def __init__(self, schedulerType, mmuType, memCompact):
self.running = False
self.kernelMode = False
self.mmu = mmuType
self.pcbt = {}
self.cpu = None
self.io = None
self.pid_counter = 0
self.current_pid = None
self.idle = True
self.scheduler = schedulerType
self.longScheduler = PLP()
self.memCompact = memCompact
class Kernel:
lock = threading.RLock()
def __init__(self, schedulerType, mmuType, memCompact):
self.running = False
self.kernelMode = False
self.mmu = mmuType
self.pcbt = {}
self.cpu = None
self.io = None
self.pid_counter = 0
self.current_pid = None
self.idle = True
self.scheduler = schedulerType
self.longScheduler = PLP()
self.memCompact = memCompact
def start(self):
print('[kernel] starting...')
if not self.running:
self.running = 1
self.cpu = CPU(self, self.mmu)
self.cpu.start()
self.io = IO(self)
self.io.start()
print('[kernel] started.')
def stop(self):
self.running = False
sys.exit(0)
def interrupt(self, signal):
self.lock.acquire()
self.kernelMode = True
try:
signal.execute(self)
finally:
self.kernelMode = False
self.lock.release()
def run(self, prog):
self.lock.acquire()
try:
pid = self.pid_counter
self.pid_counter = self.pid_counter + 1
self.pcbt[pid] = PCB(pid, prog.priority, 0)
self.longScheduler.add(self, self.pcbt[pid], prog, self.memCompact)
finally:
self.lock.release()
print('[kernel.run] pid %d: %s'%(pid, prog))
self.interrupt(SignalNew())
def setQuantum(self, quantum):
self.scheduler.setQuantum(quantum)
def getCPUburstTime(self):
return self.cpu.burstTime
def canLoad(self, prog):
return self.mmu.canLoad(prog)
def load(self, pcb, prog):
self.mmu.load(pcb, prog)
def longReschedule(self):
self.longScheduler.reschedule(self)
def reschedule(self):
self.scheduler.reschedule(self)
def compact(self):
self.mmu.compact(self.pcbt)
def exit(self):
self.cpu.pcb = None
self.mmu.unload(self.pcbt[self.current_pid])
del self.pcbt[self.current_pid]
self.current_pid = None
def set_idle(self):
self.idle = True
def features(filename,varin):
fs = varin['fs']
framesize = varin['framesize']
hopsize = varin['hopsize']
feature_select = varin['feature_select']
audio = ess.MonoLoader(downmix = 'left', filename = filename, sampleRate = fs)()
# spectrogram init
winAnalysis = 'hann'
N = 2 * framesize # padding 1 time framesize
SPECTRUM = ess.Spectrum(size=N)
WINDOW = ess.Windowing(type=winAnalysis, zeroPadding=N-framesize)
highFrequencyBound = fs/2 if fs/2<11000 else 11000
MFCC = ess.MFCC(sampleRate=fs,highFrequencyBound=highFrequencyBound,inputSize=framesize+1)
GFCC = ess.GFCC(sampleRate=fs,highFrequencyBound=highFrequencyBound)
ZCR = ess.ZeroCrossingRate()
ENERGY = ess.Energy()
mfcc = []
mfccBands = []
gfcc = []
energy = []
zcr = []
autoCorrelation = []
mX = []
print 'calculating ', feature_select, ' ... ...'
for frame in ess.FrameGenerator(audio, frameSize=framesize, hopSize=hopsize):
frame_audio = frame
frame = WINDOW(frame)
mXFrame = SPECTRUM(frame)
mX.append(mXFrame)
energyFrame = ENERGY(mXFrame)
energy.append(energyFrame)
if feature_select == 'mfcc' or feature_select == 'dmfcc':
bands,mfccFrame = MFCC(mXFrame)
mfccFrame = mfccFrame[1:]
mfcc.append(mfccFrame)
if feature_select == 'mfccBands':
bands,mfccFrame = MFCC(mXFrame)
mfccBands.append(bands)
if feature_select == 'gfcc':
bands,gfccFrame = GFCC(mXFrame)
gfccFrame = gfccFrame[1:]
gfcc.append(gfccFrame)
if feature_select == 'zcr':
zcrFrame = ZCR(frame_audio)
zcr.append(zcrFrame)
if feature_select == 'autoCorrelation':
autoCorrelationFrame = np.corrcoef(frame_audio)
autoCorrelation.append(autoCorrelationFrame)
mX = np.array(mX)
if feature_select == 'mfcc':
feature = np.array(mfcc)
elif feature_select == 'dmfcc':
dmfcc = Fdeltas(np.array(mfcc).transpose(),w=9)
ddmfcc = Fdeltas(dmfcc,w=5)
feature = np.transpose(np.vstack((dmfcc,ddmfcc)))
elif feature_select == 'mfccBands':
feature = np.array(mfccBands)
elif feature_select == 'gfcc':
feature = np.array(gfcc)
elif feature_select == 'plpcc':
feature,plp,bark = PLP(mX,modelorder=12,rasta=False)
feature = feature[:,1:]
elif feature_select == 'plp':
plpcc,feature,bark = PLP(mX,modelorder=12,rasta=False)
feature = feature[:,1:]
elif feature_select == 'rasta-plpcc':
feature,plp,bark = PLP(mX,modelorder=12,rasta=True)
feature = feature[:,1:]
elif feature_select == 'rasta-plp':
plpcc,feature,bark = PLP(mX,modelorder=12,rasta=True)
feature = feature[:,1:]
elif feature_select == 'bark':
plpcc,plp,feature = PLP(mX,modelorder=12,rasta=False)
elif feature_select == 'zcr':
feature = np.array(zcr)
elif feature_select == 'autoCorrelation':
feature = np.array(autoCorrelation)
else:
feature,d_MRCG,dd_MRCG = MRCG(audio,fs=fs)
return feature,energy,mX