fastACI toolbox

fastACI toolbox: the MATLAB toolbox for investigating auditory perception using reverse correlation.

fastACI toolbox

This is the repository of the fast Auditory Classification Images (fastACI) project. The toolbox is controlled using the command line of MATLAB. It does not have (yet) a graphical interface.

With this toolbox you can run listening experiments as used in the studies varnet2013, varnet2015, varnet2021, osses2021c, osses2022b, varnet2022a, osses2023a, osses2022b, osses2024, and carranante2023 (see the full citations in the section "References"). You can also reproduce some of the figures contained in the mentioned references.

Citation key	fastACI experiment name	Type of background noise	Target sounds
varnet2013	speechACI_varnet2013	white	/aba/-/ada/, female speaker
varnet2015	speechACI_varnet2015	white	/alda/-/alga/-/arda/-/arga/, male speaker
osses2021c	speechACI_varnet2013	speech shaped noise (SSN)	/aba/-/ada/, female speaker
varnet2022a	modulationACI	white	modulated or unmodulated tones
osses2024	speechACI_Logatome	white, bump, MPS	/aba/-/ada/, male speaker (S43M) from the OLLO database
carranante2023	speechACI_Logatome	bump	Pairs of contrasts using /aba/, /ada/, /aga/, /apa/, /ata/ from the same male speaker (S43M) in OLLO
osses2023a	segmentation	random prosody	Pairs of contrasts: /l'amie/-/la mie/, /l'appel/-/la pelle/, /l'accroche/-/la croche/, /l'alarme/-/la larme/
osses2023b	toneinnoise_ahumada1975	white	Tone-in-noise experiment with 100-ms 500-Hz sinusoids temporally centred in Gaussian noises of 500 ms

Make sure that you follow the steps indicated in the section Installation (below) the first time you use the toolbox.

How to cite this repository

This repository can be cited as follows: The fastACI toolbox was used (Osses & Varnet, 2022).

If a model version is cited (in this example: release fastACI v1.2):

A. Osses & L. Varnet (2022). "fastACI toolbox: the MATLAB toolbox for investigating auditory perception using reverse correlation (v1.2)"

If a specific commit is cited (in this example: commit cc9d9cf):

A. Osses & L. Varnet (2022). "fastACI toolbox: the MATLAB toolbox for investigating auditory perception using reverse correlation," Github commit cc9d9cf.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Running a listening experiment

Next we present the command line required to run each of the ACI experiments that are available in our toolbox. These examples assume that the listener will be named S01 (standing for Subject 01), however any character-based name can be used instead.

modulationACI: Experiment as in varnet2021

fastACI_experiment('modulationACI','S01');

speechACI_varnet2013

fastACI_experiment('speechACI_varnet2013','S01','white'); % to run it as in varnet2013
fastACI_experiment('speechACI_varnet2013','S01','SSN');   % to run it as in osses2021c

speechACI_varnet2015: Experiment as in varnet2015

Alda/Alga/Arda/Arga discrimination using a male speaker

fastACI_experiment('speechACI_varnet2015','S01','white');

speechACI_Logatome: Experiments using speech samples from the Logatome corpus in French (under construction)

Aba/Ada discrimination using a female speaker (S41F from the Logatome corpus):

fastACI_experiment('speechACI_Logatome-abda-S41F','S01','white');

Aba/Ada discrimination using a male speaker (S43M from the Logatome corpus), as in osses2022b:

fastACI_experiment('speechACI_Logatome-abda-S43M','S01','white');

Additional VCV contrasts using a male speaker (S43M from the Logatome corpus), as in carranante2023, using bump noises:

fastACI_experiment('speechACI_Logatome-abda-S43M','S01','bumpv1p2_10dB');
fastACI_experiment('speechACI_Logatome-adga-S43M','S01','bumpv1p2_10dB');
fastACI_experiment('speechACI_Logatome-apta-S43M','S01','bumpv1p2_10dB');    
fastACI_experiment('speechACI_Logatome-abpa-S43M','S01','bumpv1p2_10dB');
fastACI_experiment('speechACI_Logatome-adta-S43M','S01','bumpv1p2_10dB');    

Simulating a listening experiment

A listening experiment can be simulated using an artificial listener or, in other words, an auditory model. So far, we have validated the use of the models osses2021 (Osses & Kohlrausch, 2021), osses2022a (to be published), and king2019 (King et al., 2019). The models osses2021 and king2019 are both available within AMT 1.0 (or more recent), osses2022a is exclusively available in our toolbox.

To run simulations you only have to use the corresponding model as the subject name. To use osses2021 in the simulation of the experiment speechACI_varnet2013 using SSN noises, you need to type in MATLAB:

fastACI_experiment('speechACI_varnet2013','osses2021','SSN');   % to run it as in osses2021c

or, to use king2019:

fastACI_experiment('speechACI_varnet2013','king2019','SSN');   

More elaborate simulations can be automatically run using the scripts pres_osses2022_02_AABBA_1_sim.m and publ_osses2021c_DAGA_1_sim.m, among other scripts. In the next section, all the simulations in osses2021c can be reproduced using the osses2021 model with two different decision back ends. This is related to the script publ_osses2021c_DAGA_1_sim.m.

Demo: Obtaining the figures from osses2021c

To run the simulations from Osses & Varnet (2021, DAGA) you need to run in the MATLAB command line, and follow the instructions that will appear on the screen:

publ_osses2021c_DAGA_1_sim;

To obtain figures 1 to 4 (all the paper figures) you need to run, either of the following commands:

publ_osses2021c_DAGA_2_figs('fig1a'); % REQUIRED: manual download of experimental data (see below)
publ_osses2021c_DAGA_2_figs('fig1b'); % REQUIRED: manual download of experimental data (see below)
publ_osses2021c_DAGA_2_figs('fig2');
publ_osses2021c_DAGA_2_figs('fig3a');
publ_osses2021c_DAGA_2_figs('fig3b');
publ_osses2021c_DAGA_2_figs('fig4'); % REQUIRED: manual download of experimental data (see below)

To obtain Fig 1A or Fig 1B, you require to manually download (in advance) the experimental dataset, which is available on Zenodo (see ref. osses2021c_data).

publ_osses2021c_DAGA_0_checkdata;

Installation

The following are the general instructions to get the fastACI toolbox for MATLAB operative in your computer. The toolbox has been tested on Windows and Linux, using MATLAB (versions R2012b-R2020b).

1. Download or clone the fastACI project to your local computer (one way: press the button 'Code'->Choose 'Download ZIP' and unzip somewhere).
2. This toolbox requires the Auditory Modelling Toolbox v.1.0 (AMT 1.0 or higher) that can be downloaded from here. After the download you are not expected to do anything else, as the AMT toolbox will automatically be initialised in our next step:
3. Open and run the script startup_fastACI.m. This script will add all the paths under the fastACI toolbox to your local MATLAB path and it will run the script amt_start.m to initilise the AMT toolbox. If the AMT toolbox is not found you will be able to indicate your alternative location using a pop-up window.

References for the fastACI toolbox

The references are sorted alphabetically (first author's last name) and then more recent first.

osses2024	A. Osses, & L. Varnet (2024). A microscopic investigation of the effect of random envelope fluctuations on phoneme-in-noise perception. J. Acoust. Soc. Am. 155, p. 1469-1485 (doi: 10.1121/10.0024469, Download paper)
carranante2023	G. Carranante, M. Giavazzi, & L. Varnet (2023). Auditory reverse correlation applied to the study of place and voicing: Four new phoneme-discrimination tasks. Forum Acusticum 2023.
king2019	A. King, L. Varnet, & C. Lorenzi (2019). Accounting for masking of frequency modulation by amplitude modulation with the modulation filter-bank concept. J. Acoust. Soc. Am. 145, p. 2277-2293 (doi: 10.1121/1.5094344, Download paper)
osses2023a	A. Osses, E. Spinelli, F. Meunier, E. Gaudrain, & L. Varnet (2023). Prosodic cues to word boundaries in a segmentation task using reverse correlation. To be submitted.
osses2023a_data	A. Osses, E. Spinelli, F. Meunier, E. Gaudrain, & L. Varnet (2023). Raw and post-processed data for the study of prosodic cues to word boundaries in a segmentation task using reverse correlation (doi: 10.5281/zenodo.7865424)
osses2023b	A. Osses, & L. Varnet (2023). Using auditory models to mimic human listeners in reverse correlation experiments from the fastACI toolbox. To be presented at Forum Acusticum 2023.
osses2023b_data	A. Osses, & L. Varnet (2023). Raw and post-processing data for using auditory models to mimic human listeners in reverse correlation experiments from the fastACI toolbox (doi: 10.5281/zenodo.7886232)
osses2022d	A. Osses, C. Lorenzi, & L. Varnet (2022). Assessment of individual listening strategies in amplitude-modulation detection and phoneme categorisation tasks. International Congress on Acoustics, 24-28 October, Gyeongju, Korea (Download presentation, Download proceedings)
osses2021c	A. Osses & L. Varnet (2021). Consonant-in-noise discrimination using an auditory model with different speech-based decision devices. DAGA conference. Vienna, Austria. (Download paper)
osses2021c_data	A. Osses & L. Varnet (2021). Noise data for the study of consonant-in-noise discrimination using an auditory model with different speech-based decision devices. Experimental data for osses2021c (doi: 10.5281/zenodo.5483835)
varnet2022a	L. Varnet & C. Lorenzi (2022). Probing temporal modulation detection in white noise using intrinsic envelope fluctuations: A reverse correlation study. J. Acoust. Soc. Am. 151, p. 1356-1366 (doi: 10.1121/10.0009629)
varnet2022a_data	L. Varnet (2021). AM revcorr data. Experimental data for varnet2022a (doi: 10.5281/zenodo.5571719)
varnet2022b	L. Varnet, C. Lorenzi, & A. Osses (2022). Probing amplitude-modulation detection and phoneme categorization with auditory reverse correlation. Congrès Français d'Acoustique, 11-15 April, Marseille, France (Download presentation)
varnet2015	L. Varnet, K. Knoblauch, W. Serniclaes, F. Meunier, & M. Hoen (2015). A psychophysical imaging method evidencing auditory cue extraction during speech perception: A group analysis of auditory classification images. PLoS one 3, p. 1-23 (Download paper)
varnet2013	L. Varnet, K. Knoblauch, F. Meunier, & M. Hoen (2013). Using auditory classification images for the identification of fine acoustic cues used in speech perception. Front. Hum. Neurosci. 7, p. 1-12 (Download paper)

Other references

P. Majdak, C. Hollomey, & R. Baumgartner (2022). AMT 1.x: A toolbox for reproducible research in auditory modeling, Acta Acustica, 6, 19. (doi: 10.1051/aacus/2022011).

A. Osses, L. Varnet, L. Carney, T. Dau, I. Bruce, S. Verhulst, & P. Majdak (2022). A comparative study of eight human auditory models of monaural processing, Acta Acustica, 6, 17 (doi: 10.1051/aacus/2022008).

A. Osses & A. Kohlrausch (2021). Perceptual similarity between piano notes: Simulations with a template-based perception model. J. Acoust. Soc. Am. 149, p. 3534-3552 (doi: 10.1121/10.0004818).

Acknowledgements

The development of the fastACI toolbox was funded by the ANR grant "fastACI" attributed to Léo Varnet (ANR-20-CE28-0004) and was further supported by the "FrontCog" grant (ANR-17-EURE-0017).

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.