Speech Perception - Science topic

I am not a native speaker of English but based on my professional expertise, I can help detect distinctive realisation. let me know if I can be of help

Hi Paolo! We had a similar issue and ended up building a survey platform entirely around audio and video (phonic.ai). There's a super generous free-tier if you want to try it out. Here's a demo: https://survey.phonic.ai/5ed023df554d721ab204f049

We're also embeddable into various other research platforms if you have preferences (docs.phonic.ai).

You are especially kind and gentle, Thank you...I'll look for some dear Isabel

As a Belgium-based researcher, I can recommend asking a colleague from a university in the Dutch-speaking area (Flanders and the Netherlands) for permission to post your call for participants on one of their general message boards. KU Leuven, the University of Antwerp, and any 'Hogeschool' could be of help. Students know the importance of surveys, and while the turnout response may be fairly low, it would certainly increase your engagement and help your study. Facebook pages of certain Flemish and Dutch cities might also allow messages for research purposes. Of course, you would have to keep in mind to preserve the representational balance in your (control) group. This apply to both the age/ level of education a large number of students would represent, as to the cultural and linguistic differences between members of the Flemish and Dutch participants respectively (not to mention the significant differences between speakers from different Flemish provinces).

If I may ask, how many people would you need to comfortably conduct your study?

Isabel,

I know of many studies on when phoneme detectors develop but never thought about your question.

Are we even sure that people perform speaker normalization in the same sense that speech recognition systems do (or at least used to before deep learning)?

Since we do know that one of the early layers of human speech recognition processing produces a sequence of best phoneme guesses, and higher layers can force backtracking to try the next best guesses, it is possible to check psychophysically if there is such a normalization at lower layers by checking the processing speed, and if there is a change in processing speed after normalization takes place.

It is also an interesting question how the speaker recognition and speech recognition processes are related. Do we first recognize the speaker and then apply that speaker's phoneme recognizers? I remember reading a paper a few years ago about recognizing accents before recognizing speech (but can't find the reference).

Another clue is Hearing different accents at home impacts language processing in infants from U Buffalo (www.sciencedaily.com/releases/2017/12/171205104127.htm ) which found that infants exposed to multiple accents before 12 months develop different recognition strategies. See also the JASA reference there and

Linguistic processing of accented speech across the lifespan (www.frontiersin.org/articles/10.3389/fpsyg.2012.00479/full).

Finally Language Discrimination by English-Learning 5-Month-Olds: Effects of Rhythm and Familiarity (labfon.letras.ulisboa.pt/personal/sfrota/aeli/Nazzi_Jusczyk_Johnson_2000.pdf) may be of use.

Would love to hear what you learn!

Y(J)S

Nice Dear Gerry Leisman

I hope that the References of the manuscript that I posted in the Attachment will be useful to you.

Hi, I have seen this software. One of my colleagues (Dr. Tayebeh Ahmadi) has conducted a research with this software on CAPD children. She is now preparing the article.

She conducted TFS test on me. It was interesting.

Please try https://www.mathworks.com/matlabcentral/fileexchange/47333-pesq-matlab-driver?requestedDomain=www.mathworks.com or https://www.mathworks.com/matlabcentral/fileexchange/33820-pesq-matlab-wrapper?focused=5203883&tab=function

Andy Clark's Mindware (OUP, 2001), while billed as an introduction to the 'Philosophy of Cognitive Science,' is worth a look. It isn't bogged down in history (although he's certainly aware of it, and mentions it when it helps explain things), and favours trying to give a sense of current (as of 2001) theoretical issues and research trends.

Its a nice question and actually a doubt comes to a researcher when it comes to the concept of 'test retest reliability". I feel to measure reliability you need to have the tests repeated soon after a first one and preferably within a span of two weeks or so. To make sure that it covers the inter session and intra session reliability. Also not giving much of the time for the individual to adjust for the test patterns. In one of my research I have repeated the second recording after half an hour of first recording and the third recording after a span of 6 days of first recording. Without allowing any training within these 6 days.

You're looking at more of a psychological phenomenon than an acoustic one. It's similar to the "phonetic restoration" effect that's been studied in the past.  

If you think of  the human auditory system as actively seeking evidence for a particular speech event and finding sufficient evidence for it in the sound then you get the observed phantom percept.  A different version of the effect can be observed in "babble", what people will hear in recordings of superimposed voices.

Actually, if you want to pursue this systematically it could get interesting. For example, can you find sounds that, across listeners, appear to be fertile sources of illusion? What are their characteristics?

Dear Giannis, there are no published studies that have compared all the three companies. However, based on understanding technology and physiology the benefit will be approximately similar. 

There are some published studies comparing various technologies available out-there.

Vijay  

Hello all

I really appreciate all your support. 

Regards

Fateme

but I didn't find it.?   does not make sense.

You say that you downloaded it. So, perhaps you had difficulties to install and use it?  Then your question should be "How can I install and use it".  Just puzzled.

Dear Mr Yang Zhang,

  Thanks indeed for your comments and suggestions.

  I finally used Praat and trained experimenters in Praat at the designated locations. The experiments were successfully finished

Dear Anna,

I agree with the other contributions that there is a complex interplay of factors in speech perception and in particular in Second Language Acquisition research. The issue of individual differences is therefore very interesting; Antje Meyer at the MPI in Nijmegen has her research focus on individual differences in speech processing - you also might find some publications there (for instance an edited book by Leah Roberts and Antje Meyer (2012) on Individual Differences in Second Language Learning, published by Wiley: http://eu.wiley.com/WileyCDA/WileyTitle/productCd-111848634X.html).

If I understand your question correctly, you have a group of participants (or more groups) that behaved in a particular way and you would like to make a statement for the group (or groups) as a whole. One option would be to inspect the 95% confidence interval for each participant individually (calculated from the different trials that the participants gave responses to). That would give you a first indication on the differences across participants in your group(s). 

Best,

Bettina

Best, C. T., McRoberts, G. W., & Goodell, E. (2001). Discrimination of non-native consonant contrasts varying in perceptual assimilation to the listener's native phonological system. Journal of the Acoustical Society of America, 109(2), 775-794.

Antoniou, M., Best, C. T., & Tyler, M. D. (2013). Focusing the lens of language experience: Perception of Ma'di stops by Greek and English bilinguals and monolinguals. Journal of the Acoustical Society of America, 133(4), 2397-2411.

I suggest that you read Chapter 1 in Volume 1 of the Second Edition of Steven's Handbook of Experimental Psychology. 1988. Published by John Wiley. 

It was written by Duncan Luce and Carol Krumhansl. Carol was a student of Roger Shepard's who developed multidimensional scaling. The answer by Morgan above is ver important. This is because it gives you the assumptions that have to be met before you can use this scaling method. 

The most common fluency disorders are stuttering and cluttering. Cluttering is also described as a language disorder. Cluttering is usually not accompanied by fright or fear, or identification of problems with specific words or sounds. A person who clutters does not perceive the speech as deviant.

Question 1: If your research question is to what degree you can use discrimination performance as a predictor for speaking/listening performance, then it seems like you might not be able to. If subjects are at chance level, this can't act as a predictor. However, individuals might vary here and for certain individuals, it might be predictive. How much inter-subject variability was observed?

Question 2: Identification typically requires long-term memory representations (though, this can depend on the task). How were naive listeners/speakers trained in the Chinese tone categories prior to the experiment? By contrast, discrimination relies on both psycho-acoustic and phonetic cues. It's perfectly possible for a person to be able to distinguish two tones very well but not know the categories.

Question 3: If the discrimination results were at chance level and you are saying that they showed a significant relation to the speaking/listening test, then one can only conclude that subjects did extremely poorly in speaking/listening. It's not clear to me that you can conclude anything from this.

As far as your last question "What should be the fair conclusion when only one of the four measures shows the statistical significancy?", I am a bit confused. You only mentioned three possible measures (identification accuracy, discrimination accuracy, and RT for discrimination). What was the statistical test? What was the relation that you examined?

A typical example of this might be the McGurk effect. Depending on which articulatory movement subjects see, their perception of the auditory stimulus (usually b,d,g) changes. You will find a lot of references for the McGurk effect in the literature. They discuss auditory-visual fusion. (see for instance work by Jean Luc Schwartz, Grenoble).

Not sure how much you want to look at pictures or movements. McGurk relates to movement and its perception, not to static pictures.

A good review of information about speech banana and how the speech banana is derived. How the consonants and vowels placed is given in speech banana is given in review articles.

Olsen WO, Hawkins DB, Van Tasell DJ. (1987). Representations

of the long-term spectra of speech. Ear Hear,8:1008-108S .

It is very difficult to get commercially available linear hearing aids. Phonak still offers a limited number but the stock dates back many years. Their current product line along with Siemens, Oticon, Widex, Starkey, Unitron and Bernafon do not include a linear hearing aid. Some hearing aids can be programmed to be linear in gain, Starkey and Siemens and Phonak but sound cleaning tools will still be active.

personally I like

"Hearing Instrument Technology for the Hearing Health Care Professional by Andi Vonlanthen and Horst Arndt" as a very good overview of hearing aid technology and fitting strategies. However if you are looking at the different fitting strategies, i also suggest you read the how they were derived. I have been less than impressed by the methodology utilized by some.

Thanks you very much for your help

Hi there,

I would look into the work of Dr. Peter Assman at the University of Texas at Dallas (my old alma mater). Thomas Powell at Louisiana State has also been doing work in clinical phonetics and speech perception. He also has an account here, so you might want to reach out to him.

Attached are an article by Gregory Hickok.

Hope it helps.

Brian Roper