def run_main_experiment():
time_start = core.getTime()
time_play = time_start
order = Exp.make_random_stim_order()
Nonethird = np.floor(len(order)/3)
Ntwothird = np.floor(2*len(order)/3)
Exp = exp_param.exp_param()
rate = Exp.rate
a = 1000 # amplitude of sounds to play (scaling sounds of rms 1)
# plotting and playing stimuli
R_inf = 10000
A_broad = sb.make_act_target(Exp.RATIO[4],Exp.f0,Exp.BROAD,Exp)
A_dark = sb.make_act_target(Exp.RATIO[4],Exp.f0,Exp.DARK,Exp)
B_broad = sb.make_act_target(R_inf,Exp.f0,Exp.BROAD,Exp)
B_dark = sb.make_act_target(R_inf,Exp.f0,Exp.DARK,Exp)
couples = [ [A_broad,B_broad],[A_dark,B_dark]]
t = 0
for i in np.arange(0,100):
t = t+1
snr_type = random.randint(0,1)
timbre_type = random.randint(0,1)
AXB_type = random.randint(0,1)
WXZ_type = random.randint(0,1)
pair = couples[snr_type]
A = pair[int(WXZ_type)]
B = pair[int(not WXZ_type)]
X = [A,B][AXB_type]
blank = np.zeros((Exp.rate,1),float)
seq = np.append(blank,A)
seq = np.append(seq,blank)
seq = np.append(seq,X)
seq = np.append(seq,blank)
seq = np.append(seq,B)
seq = np.append(seq,blank)
snrdb = [0,-6][snr_type]
s = sb.add_lpnoise_at_snrdb(seq,snrdb,Exp)
s = sound_build.make_noisy_stim(i,Exp)
scaled = np.int16(s/np.max(np.abs(s)) * 32767)
write('test.wav', 44100, scaled)
core.wait(time_play - core.getTime())
set_msg('AAB or ABB?','MAIN')
win.flip()
playsound(s,vol)
core.wait(0.1)
#core.wait(0.5) #wait 500ms; but use a loop of x frames for more accurate timing in fullscreen
thisResp = get_response()
iscorrect = (get_response()==AXB_type)
time_play = core.getTime() + iti
dataFile.write('%i,%i,%i,%i,%i,%i\n' %(snr_type,timbre_type,AXB_type,WXZ_type, thisResp,iscorrect))
dataFile.close()
def perform_analysis(filename):
#path = '../data/'
#filename = 'gui_short2014_Jan_03_1512'
#filename = 'remi_short2014_Jan_03_1512'
#filename = 'VINCENT5JAN'
#with open(path+filename+'.csv', 'r') as f:
with open(filename, 'r') as f:
reader = csv.reader(f)
reader.next()
x = list(reader)
y = [[int(string) for string in inner] for inner in x]
INDEX_hct = [15,16,17,18]
CORRECT_hct = [0,0,1,1]
counts_hct = np.zeros((len(INDEX_hct),3),float) # correct, incorrect, missed
INDEX_control = [19,20,21,22]
CORRECT_control = [0,0,1,1]
counts_control = np.zeros((len(INDEX_control),3),float) # correct, incorrect, missed
# Extract counts for HCT
i_index = -1
for index in INDEX_hct:
i_index += 1
for row in y:
if (row[0] == index):
if (row[2] == 1): #correct
counts_hct[i_index,0] += 1
elif (row[2] == 0): #incorrect
counts_hct[i_index,1] += 1
elif (row[2] == -1):#missed
counts_hct[i_index,2] += 1
# Compute Accuracy for HCT
accuracy_hct = np.zeros((4),float)
i_index = -1
for row in counts_hct:
i_index += 1
a = row[1]/(row[0]+row[1])
x = CORRECT_hct[i_index]
accuracy_hct[i_index] = a**x*(1-a)**(1-x)
# Extract counts for CONTROLS (counts of correct answers)
i_index = -1
for index in INDEX_control:
i_index += 1
for row in y:
if (row[0] == index):
if (row[2] == 1): #correct
counts_control[i_index,0] += 1
elif (row[2] == 0): #incorrect
counts_control[i_index,1] += 1
elif (row[2] == -1):#missed
counts_control[i_index,2] += 1
# Compute Accuracy for HCT
accuracy_control = np.zeros((4),float)
i_index = -1
for row in counts_control:
i_index += 1
a = row[1]/(row[0]+row[1])
x = CORRECT_hct[i_index]
accuracy_control[i_index] = a**x*(1-a)**(1-x)
# Make psychometric curve for mains
Exp = exp_param.exp_param()
INDEX_ambi = Exp.Ambi_exp_sounds # [0...14] for ACT
N_ambi = len(INDEX_ambi)
counts_ambi = np.zeros((N_ambi,3),float) # 3 timbre/snr, 5 ratios, 3 possible responses
# Extract counts for mains
i_index = -1
for index in INDEX_ambi:
i_index += 1
for row in y:
if (row[0] == index):
if (row[1] == 1): #pressed up
counts_ambi[i_index,0] += 1
elif (row[1] == -1): #pressed down
counts_ambi[i_index,1] += 1
else:# pressed other key (miss)#psy_curve = np.zeros((3,5)) # cond * ratio
counts_ambi[i_index,2] += 1
# Compute Accuracy for main
accuracy_ambi = np.zeros((len(INDEX_ambi)),float)
i_index = -1
for row in counts_ambi:
i_index += 1
accuracy_ambi[i_index] = row[0]/(row[0]+row[1])
# Merge accuracies to build psychometric curves
# indices of sounds of increasing ratio for the 3 conditions
ind_by_cond = [np.arange(0,N_ambi,3),
np.arange(1,N_ambi,3),
np.arange(2,N_ambi,3)]
psy_curve = [[accuracy_ambi[index] for index in row] for row in ind_by_cond]
psy_curve_std = [[accuracy_ambi[index]*(1-accuracy_ambi[index]) for index in row] for row in ind_by_cond]
# compute std
std_control = ((1-accuracy_control)*accuracy_control)
std_hct = ((1- accuracy_hct)*accuracy_hct)
std_ambi= ((1- accuracy_ambi)*accuracy_ambi)
#########################################
N = 2
ind = np.arange(N) # the x locations for the groups
width = 0.35 # the width of the bars
hctf250m = accuracy_hct[[0,2]]
hctf250std = std_hct[[0,2]]
hctf500m = accuracy_hct[[1,3]]
hctf500std = std_hct[[1,3]]
controlf250m = accuracy_control[[0,2]]
controlf250std = std_control[[0,2]]
controlf500m = accuracy_control[[1,3]]
controlf500std = std_control[[1,3]]
print(counts_ambi)
print(counts_hct)
print(counts_control)
print(accuracy_ambi)
print(accuracy_hct)
print(accuracy_control)
print(hctf250m)
print(hctf500m)
print(controlf250m)
print(controlf500m)
RATIO = Exp.RATIO
def press(event):
print('press', event.key)
sys.stdout.flush()
if event.key=='enter':
sys.exit()
#########################################
fig1, ax1 = plt.subplots()
rects1 = ax1.bar(ind, hctf500m, width, color='r', yerr=hctf500std)
rects2 = ax1.bar(ind+width, hctf250m, width, color='y', yerr=hctf250std)
ax1.set_ylabel('fraction high')
ax1.set_xticks(ind+width)
ax1.set_xticklabels( ('500 Hz', '250 Hz') )
ax1.legend( (rects1[0], rects2[0]), ('Broad', 'Dark') )
ax1.set_title('HCT')
plt.ylim(0,1)
#########################################
fig2, ax2 = plt.subplots()
rects1 = ax2.bar(ind, controlf500m, width, color='r', yerr=controlf500std #accuracy_control[i_index] = row[0]/(row[0]+row[1])
)
rects2 = ax2.bar(ind+width, controlf250m, width, color='y', yerr=controlf250std)
ax2.set_ylabel('fraction high')
ax2.set_xticks(ind+width)
ax2.set_xticklabels( ('500 Hz', '250 Hz') )
ax2.legend( (rects1[0], rects2[0]), ('Broad', 'Dark') )
ax2.set_title('Controls')
plt.ylim(0,1)
#########################################
fig3, ax3 = plt.subplots()
for i in np.arange(0,np.shape(psy_curve)[0]):
ax3.errorbar(RATIO, psy_curve[i], yerr=psy_curve_std[i] , linestyle="dashed", marker="o")
ax3.legend(['0db dark','0db', '-6dB'])
ax3.set_xlabel('ratio')
ax3.set_ylabel('fraction high')
ax3.grid(True)
ax3.set_title('Timbral effects, main')
plt.ioff()
plt.ylim(0,1)
fig1.show()
fig1.canvas.mpl_connect('key_press_event', press)
fig2.show()
fig2.canvas.mpl_connect('key_press_event', press)
fig3.show()
fig3.canvas.mpl_connect('key_press_event', press)
plt.show()
from psychopy.misc import fromFile, toFile
from scipy.io.wavfile import write
import numpy as np
import pygame
import random
import os, sys
# realpath() with make your script run, even if you symlink it :)
#cmd_subfolder = os.path.realpath(os.path.abspath(os.path.split(inspect.getfile( inspect.currentframe() ))[0]) + "/scripts")
cmd_subfolder = os.path.realpath('..')+'/scripts'
if cmd_subfolder not in sys.path:
sys.path.insert(0, cmd_subfolder)
import sound_build, exp_param
#------------ load global experiment parameters
Exp = exp_param.exp_param()
iti = Exp.iti
# constants initialization
TOP_POS = [0,+3]
CENTER_POS = [0,0]
BOTTOM_POS = [0,-4]
# function definition
def get_subject_info():
try:
expInfo = fromFile('../data/lastParams.pickle') # check existence of previous parameter file
except:
expInfo = {'pseudo':'pseudo'}
expInfo['date']= data.getDateStr() #add current time
#present a dialog box to get user info