Synchronize sound and trigger

Often, a trigger must be emitted simultenously with a sound onset, with as little delay and jitter as possible. With stimuli, similarly to psychtoolbox, the key concept is to schedule the sound.

First, let’s have a look at our default device latency.

import sounddevice as sd

from stimuli.audio import SoundAM
from stimuli.time import sleep
from stimuli.trigger import MockTrigger

idx = sd.default.device["output"]
print(f"Low-latency (s): {sd.query_devices()[idx]['default_low_output_latency']}")

Low-latency (s): 0.008684807256235827

Next, we can schedule a sound with the argument when which will use the clock provided in the argument clock to schedule the sound. Then, we wait for this duration to elapse before sending the trigger.

trigger = MockTrigger()  # replace with your trigger object
sound = SoundAM(
    frequency_carrier=1000,
    frequency_modulation=40,
    method="dsbsc",
    volume=10,
    duration=1,
)
sound.play(when=0.5, blocking=False)
sleep(0.5)
trigger.signal(1)

[mock.signal] INFO: Mock set to 1.

A quick measurement on a dual-boot Windows/Linux computer connected to a USB Crimson 3 soundcard shows that the delay and jitter of stimuli are similar to psychtoolbox on linux.

On different computers with different soundcards, the performance may vary. For instance, with on-board soundcards on Linux, both psychtoolbox and stimuli are usually perfectly synchronized with the trigger. In the end, the performance should always be measured on the target system.

from stimuli.audio import Tone
from stimuli.time import sleep
from stimuli.trigger import ParallelPortTrigger

sound = Tone(frequency=440, volume=100, duration=0.3)
trigger = ParallelPortTrigger("/dev/parport0")

for k in range(10):
    sound.play(when=0.2)
    sleep(0.2)
    trigger.signal(1)
    sleep(0.5)

Estimated memory usage: 147 MB