SenseLabOnline is a web application that functions as a complete listening test facility.
Web application for listening tests
SenseLabOnline handles all steps included in performing a best-practise listening test:
- Experimental setup and randomisation
- Presentation of test stimuli and gathering of data
- Automatic statistical analysis
- Access to external expert listening panels or consumers.
Once you have registered as a user of SenseLabOnline you can sign in and build a range of standard-compliant, as well as custom listening tests.
In-house tool for facilitating listening tests
SenseLabOnline will serve as an in-house tool for facilitating listening tests.
Because the system is built to run in a browser, you can distribute a listening test across physical locations – even to departments in other countries. All data transformation is done in Secure Socket Layer (SSL), and the application is approved with industry-standard code signing certificates. Currently, SenseLabOnline serves a number of clients within:
- Audio coding
- Hearing aid industry
- Automotive industry
- Consumer products (product sound quality)
- Sound/audio branding agencies
- Building acoustics.
Available test types of listening tests
SenseLabOnline offers a number of listening test in our online service.
Adapted sensory evaluation methods in SenseLabOnline
Virtually any type of listening test can be performed using SenseLabOnline. When you register as a client you will get a start package that includes the following test types:
- ITU-R BS.1116-3
- ITU-R BS.1534-2 (MUSHRA)
- ITU-T P.835
- ITU-T P.800 ACR/CCR/DCR
- Triangle test (ISO 4120)
- Pairwise comparison (ISO 5495)
- ABX/Duo-trio test (ISO 10399)
- Multiple Stimulus Comparison Rating
- Adapted sensory evaluation methodologies (attribute testing)
- Music in noise (MiN)
- Ideal Profile Method (IPM)
- Annoyance test
- Suitability test.
Advanced or custom listening tests
If you are interested in more advanced test types or custom versions of known methodologies this is easily achievable. Just contact the SenseLab team with your specifications and you will have a template ready (with complete statistical module) within 24 hours.
In SenseLabOnline you can test an unlimited number of systems (codecs, algorithms, etc.) in 16 different conditions (e.g. background noise scenarios, SNR etc.) with 24 different samples (source material). However, we always recommend staying within the normal recommendations regarding test duration to avoid listener fatigue.
Pushing the limits for online listening tests
The sound player used by SenseLabOnline has been developed exclusively for SenseLabOnline to meet the high-performance criteria specified in the international standards for audio evaluation testing. As well as meeting high-performance criteria the sound player runs exclusively via your web browser, meaning that no additional programs need to be installed in order to build or run tests.
- Windows (XP, Vista, 7), Mac OS (10.7.x, 10.8.x)
- Safari (latest version), Chrome (latest version, Windows only), Internet Explorer (7, 8, 9)
- Microsoft® Silverlight (free browser plug-in)
- 1280×800 screen resolution (minimum)
- Dual-core processor (minimum)
- 2 GB ram (minimum).
- Supported format: WAV
- Supported sample rates: 8 – 96 kHz
- Resolution: 16 bit
- Channels: Mono or stereo
- Maximum file length: 45 s
- Sounds per screen/trial: up to 13
- Crossfade time between samples: 40 ms.
Automatic statistical analysis
SenseLabOnline has implemented automatic statistical analysis procedures. Any time during a test you can get an instant and full statistical analysis of the preliminary results.
The automatic routines check data quality and basic assumptions before proceeding to plotting of the results.
Basic analysis methods like 2-way Analysis of Variance (ANOVA), showing mean values with confidence intervals, plotting of interaction between experimental variables (etc.) are always included.
For more advanced test types e.g. tests with multiple test questions/attributes, you will get a complete overview of the results using multivariate data analysis methods such as Principal Component Analysis (PCA), and Hierarchical Multiple Factorial Analysis (HMFA).
Current available statistical modules include
- Analysis of Variance (ANOVA)
- TUKEY HSD
- Mean scores with confidence intervals
- Interaction plots
- Assessor feedback
- Assessor performance (eGauge)
- Hierarchical multiple factorial analysis (HMFA)
- Multi Dimensional Scaling (MDS)
- Factorial Analysis on Distance Tables (FADT)
- Check-All-That-Apply (CATA) analysis
- Density plots
- Spider/radar plots.
You will receive the results as jpg files, and have access to raw data file for all your experiments.
- ITU-R BS. 116-3 Evaluation of high quality audio systems having small impairments
- ITU-R BS.1534-2 – MUSHRA - Assessment of sound quality for multiple systems
- ITU-T P.800 methods - Methods for subjective determination of transmission quality
- ITU-T P.835 - Methodology for evaluating speech communication systems that include noise suppression
- Sensory profiling - Methodology for objectively characterizing and measuring product characteristics