Pygame.mixer.sound的声音持续时间不够准确

Question

I want to generate a series of audio cues at specific frequency and duration for my neuroscience research. After some googling, I find a Python module called Pygame that seems able to get the job done.

sound = pygame.sndarray.make_sound(sample_array)   #sample_array is a one-second sine wave I created
sound.play(maxtime = 100)   #play the sample_array for 100ms
pygame.time.delay(1000)     #pause for 1000ms

I run these code in a loop for many times. However, there is a small but noticeable difference in length of the sounds. I recorded the sound with my phone, and analyzed the signal in Matlab. Some of the pulses last for as long as 180ms, some short ones last only 80ms, while they should all last 100ms in theory.

This kind of variation is too big for my application. Is there anything I can do to improve it? Or it'll also be great if someone can confirm that this kind of precision is beyond Pygame's capacity, and maybe suggest another library to use.

Answer 1

It looks like this question doesn't attract much attention :(

I'm posting my workaround here in case someone wants to know. If I use sound.play(maxtime = 100) , or sound.play(-1) then time.delay(100) , the actual duration of the sound is never precisely 100ms. My workaround is to generate a wav file beforehand which has exactly the length I need, and play this wav file.

sound = pygame.mixer.Sound("foo.wav")
sound.play()

It's not the best fix but this works for me.