Analysing a Soundfile with speech

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

Analysing a Soundfile with speech

dufay
Hello everybody,

can someone help me with the following:

i've a soundfile with speech: some words and breaks between them.
now i want to analyse the file and get an Array with the start-times
of the words.
what's the best way to handle it?

thanks a lot for help.

Johannes



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Analysing a Soundfile with speech

altern2
You could try DetectSilence and play around with the amp parameter to
find out the silence/signal changes. Another option might be Onsets. I
am not sure about which one would work best in this case

enrike

og., 2018.eko urtren 11a 12:10(e)an, [hidden email] igorleak
idatzi zuen:

> Hello everybody,
>
> can someone help me with the following:
>
> i've a soundfile with speech: some words and breaks between them.
> now i want to analyse the file and get an Array with the start-times
> of the words.
> what's the best way to handle it?
>
> thanks a lot for help.
>
> Johannes
>
>
>
> --
> Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html
>
> _______________________________________________
> sc-users mailing list
>
> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
>


_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Analysing a Soundfile with speech

Josh Parmenter
There isn’t always silence between words. In fact, most speech doesn’t have a pause in it, really, until a sentence break or pause.

In English, looking for noisier parts of the signal can sometimes help show where syllables are (since that implies a consonant sound), but even that would be a problem for word detection and words without hard consonants.

Best,
Josh

> On Jan 11, 2018, at 3:24 AM, [hidden email] wrote:
>
> You could try DetectSilence and play around with the amp parameter to find out the silence/signal changes. Another option might be Onsets. I am not sure about which one would work best in this case
>
> enrike
>
> og., 2018.eko urtren 11a 12:10(e)an, [hidden email] igorleak idatzi zuen:
>> Hello everybody,
>> can someone help me with the following:
>> i've a soundfile with speech: some words and breaks between them.
>> now i want to analyse the file and get an Array with the start-times
>> of the words.
>> what's the best way to handle it?
>> thanks a lot for help.
>> Johannes
>> --
>> Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html
>> _______________________________________________
>> sc-users mailing list
>> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
>> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
>> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
>
>
> _______________________________________________
> sc-users mailing list
>
> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/


_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

RE: Analysing a Soundfile with speech

Raul Islas
Try Praat . Download it from Sourceforge. It is helpful when analyzing vowels and consonants, words and phrases as well. Give it a try. 

From: [hidden email]
Sent: ‎1/‎11/‎2018 8:27
To: [hidden email]
Subject: Re: [sc-users] Analysing a Soundfile with speech

There isn’t always silence between words. In fact, most speech doesn’t have a pause in it, really, until a sentence break or pause.

In English, looking for noisier parts of the signal can sometimes help show where syllables are (since that implies a consonant sound), but even that would be a problem for word detection and words without hard consonants.

Best,
Josh

> On Jan 11, 2018, at 3:24 AM, [hidden email] wrote:
>
> You could try DetectSilence and play around with the amp parameter to find out the silence/signal changes. Another option might be Onsets. I am not sure about which one would work best in this case
>
> enrike
>
> og., 2018.eko urtren 11a 12:10(e)an, [hidden email] igorleak idatzi zuen:
>> Hello everybody,
>> can someone help me with the following:
>> i've a soundfile with speech: some words and breaks between them.
>> now i want to analyse the file and get an Array with the start-times
>> of the words.
>> what's the best way to handle it?
>> thanks a lot for help.
>> Johannes
>> --
>> Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html
>> _______________________________________________
>> sc-users mailing list
>> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
>> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
>> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
>
>
> _______________________________________________
> sc-users mailing list
>
> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/


_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

RE: Analysing a Soundfile with speech

dufay

Try Praat . Download it from Sourceforge. It is helpful when analyzing
vowels and consonants, words and phrases as well. Give it a try.  

Ok, i will. Thanks!



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

RE: Analysing a Soundfile with speech

Scott Carver
I did a pretty elaborate project using PRAAT + SC a year or two ago - PRAAT is a huge pain to use, but you can kludge a workflow together if absolutely necessary. 

In particular, I'd recommend the prosogram tools here: http://bach.arts.kuleuven.be/pmertens/prosogram/download.html
If you run prosogram.praat, and use the Calculate Intermediate Data Files task on your sound files, you should end up with some "_styl.txt" files that have syllable information. I don't think they do word segmentation, since that requires analysis of the language content, but you can easily break things up into syllable clusters with e.g. more than 0.3 seconds of space in between.

- S

On Fri, Jan 12, 2018 at 10:53 AM <[hidden email]> wrote:

Try Praat . Download it from Sourceforge. It is helpful when analyzing
vowels and consonants, words and phrases as well. Give it a try.

Ok, i will. Thanks!



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Analysing a Soundfile with speech

George Edward Thomas
In reply to this post by dufay
Hello Johannes,

have you tried using SCMIR (SuperCollider Music Information Retrieval
Library)? Its part of the sc3-plugins written by Nick Collins.

https://composerprogrammer.com/research/scmir.pdf

all the best!

Jorge.



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/