eprime capabilities

[Michiel Spape]

Hiya,

Thanks for the clarification. I understand what you want e-prime, or some other psycho-prog to do, but if so, you may still ask yourself HOW MUCH you want to do this - because it will be difficult. The thing is, to my knowledge, nobody has written an automated script to do the thing you want, which would be:
1. Recording vocal responses to buffer (to .wav would be probably too slow - real signal analysis software can do this sort of thing on the fly, but I'd imagine you just might be able to come up with something based on a flow of save-to-wav then read-from-disk then analysis then feedback, but read on)
2. Analysis using sophisticated signal analysis. Language labs may have that kind of thing, but those I know prefer to do this off-line (outside the experiment) because it is VERY tricky indeed. For example, you might (at least most people do) think syllables can be easily observed in a wave-form, so that the word 'programming' has a wave-form of roughly 3 distinct waves 'pro' 'gram' 'ming'. After finding out what the first wave is (thus, the first syllable), you will easily see which was has the higher volume, and thus detect stress, right? (I see PVI detects it based on syllable duration, but the point is merely to illustrate the difficulty of connecting the higher to the lower order - loudness to wave amplitude, here) 
No, the spaces seldom exist, words, as waveforms, are often jumbled together, syllables may be bigger points of empty sounds than actual spaces, or vice versa (thus two words will be detected as one, or vice versa), there's _massive_ variance between participants, and, more annoyingly, between separate phonemes (P sounds produce big 'pops', for instance). Then, the 'high volume' means either high peaks, high absolute averages (envelope), high RMS? A P-pop, for instance, has a much higher peak than the (very subtle) difference between stress, therefore making it really quite difficult to 'see' stress in syllables, let alone programming it so a pc can do it automatically. Having produced quite a bit of spoken word music, myself, I lack the ability to see stress in wave-forms, actually... 

I am almost certain that normal stimulus presentation software (e-prime, DMDX, superlab, etc) might be able to do what you want it to do - but you'd need great amounts of experience both in advanced signal processing and (to use such knowledge in the way you want it to) lots in programming languages. You might, if you go that path, find it much easier to work using matlab, which has signal processing scripts pre-written (and, presumably, language labs doing the sort of thing you want, have scripts to work out), which should make things a lot easier. Also, PsychoPy might be able to help you, since you could presumably incorporate pre-written Python scripts. 

My apologies if this all sounds very drastic, I'm mainly trying to show you that, as David pointed out, it is more difficult than you may think. Perhaps this is because E-Prime (and similar software) seems to have been designed for experimental psychology (priming experiments, etc) rather than as a piece of all-purpose experimentation/analysis software. What you want, then, sounds a bit like me hoping E-Prime (or such software) will not only present my stimuli, but also analyse EEG and give biofeedback. The easiest way I could think of to have E-Prime do what you want is to employ a similar setup: have one computer (with the speech analysis software running) connected to another one (serial cable) and another running E-Prime (granted, it's not a *very* easy solution). 

In summary then, three thoughts: try Matlab (and see if you can get a speech analysis toolkit for matlab), try psychopy (and see if you can get a speech toolkit in Python) and try to find a dedicated, automated speech analysis toolkit from some place that might work in conjunction with other bits of software.

Cheers,
Mich

Michiel Spapé
Research Fellow
Perception & Action group
University of Nottingham
School of Psychology
www.cognitology.eu

-----Original Message-----
From: e-prime at googlegroups.com [mailto:e-prime at googlegroups.com] On Behalf Of thha4757 at uni.sydney.edu.au
Sent: 10 December 2010 06:50
To: e-prime at googlegroups.com
Subject: RE: eprime capabilities

Sorry for the confusion. 

I want e-prime to record vocal responses, and determine whether the duration of the first syllable is longer than the second syllable, and vice versa. So yes, i wanted e-prime to do real-time analysis and provide instant feedback on the responses.
PVI is a calculation equation about the duration of syllables to determine stress, that I assume would have to be part of the programming script in order for it to be able to do. Currently I am fiddling with DMDX to see if it is able to do this kind of analysis. 

Thanks for your help.
________________________________________
From: e-prime at googlegroups.com [e-prime at googlegroups.com] on behalf of David McFarlane [mcfarla9 at msu.edu]
Sent: Friday, December 10, 2010 2:52 AM
To: e-prime at googlegroups.com
Subject: RE: eprime capabilities

Merry,

I had pretty much the same reaction as did Mich
to your question.  EP can readily present your
visual and auditory stimuli and present feedback,
and might even collect the subjects' responses
(EP2 can even record subjects' vocalizations into
.wav files).  But by "have the computer judge ...
whether they correctly used weak-strong (WS) or
strong-weak (SW) stress" do I correctly
understand that you wish EP to do real-time vocal
recognition & analysis?  That *is* a tall
order!  Do you know of any other current software
or hardware that can not only recognize WS vs. SW
stress, but do so reliably enough to pass
scientific muster?  Or is that what the PVI does,
based merely on the raw sound-pressure waveforms?

To be sure, in principle EP might well be
*capable* of "judging" WS vs. SW stress, but it
would pose a formidable programming challenge and so would not be *practical*.

But perhaps I misunderstand and you have something simpler in mind.

-- David McFarlane, Professional Faultfinder


At 12/9/2010 07:27 AM Thursday, Michiel Spape wrote:
>Hi,
>What exactly do you wish E-Prime to calculate?
>"whether they correctly used weak-strong (WS) or
>strong-weak (SW) stress (this could be
>calculated using a pairwise variability index
>(PVI) - a positive PVI would indicate SW, and a
>negative PVI would indicate weak-strong stress)"
>would mean that the computer A) records vocal
>responses (possible in E-Prime), B) that vocal
>responses are classified as WS/SW, right? I'd
>imagine if you would do the test semi-automatic
>- let e-prime record responses and so on, and
>then have the experimenter classify the words
>immediately after the participant says them (I
>mean, after every trial, for instance), this
>would be possible. If you wish to know whether
>E-Prime can do any sort of higher-order
>analysis, the answer is no (unless you are
>REALLY REALLY good with programming, in which
>case you could do the same with much more ease
>in .NET, c#, Matlab, etc). It cannot do signal
>analysis at all (therefore no acoustic analysis
>either), nor is it (I imagine) intended to do such things.
>
>BUT, the semi-automatic way is probably much
>easier for anyone involved. I mean, the few
>times I did a Stroop task with vocal responses,
>I classified the subjects' responses (tip for
>those thinking of it: classify the first letter
>of their literal response ('g' if green, for
>instance), not whether it is correct, or _you_
>will have a Stroop effect as well as the
>subjects!) as well. You will have to endure
>[number of subjects * boredom involved in task = ] total boredom though.
>
>Hope that helps,
>Mich
>
>
>Michiel Spapé
>Research Fellow
>Perception & Action group
>University of Nottingham
>School of Psychology
>www.cognitology.eu
>
>From: e-prime at googlegroups.com
>[mailto:e-prime at googlegroups.com] On Behalf Of Merry Ha
>Sent: 09 December 2010 11:32
>To: e-prime at googlegroups.com
>Subject: eprime capabilities
>
>Hi,
>I'm a speech pathology honours student from the
>University of Sydney working on my project at
>the moment using e-prime for the first time.  My
>honours project will be used for children with
>Childhood Apraxia of Speech (CAS).
>
>We will be presenting novel word stimuli on the
>screen visually and auditorily, and then have
>the child to imitate the computer, or read off
>the screen. Children with CAS tend to say words
>with equal stress, and the aim of the experiment
>was to have the computer judge and provide
>feedback whether they correctly used weak-strong
>(WS) or strong-weak (SW) stress (this could be
>calculated using a pairwise variability index
>(PVI) - a positive PVI would indicate SW, and a
>negative PVI would indicate weak-strong stress).
>Do you know if this programming is possible for e-prime?
>
>Merry

--
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.

This message and any attachment are intended solely for the addressee and may contain confidential information. If you have received this message in error, please send it back to me, and immediately delete it.   Please do not use, copy or disclose the information contained in this message or in any attachment.  Any views or opinions expressed by the author of this email do not necessarily reflect the views of the University of Nottingham.

This message has been checked for viruses but the contents of an attachment
may still contain software viruses which could damage your computer system:
you are advised to perform your own checks. Email communications with the
University of Nottingham may be monitored as permitted by UK legislation.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.