def __init__(self, root):
Frame.__init__(self, root)
#========================================================================
# Fast Fourier Transform
self.processor = FFTVoiceAnalyzer(path_to_examples, np.fft.fft, WavFile(path_to_silence))
# Fast Fourier Transform
#========================================================================
#========================================================================
# Naive Bayes Classifier
self.nbc = NBC()
self.nbc.initialize()
# Naive Bayes Classifier
#========================================================================
#========================================================================
# SPro 5 (MFCC classifierd)
self.s = SPro5()
self.s.learn()
# SPro 5 (MFCC classifierd)
#========================================================================
#get general properties
self.title = cf.get("general", "title")
self.author = cf.get("general", "author")
self.link_color = cf.get("general", "link_color")
self.link_cursor = cf.get("general", "link_cursor")
self.font_family = cf.get("general", "font_family")
self.author_email = cf.get("general", "author_email")
self.copyright_year = cf.get("general", "copyright_year")
self.author_google_plus = cf.get("general", "author_google_plus")
self.license_agree_link = cf.get("general", "license_agree_link")
#get general properties
#get program properties
self.min_audio_time = int(cf.get("program", "min_audio_time"))
self.max_audio_time = int(cf.get("program", "max_audio_time"))
self.default_audio_time = int(cf.get("program", "default_audio_time"))
self.min_test_audio_time = int(cf.get("program", "min_test_audio_time"))
self.max_test_audio_time = int(cf.get("program", "max_test_audio_time"))
self.default_test_audio_time = int(cf.get("program", "default_test_audio_time"))
#get program properties
self.root = root
self.root.title(self.title)
self.root.resizable(FALSE, FALSE)
self.root.protocol("WM_DELETE_WINDOW", self.close)
root.wm_iconbitmap(bitmap=cf.get("main_window", "icon_file"))
self.top_frame = None
self.main_frame = None
self.bottom_frame = None
self.init_ui()
self.pack()
class Application(Frame):
def __init__(self, root):
Frame.__init__(self, root)
#========================================================================
# Fast Fourier Transform
self.processor = FFTVoiceAnalyzer(path_to_examples, np.fft.fft, WavFile(path_to_silence))
# Fast Fourier Transform
#========================================================================
#========================================================================
# Naive Bayes Classifier
self.nbc = NBC()
self.nbc.initialize()
# Naive Bayes Classifier
#========================================================================
#========================================================================
# SPro 5 (MFCC classifierd)
self.s = SPro5()
self.s.learn()
# SPro 5 (MFCC classifierd)
#========================================================================
#get general properties
self.title = cf.get("general", "title")
self.author = cf.get("general", "author")
self.link_color = cf.get("general", "link_color")
self.link_cursor = cf.get("general", "link_cursor")
self.font_family = cf.get("general", "font_family")
self.author_email = cf.get("general", "author_email")
self.copyright_year = cf.get("general", "copyright_year")
self.author_google_plus = cf.get("general", "author_google_plus")
self.license_agree_link = cf.get("general", "license_agree_link")
#get general properties
#get program properties
self.min_audio_time = int(cf.get("program", "min_audio_time"))
self.max_audio_time = int(cf.get("program", "max_audio_time"))
self.default_audio_time = int(cf.get("program", "default_audio_time"))
self.min_test_audio_time = int(cf.get("program", "min_test_audio_time"))
self.max_test_audio_time = int(cf.get("program", "max_test_audio_time"))
self.default_test_audio_time = int(cf.get("program", "default_test_audio_time"))
#get program properties
self.root = root
self.root.title(self.title)
self.root.resizable(FALSE, FALSE)
self.root.protocol("WM_DELETE_WINDOW", self.close)
root.wm_iconbitmap(bitmap=cf.get("main_window", "icon_file"))
self.top_frame = None
self.main_frame = None
self.bottom_frame = None
self.init_ui()
self.pack()
def init_ui(self):
"""
initialize all GUI
"""
self.style = Style()
self.style.theme_use('default')
self.init_menu()
self.init_top_frame()
self.main_frame = Frame(self, relief=SUNKEN)
self.make_main_frame()
self.init_bottom_frame()
def init_menu(self):
"""
initialize menu (now menu has only 'About' element)
"""
menu_bar = Menu(self.root)
help_menu = Menu(menu_bar, tearoff=0)
help_menu.add_command(label="About", command=self.about)
menu_bar.add_cascade(label="Help", menu=help_menu)
self.root.config(menu=menu_bar)
def about(self):
"""
makes simple GUI of About window
"""
window = "about_window"
ok_hover_bg = cf.get(window, "ok_hover_bg")
ok_default_bg = cf.get(window, "ok_default_bg")
font = Font(family=self.font_family, size=10)
#main frame
top = Toplevel(self)
top.resizable(FALSE, FALSE)
top.title("About " + self.title)
top.wm_iconbitmap(bitmap=cf.get(window, "icon_file"))
#info frame in main frame
frame = Frame(top, relief=SUNKEN, bg="white", borderwidth=1)
Label(frame, text="Voice Analyzer. Simple program for speech recognition, created for qualification work",
bg="white", font=font).pack(padx=5, pady=5)
Label(frame, text="Copyright " + self.copyright_year + ". GNU License. All rights reserved.",
bg="white", font=font).pack(padx=5, pady=5)
#author
link_info = Message(frame, text="Author: " + self.author, bg="white", font=font, width=250)
link_info.pack(side=LEFT, anchor=E, padx=5, pady=5)
#link to author page (somewhere)
link = Label(frame, text="Google Plus", bg="white", fg=self.link_color, font=font, cursor=self.link_cursor)
link.bind("<Button-1>", lambda x: webbrowser.open_new_tab(self.author_google_plus))
link.pack(side=LEFT, anchor=S, pady=5)
frame.pack()
#OK button in main frame
button = Button(top, text="OK", command=top.destroy, width=10, bg=ok_default_bg, borderwidth=1)
button.bind("<Enter>", lambda event, h=button: self.change_bg(h, ok_hover_bg))
button.bind("<Leave>", lambda event, h=button: self.change_bg(h, ok_default_bg))
button.pack(side=RIGHT, anchor=S, padx=5, pady=5)
@staticmethod
def change_bg(item, new_bg):
item.configure(bg=new_bg)
item.pack()
@staticmethod
def clear_frame(frame):
for child in frame.winfo_children():
child.destroy()
#==================================================================================================================
#========================================================================
# Record Audio
def record_audio(self, time, path=path_to_records):
"""
record audio data to file
@param time: length of file in sec
"""
file_name = filedialog.asksaveasfilename(filetypes=[("Wave audio files", "*.wav *.wave")],
defaultextension=".wav", initialdir=path)
if len(file_name) == 0:
messagebox.showwarning("Warning", "You must input name of new file, and save it!")
return None
else:
wav = self.processor.recorder.record_audio_to_file_and_get_wav(time=time, file_name=file_name)
messagebox.showinfo("File Saved", "Audio has recorded.")
return wav
def record_test_audio(self, time):
"""
records audio from microphone and uses this audio for analyzing
@param time: length of test file
"""
if time.get() == 0:
time.set(self.default_test_audio_time)
return self.processor.recorder.record_and_get_wav(time.get())
# Record Audio
#========================================================================
#========================================================================
# Fast Fourier Transform Voice Analyzer
def fft_analyzer_record_to_lib(self, time):
"""
records short audio file (file example with some word)
and adds it to library
@param time: length of test file
"""
if time.get() == 0:
time.set(self.default_audio_time)
wav = self.record_audio(time.get())
self.processor.lib.create_and_add_item_from_wave(wav)
def fft_analyzer_record(self, time):
if time.get() == 0:
time.set(self.default_audio_time)
wav = self.record_test_audio(time)
if not wav is None:
result_str = FFTVoiceAnalyzer.analyze(wav, self.processor)
messagebox.showinfo("Result", result_str)
if show_plots:
plot = Plotter("DRA")
plot.add_sub_plot_data("Digitized Recorded Audio", wav.get_one_channel_data(), x_label="Samples",
y_label="Amplitude")
plot.sub_plot_all_horizontal(show=False, save=True)
def fft_analyzer_select(self, path=path_to_test):
askopenfile = filedialog.askopenfile(filetypes=[("Wave audio files", "*.wav *.wave")], defaultextension=".wav",
initialdir=path)
if not askopenfile is None:
wav = WavFile(askopenfile.name)
result_str = FFTVoiceAnalyzer.analyze(wav, self.processor)
messagebox.showinfo("Result", result_str)
if show_plots:
plot = Plotter("DRA")
plot.add_sub_plot_data("Digitized Recorded Audio", wav.get_one_channel_data(), x_label="Samples",
y_label="Amplitude")
plot.sub_plot_all_horizontal(show=False, save=True)
else:
messagebox.showwarning("Warning", "You should select one file. Please, try again")
# Fast Fourier Transform Voice Analyzer
#========================================================================
#========================================================================
# Naive Bayes Classifier
def nbc_add_file_select(self, clazz, path=path_to_records):
askopenfile = filedialog.askopenfile(filetypes=[("Wave audio files", "*.wav *.wave")], defaultextension=".wav",
initialdir=path)
if not askopenfile is None:
self.nbc.add_one_audio_file(clazz, path_to_file=askopenfile.name)
self.nbc.teach_classifier()
def nbc_classify_file_select(self, path=path_to_test):
askopenfile = filedialog.askopenfile(filetypes=[("Wave audio files", "*.wav *.wave")], defaultextension=".wav",
initialdir=path)
if not askopenfile is None:
print("Teaching NBC classifier with added examples")
self.nbc.teach_classifier()
print("Teaching NBC classifier finished")
classes = self.nbc.get_classes(self.nbc.classify(WavFile(askopenfile.name)))
mess = ""
for i in classes.keys():
for j in classes[i].keys():
mess += str(i) + ": " + str(j) + "\n"
messagebox.showinfo("Classification results", mess)
else:
messagebox.showwarning("Warning", "You should select one file. Please, try again")
# Naive Bayes Classifier
#========================================================================
#========================================================================
# Voice Activity Detection
def show_test_vad_open(self, path=path_to_test):
askopenfile = filedialog.askopenfile(filetypes=[("Wave audio files", "*.wav *.wave")], defaultextension=".wav",
initialdir=path)
if not askopenfile is None:
test(WavFile(askopenfile.name), self.nbc)
else:
messagebox.showwarning("Warning", "You should select one file. Please, try again")
def show_test_vad_record(self, time):
wav = self.record_test_audio(time)
if not wav is None:
test(wav, self.nbc)
if show_plots:
plot = Plotter("DRA")
plot.add_sub_plot_data("Digitized Recorded Audio", wav.get_one_channel_data(), x_label="Samples",
y_label="Amplitude")
plot.sub_plot_all_horizontal(show=False, save=True)
# Voice Activity Detection
#========================================================================
#========================================================================
# MFCC
def add_mfcc_file(self, file_type, path=path_to_mfcc):
if not self.record_audio(5, path=path + "waves/" + file_type) is None:
self.s.learn()
def show_test_mfcc(self):
print("Start MFCC")
self.s.test()
str_res = SPro5.get_results()
messagebox.showinfo("Results MFCC", str_res)
print("MFCC finished successful")
# MFCC
#========================================================================
#==================================================================================================================
def make_main_frame(self):
"""
makes simple GUI of main window with 2 buttons record example file (Teach Program)
and record test audio and analyze it (Record and Analyze Sound)
"""
self.clear_frame(self.main_frame)
#========================================================================
# Record Audio
Label(self.main_frame, text="Record Audio File",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, pady=5, padx=5)
Button(self.main_frame, text="Record",
command=lambda: self.make_record_frame(self.min_audio_time, self.max_audio_time,
self.record_audio), width=15).pack()
# Record Audio
#========================================================================
#========================================================================
# Fast Fourier Transform Voice Analyzer
Label(self.main_frame, text="FFT Voice Analyzer",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, pady=5, padx=5)
Button(self.main_frame, text="Record and Teach Program",
command=lambda: self.make_record_frame(self.min_audio_time, self.max_audio_time,
self.fft_analyzer_record_to_lib), width=30).pack()
Button(self.main_frame, text="Record and Analyze Sound",
command=lambda: self.make_record_frame(self.min_test_audio_time, self.max_test_audio_time,
self.fft_analyzer_record),
width=30).pack()
Button(self.main_frame, text="Choose and Analyze Sound",
command=lambda: self.fft_analyzer_select(), width=30).pack()
# Fast Fourier Transform Voice Analyzer
#========================================================================
#========================================================================
# Naive Bayes Classifier
Label(self.main_frame, text="Naive Bayes Classifier",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, pady=5, padx=5)
Button(self.main_frame, text="Add Speech File for NBC", command=lambda: self.nbc_add_file_select("speech"),
width=30).pack()
Button(self.main_frame, text="Add Non Speech File for NBC",
command=lambda: self.nbc_add_file_select("non_speech"),
width=30).pack()
Button(self.main_frame, text="Choose and Classify File", command=lambda: self.nbc_classify_file_select(),
width=30).pack()
# Naive Bayes Classifier
#========================================================================
#========================================================================
# Voice Activity Detection
Label(self.main_frame, text="Voice Activity Detection",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, pady=5, padx=5)
Button(self.main_frame, text="Record and Analyze",
command=lambda: self.make_record_frame(self.min_audio_time, self.max_audio_time,
self.show_test_vad_record), width=30).pack()
Button(self.main_frame, text="Choose and Analyze", command=lambda: self.show_test_vad_open(), width=30).pack()
# Voice Activity Detection
#========================================================================
#========================================================================
# SPro 5 (MFCC classifier)
Label(self.main_frame, text="MFCC",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, pady=5, padx=5)
Button(self.main_frame, text="Test", command=lambda: self.show_test_mfcc(), width=30).pack()
Button(self.main_frame, text="Record Test File", command=lambda: self.add_mfcc_file("test"), width=30).pack()
Button(self.main_frame, text="Record Learn File", command=lambda: self.add_mfcc_file("learn"), width=30).pack()
# SPro 5 (MFCC classifier)
#========================================================================
self.main_frame.pack(side=TOP)
def init_top_frame(self):
"""
initialize head of main window
"""
self.top_frame = Frame(self, relief=RIDGE, height=50, bg="white", borderwidth=1)
Label(self.top_frame, text="Welcome to " + self.title, bg="white",
font=Font(family=self.font_family, size=10, weight="bold")).pack(side=TOP, anchor=W, pady=5, padx=5)
Label(self.top_frame, text="You can teach program or use default data", bg="white",
font=Font(family=self.font_family, size=8)).pack(side=TOP, anchor=W, padx=10)
self.top_frame.pack(side=TOP, fill=BOTH)
def init_bottom_frame(self):
"""
initialize footer of main window
"""
self.bottom_frame = Frame(self, relief=GROOVE, height=50, borderwidth=1)
Button(self.bottom_frame, text="Cancel", command=self.close, width=10).pack(side=RIGHT, padx=5, pady=5)
font = Font(family=self.font_family, size=8)
link = Label(self.bottom_frame, text="License Agreement:", fg=self.link_color, font=font,
cursor=self.link_cursor)
link.bind("<Button-1>", lambda x: webbrowser.open_new_tab(self.license_agree_link))
link.pack(side=LEFT, pady=5)
text = Label(self.bottom_frame, text="By continuing, you agree to the term of the license agreement.",
font=font)
text.pack(side=LEFT, pady=5)
self.bottom_frame.pack(side=BOTTOM, fill=BOTH)
def close(self):
"""
close program dialog
"""
answer = messagebox.askyesno(parent=self, message='Are you sure you want to close program?',
icon='question', title='Close')
if answer:
self.root.destroy()
def make_record_frame(self, min_time, max_time, record_function):
"""
make simple GUI of record window
@param min_time: minimum length of audio which will record
@param max_time: maximum length of audio which will record
@param record_function: command for record button (now: record example audio or test audio)
"""
self.clear_frame(self.main_frame)
time = IntVar(self.main_frame)
Label(self.main_frame, text="Time:").pack(side=LEFT)
Spinbox(self.main_frame, from_=min_time, to=max_time, textvariable=time).pack(side=LEFT)
Button(self.main_frame, text="Record audio", command=lambda: record_function(time), width=20).pack(side=LEFT)
Button(self.main_frame, text="Back", command=self.make_main_frame, width=20).pack(side=LEFT)
self.main_frame.pack()
