Platforms /


Hardware and Calibration

Platform Who provides the hardware for this platform? Please describe any hardware that must be provided by the experimenter:
Amazon MTurk Subject None.
Cognition.Run Experimenter or subject None required.
Django Experimenter or subject A web server machine is needed to run Django.
Gorilla Experimenter or subject Hardware by experiments is optional.
hearX Experimenter A Samsung Galaxy tablet and Sennheiser headphones (both vary by price).
ispring Experimenter or subject None required.
Jacoti Experimenter or subject iPhone or iPad, Apple EarPods (likely, Apple AirPods will be an option in the very near future).
jspsych Experimenter or subject None. This is run on a browser. However, the experimenter could choose to provide hardware to standardize screen-size, headphones calibration etc., but that’s an issue that is not specific to jspsych.
MATLAB Experimenter or subject The MATLAB web app server must be provided by the experimenter to allow online access to web apps.
PART/BGC Science Experimenter or subject Experimenter can provide a calibrated device (tablet, computer, phone) running an appropriate operating system and headphones associated with the calibration settings, or it can be deployed by allowing the participant to download on their own device, use their own headphones, and use only perceptual calibration. Mixtures are possible as well.
Prolific Experimenter or subject None, but may be required by the platform used to conduct experiments. Prolific is only a participant recruitment/management tool.
Psychstudio Subject Cloud service. Experimenter uses own device and web browser to access account on the service. Participant/subject uses own device and web browser to participate in study.
PsyToolkit Experimenter or subject None.
Qualtrics Experimenter or subject None.
SHOEBOX Experimenter It is unclear if researchers or patients may use their own iPads or if they must purchase them with the SHOEBOX software. Compatible with most iPads, recommended transducers are RadioEar DD450 and E-A-RTONE 3A inserts
SpeakPipe Experimenter or subject None.
TabSINT Experimenter or subject
TeamHearing Experimenter or subject None needed for calibration by sensation level; calibrated equipment needed for calibration by pressure level.
Platform Cost for hardware:
Amazon MTurk None.
Cognition.Run None.
Django Django needs to run on a server (linux typically). This could be a machine in the lab or virtual private server in the cloud (like AWS/Linode/GCP). If renting a cloud server, there will be monthly costs.
Gorilla None associated with maintenance. Only optional hardware provided by the investigator.
hearX Hardware (tablet and headphones): $860-1330 ($1920 for system with inserts, $1400 for system with high-frequency testing) Video Otoscope: $200 – 730
ispring No specific costs.
Jacoti Cost of an iPhone or iPad with iOS 12.x or 13.x and a pair of Apple EarPods
jspsych Not applicable
MATLAB Recommended hardware for the MATLAB web app server is 1 logical CPU core and 2GB of ram per simultaneously connected research participant. Our current hardware needs are supported at no cost by institutional research IT group.
PART/BGC Science Variable, based on platform (roughly $400-$3000)
Prolific Not applicable
Psychstudio None.
PsyToolkit None.
Qualtrics None.
SHOEBOX See above.
SpeakPipe None.
TabSINT The cheapest known calibrated system runs on a Nexus 7 tablet and Vic Firth SIH2 headphones with a total cost of around $200 total. Experimenter calibrated systems could be cheaper.
TeamHearing Hardware not supplied, but software calibration file can be generated and incorporated into experiment.
Platform How is hardware calibration implemented on this platform? (Experimenter side, Subject side, Either side) What procedure is used for hardware calibration? (Electroacoustic, Perceptual, Both/Either)
Amazon MTurk None available Not applicable
Cognition.Run None available
Django None available Not provided by the platform per se, but not constrained by the platform either.
Gorilla Either side Both/Either will be closers. I want to provide more details here. It depends on whether the experimenter has access to the testing hardware. If the experimenter provides hardware, it can use electroacoustic-based calibration. If participants provide hardware, the best approach will be a perceptual-based calibration.
hearX None available Environmental noise is monitored in the testing environment. Company provides calibration of headphones every 12 months with an indicator that appears on the tablet to notify you when to send system back to the company for calibration.
ispring Subject side Perceptual
Jacoti Subject side Only measures ambient noise prior to testing.
jspsych Either side If the experimenter has access to hardware, electroacoustic calibrations are possible. On the subject side, the platform allows for mic input from the subject. If the approximate specs for the subject hardware are known (e.g., average specs for the brand of mic/device), it may be possible to get approximate calibrations from highly co-operative subjects. Perceptual/relative calibrations are likely the best option for subject-side implementation.
MATLAB Either side Both/Either
PART/BGC Science Either side Both/Either
Prolific None available Not applicable
Psychstudio Subject side Perceptual
PsyToolkit None available Not applicable
Qualtrics Subject side The subject can be asked to adjust their volume to a suitable level while an example stimulus (e.g., white noise) plays. There is no actual calibration performed, but the experimenter can create questions and screening tasks to check if the participant’s sound is audible.
SHOEBOX None available Ambient noise is monitored but calibration is done by the company, annually.
SpeakPipe None available Not applicable
TabSINT Experimenter side Electroacoustic
TeamHearing Either side Both/Either
Platform Can this platform document the hearing sensitivity of the subject (capable of threshold audiometry)? Can this platform quantify ambient noise levels?
Amazon MTurk Not possible No
Cognition.Run Not possible No
Django Relative thresholds (dB SL) Can be implemented
Gorilla Relative thresholds (dB SL) Yes – no audio stored
hearX Absolute thresholds (dB HL/SPL) Yes – no audio stored
ispring Not possible Yes – no audio stored
Jacoti Absolute thresholds (dB HL/SPL) Yes – no audio stored
jspsych Relative thresholds (dB SL) Yes – audio can be stored for offline analysis or not, as desired
MATLAB Relative thresholds (dB SL) Yes – no audio stored
PART/BGC Science Absolute thresholds (dB HL/SPL) No
Prolific Not possible No
Psychstudio Relative thresholds (dB SL) No
PsyToolkit Not possible No
Qualtrics Not possible No
SHOEBOX Absolute thresholds (dB HL/SPL) Yes – no audio stored
SpeakPipe Not possible No
TabSINT Absolute thresholds (dB HL/SPL) Yes – no audio stored, can be calibrated in dB SPL
TeamHearing Absolute thresholds (dB HL/SPL) No
Platform Please describe any details of the hardware and calibration interactions that are not captured by the questions above, including expected usage scenarios, if useful:
Amazon MTurk Doing any sort of audio recording, i.e. to check ambient levels, might be prohibited. Similarly with video of the subject.
Django These tasks will be front-end dependent. Django provides back end control. A linux computer running Django is the typical way to setup web servers. However, unlike old-school web app platforms like PHP, Django comes with many commonly-needed functions out-of-the-box and makes is possible for non-technical users to setup flexible web servers.
Gorilla In Q26, Gorilla runs on the Web Audio API which allows audio recording in the background. It’s a potential (not yet confirmed) functionality to allow analysis on noise fluctuation during test session but not directly on dB SPL. This may be helpful to identify how dynamic (not level) background noise is in the home environment. The quality of the hardware chain [computer/laptop/smartphone with soundcard + headphones/earbuds] is at the discretion of the investigator if participants are allowed to use own equipment at home.
Jacoti Intent of this product is to measure pure tone audiogram. Evaluates noise, then performs test assuming known equipment (platform of iPhone or iPad and AppleEarPods. Output level 10dB to 85 dB HL. Warning statement that "Jacoti Hearing Center requires a silent environment for taking the test. Please avoid places with air conditioning machines, fans and similar noise sources."
jspsych A long calibration sound may be used to ask subjects to set their computer volume to a comfortable but clearly audible level. Other testing sounds can be calibrated relative of this reference. This may be sufficient for supra threshold experiments. jsPsych allows mic input to be recorded. Clever ways to use that with subject consent may be considered by the users.
MATLAB By default there is no provided calibration support on this platform. Experimenters will need to develop their own calibration techniques that are suitable for their study. MATLAB web app experiments that require precise calibration should likely require the use of dedicated sound hardware that is distributed to participants in order to ensure calibration across users. Studies making use of perceptual calibration methods could likely be implemented across a large range of participant supplied hardware. Custom HTML/JS code could be used to capture a microphone signal for use in calibration. If supplied with specific recording hardware, SPL calibration could be possible.
PART/BGC Science Experimenter can provide a calibrated device (tablet, computer, phone) running an appropriate operating system and headphones associated with the calibration settings, or it can be deployed by allowing the participant to download on their own device, use their own headphones, and use only perceptual calibration. Mixtures are possible as well.
Prolific Not applicable
SHOEBOX Compatible with most iPads, recommended transducers are RadioEar DD450 and E-A-RTONE 3A inserts
TeamHearing Ambient noise level estimation could be easily added.