Microphone Calibration
In experiments studying human speech, the captured data is often microphone data of the participant’s voice as they perform different behavioral tasks. As part of these experiments, participants will wear headphones that at times provide a stimulus the participant must react to, at others provide a mix of the participant’s voice in real-time, or even a combination of the two. A crucial element of the hardware setup is the calibration of the microphone and headphones so that the levels are controlled for, the experimentation can be replicated, and the validity of the scientific methods can be measured.
In these experiments, it is often important to control for different acoustic properties of both the produced audio recorded by the microphone and the auditory feedback supplied by the headphones. Different behavioral conditions can be met if the participant hits a threshold of loudness, pitch, or pronunciation. Similarly, participants might adjust their production based on the types of feedback provided. Validation of the hardware levels and operating patterns can allow for more precise experimental conditions and hone on in specific behavioral patterns.
Outside of the in-experiment conditions, sample recordings of the ambient environment can validate the scientific process that was included. Since sensation is a very subjective process, it is necessary to confirm that the microphone and headphones are indeed acting a manner as described by the protocol. Taking samples of “known-good” audio samples and measuring how their digital signals can be adjusted with the hardware being used are good sanity checks for post-analysis and peer-review.
Play a standardized, “known-good”, recorded audio tone from a handheld audio recorder. The specifics of the tone are not extremely important, just as long as the waveform is uniform, and the same tone is used for each calibration, and at the same loudness. Place a sound-level meter (SLM) at a distance and angle to the audio recorder that is identical to the position that the microphone will be in relation to the participant’s mouth. At this distance and angle, record the sound pressure level (SPL) of the recorded tone.
- Handheld audio recorder with recording of a pure tone (Output: Audio in air)
- Sound-level meter (SLM) (Input: Audio in air; Output: SPL of audio)
- Value of the average SPL (decibels, dB) recorded by the SLM from the recorder
Remove the SLM from the testing area and replace it with the microphone to be used during testing. Set up the microphone at a distance and angle to the recorder that is identical to how the microphone will be place in relation to the participant’s mouth. Set up the necessary microphone amplifier and sound card, and make sure that the experimental computer is set up to record audio. Play the audio tone from the recorder at the same loudness as before and record a sample (~5s) of the audio using the microphone and the experimental computer. Adjust the gain of the microphone to account for things like low or high amplitude recordings and verify lack of artifacts caused by hardware or software. If adjustments are made, re-record a sample of the tone using the microphone. When a sample of the tone has been recorded, create a calibration ratio of the average peak-to-peak amplitude with the SLP measured in the previous step. This will be used to determine loudness of the participant during the experimental task, and their subsequent data
- Handheld audio recorder with recording of a pure tone (Output: Audio in Air)
- Experimental microphone (Mic) (Input: Audio in air)
- Microphone amplifier/gain stages (Input: raw audio; Output: Amplified audio)
- Audio card (Input: Amplified audio, Output: Processed audio)
- Computer (Input: Processed audio; Output: digital audio recording, calibration ratio)
- Value of the average SPL (decibels, dB) recorded by the SLM from the recorder
- Sampled audio tone (~5s) recorded by microphone
- Calibration ratio of audio waveform with average SPL
Identify the types of headphones to be used in the experiment and attach a coupler to the SLM which mimics the acoustics space between the headphones and the ear canal. Attached the headphones to the SLM coupler and play the sampled audio tone collected in the last step. Adjust the headphone gain stages until the SLM reads a SPL value for the target experiment. Sometimes the SPL value is the same as the microphone recording, sometimes greater or lesser.
- Sound-level meter (SLM) (Input: Audio in coupler space; Output: SPL of audio)
- Headphone coupler
- Experimental headphones (Input: digital audio recording; Output: audio in coupler space)
- Headphones gain stages (Input: digital audio recording; Output digital audio recording)
- Audio card (Input: digital audio recording; Output digital audio recording)
- Computer (Input: digital audio recording; Output digital audio recording)
- SPL of audio played through headphones.
- Sampled audio tone (~5s) recorded by the microphone
