, None) # 使用DirectSound播放声音
d.SetCooperativeLevel(None, directsound.DSSCL_PRIORITY)
sdesc = directsound.DSBUFFERDESC()
sdesc.dwFlags = (directsound.DSBCAPS_STICKYFOCUS
| directsound.DSBCAPS_CTRLPOSITIONNOTIFY)
sdesc.dwBufferBytes = size
sdesc.lpwfxFormat = wfx
self.buffer = buffer = d.CreateSoundBuffer(sdesc, None)
buffer.Update(0, f.read(size))
buffer.Play(0)
def test_2_Attr(self):
'WAVEFORMATEX attribute access'
# A wav header for a soundfile from a CD should look like this...
w = pywintypes.WAVEFORMATEX()
w.wFormatTag = pywintypes.WAVE_FORMAT_PCM
w.nChannels = 2
w.nSamplesPerSec = 44100
w.nAvgBytesPerSec = 176400
w.nBlockAlign = 4
w.wBitsPerSample = 16
self.failUnless(w.wFormatTag == 1)
self.failUnless(w.nChannels == 2)
self.failUnless(w.nSamplesPerSec == 44100)
self.failUnless(w.nAvgBytesPerSec == 176400)
self.failUnless(w.nBlockAlign == 4)
self.failUnless(w.wBitsPerSample == 16)
def test_2_Attr(self):
'DSCBUFFERDESC attribute access'
c = ds.DSCBUFFERDESC()
c.dwFlags = 1
c.dwBufferBytes = 2
c.lpwfxFormat = pywintypes.WAVEFORMATEX()
c.lpwfxFormat.wFormatTag = pywintypes.WAVE_FORMAT_PCM
c.lpwfxFormat.nChannels = 2
c.lpwfxFormat.nSamplesPerSec = 44100
c.lpwfxFormat.nAvgBytesPerSec = 176400
c.lpwfxFormat.nBlockAlign = 4
c.lpwfxFormat.wBitsPerSample = 16
self.failUnless(c.dwFlags == 1)
self.failUnless(c.dwBufferBytes == 2)
self.failUnless(c.lpwfxFormat.wFormatTag == 1)
self.failUnless(c.lpwfxFormat.nChannels == 2)
self.failUnless(c.lpwfxFormat.nSamplesPerSec == 44100)
self.failUnless(c.lpwfxFormat.nAvgBytesPerSec == 176400)
self.failUnless(c.lpwfxFormat.nBlockAlign == 4)
self.failUnless(c.lpwfxFormat.wBitsPerSample == 16)
def wav_header_unpack(self, data):
print("Start reading wave file.")
#unpack wav header information
(riff, riffsize, wave, fmt, fmtsize, format, nchannels, samplespersecond, \
datarate, blockalign, bitspersample, data, datalength) = struct.unpack('<4si4s4sihhiihh4si', data)
if riff != b'RIFF' or fmtsize != 16 or fmt != b'fmt ' or data != b'data':
print("Error! Please check your file!")
exit(1)
print("File size: " + str(riffsize))
wfx = pywintypes.WAVEFORMATEX()
wfx.wFormatTag = format
print("Audio format code: " + str(format))
wfx.nChannels = nchannels
print("Number of channels: " + str(nchannels))
wfx.nSamplesPerSec = samplespersecond
print("Sample rate: " + str(samplespersecond))
wfx.nAvgBytesPerSec = datarate
print("Byte rate: " + str(datarate))
wfx.nBlockAlign = blockalign
print("Block align: " + str(blockalign))
wfx.wBitsPerSample = bitspersample
print("Bits per sample: " + str(bitspersample))
return wfx, datalength
def loadwav(wavefname):
d = ds.DirectSoundCreate(None, None)
d.SetCooperativeLevel(None, ds.DSSCL_PRIORITY)
w = wave.open(wavefname, "rb")
sdesc = ds.DSBUFFERDESC()
sdesc.lpwfxFormat = pywintypes.WAVEFORMATEX()
sdesc.lpwfxFormat.wFormatTag = pywintypes.WAVE_FORMAT_PCM
sdesc.lpwfxFormat.nChannels = w.getnchannels()
sdesc.lpwfxFormat.nSamplesPerSec = w.getframerate()
sdesc.lpwfxFormat.nAvgBytesPerSec = w.getnchannels() * w.getsampwidth(
) * w.getframerate()
sdesc.lpwfxFormat.nBlockAlign = w.getnchannels() * w.getsampwidth()
sdesc.lpwfxFormat.wBitsPerSample = w.getsampwidth() * 8
sdesc.dwBufferBytes = w.getnframes() * w.getnchannels() * w.getsampwidth()
sdesc.dwFlags = ds.DSBCAPS_STICKYFOCUS | ds.DSBCAPS_CTRLPOSITIONNOTIFY
buffer = d.CreateSoundBuffer(sdesc, None)
buffer.Update(0, w.readframes(w.getnframes()))
w.close()
return buffer
def wav_header_pack(wfx, datasize):
return struct.pack('<4sl4s4slhhllhh4sl', 'RIFF', 36 + datasize, 'WAVE',
'fmt ', 16, wfx.wFormatTag, wfx.nChannels,
wfx.nSamplesPerSec, wfx.nAvgBytesPerSec,
wfx.nBlockAlign, wfx.wBitsPerSample, 'data', datasize)
d = ds.DirectSoundCaptureCreate(None, None)
sdesc = ds.DSCBUFFERDESC()
sdesc.dwBufferBytes = 352800 # 2 seconds
sdesc.lpwfxFormat = pywintypes.WAVEFORMATEX()
sdesc.lpwfxFormat.wFormatTag = pywintypes.WAVE_FORMAT_PCM
sdesc.lpwfxFormat.nChannels = 2
sdesc.lpwfxFormat.nSamplesPerSec = 44100
sdesc.lpwfxFormat.nAvgBytesPerSec = 176400
sdesc.lpwfxFormat.nBlockAlign = 4
sdesc.lpwfxFormat.wBitsPerSample = 16
buffer = d.CreateCaptureBuffer(sdesc)
def test_1_Type(self):
'WAVEFORMATEX type'
w = pywintypes.WAVEFORMATEX()
self.failUnless(type(w) == pywintypes.WAVEFORMATEXType)
def test_1_Type(self):
"WAVEFORMATEX type"
w = pywintypes.WAVEFORMATEX()
self.assertTrue(type(w) == pywintypes.WAVEFORMATEXType)
def test_1_Type(self):
'WAVEFORMATEX type'
w = pywintypes.WAVEFORMATEX()
self.assertTrue(isinstance(w, pywintypes.WAVEFORMATEXType))