Hi everyone.
I'm new here and I'm so glad I found this website 'cause I was really having trouble until now.

Ok, I'm not into making musics as such.
I'm actually learning English and doing some experiment on linguistics stuff especially on prosody, you know like intonation and stress in a language.

Ok, I'm simulating a situation where a baby listens to speech sound in surroundings while the baby doesn't recognise what the words actually mean.

Babies at first listen to the rhythm of the language and familiarise with them and I want to do this as a part of English training.

Basically, I need to modify a speech file so I can listen to the rhythm of English language (intonation, stress, high and low pitch) and rhythm only. And I shouldn't be able to make out what the speak is actually saying.

Say, I have an audio file which contains a native English speaker's speech, I need to modify the audio file so that I can hear the speed, pause, the pitch and all of that except the actual subtle differences in consonants or even vowels.
In other words, I shouldn't be able to make out A from B from C.

So, if a speaker says "how do you do", it should to me sound like "ouyouuu~~" or somthing like that ^^;
Actually, it doesn't matter how it would sound as long as I can hear the rhythm and pitch correctly.

I guess it's like converting speech file into a musical notes as all the sound remains intact except meanings.

Can someone suggest how I should go about doing this.
I haven't actually decided which (free) host program I should be using as I'm a total newbie in this.
I'm not even sure I'm in the right forum!

Please help..

I'm guessing what you might want is a transient modulator ... I know Sony Oxford made a plugin that does that (called "Transient Modulator", no less), but I don't know of any free ones. You should be able to remove consonant sounds with that.

What you could do is to get a software (you can find a few free online) which convert audio to midi (If you dont know what midi is do a search).

This will give you a midi file which can be played back in windows as a piano (or whatever instrument you like).

Im not sure if its what you are looking for.

Do a google search for "audio to midi" for the software.

Hope it helps you out somewhat

Wow, that sounds extremely difficult. I'll have a play around with some free effects later and see if I can come up with anything, though someone will probably beat me to it.

Interesting application. You might want to check out a plug-in called Kantos from www.antarestech.com . It’s a voice-controlled synth that can extract pitch, dynamics, harmonic content and formant information from an audio file. There are a number of mp3 examples at the website.

If you don’t already have it you should also have a look at Praat ( www.praat.org ) which is a program for analysis, synthesis and manipulation of speech.

/Yoss

What you can do is download a trial version of Audiomulch http://www.audiomulch.com/

Not really easy although...
With sound in you can open your file. As sound effect you can open DigiGrunge or DLGranulator.
With sound out you can save your file.

Or this:
http://hem.passagen.se/rasmuse/Granny.htm

It is a standalone free granulator.
But you have to choose the right knob to tweak...

maybe it will be enough to simply lowpass filter the speech with a steep filter to get rid of the formants and all other spectrum-based information?

You might start with Spechstimme notation--it was developed by Arnold Schoenberg for "Pierrot Lunaire" around 1900. It would be probably serve you well.

Do a Goggle search on SPECHSTIMME.

You might start with Spechstimme notation--it was developed by Arnold Schoenberg for "Pierrot Lunaire" around 1900. It would be probably serve you well.

Do a Goggle search on SPECHSTIMME.

You might start with Spechstimme notation--it was developed by Arnold Schoenberg for "Pierrot Lunaire" around 1900. It would be probably serve you well.

Do a Goggle search on SPECHSTIMME.

Sorry about the multiple posts. Site wasn't responding normally.

bugs wrote:

Do a Goggle search on SPECHSTIMME.

Actually, "Sprechstimme." Interesting search, bugs. Thanks for mentioning this.

Personally, I agree with those who suggest using granular resynthesis to scramble the incoming audio. Some granular setups are very flexible, so you could choose just how strongly to alter the sound.

You have a fascinating research project -- best of luck!

Meffy

Sorry for the misspelling--its was early here.

bugs wrote:

Sorry for the misspelling--its was early here.

Ah, just playing. I knew what you meant, and that you knew your stuff too. Over here it's either early or late, depending on time-scale and criteria.

Meffy

What about using Melodyne to map out the pitch and rhythm of the speech, and then impose that on musical tones?

http://www.celemony.com/

Radical formant shifting might render apeach* unintelligible while retaining overall pitch and timing.
I can't think of such an effect but others may.

*edit: I should correct this to speech but it made me smile.

Melodyne works like a charm, you will not find anything that works that good with such simplicity.

Wow.. this is overwhelming.
I'm trying out your sugggestions one by one and it's taking time since I know nothing.. and have no idea which knob to turn.

So far, I haven't achieved much but I'm sure I'll get there in time and I'll let you guys know how it went. In the mean time, well thank you.

> Melodyne works like a charm, you will not find anything that works that good with such simplicity.

I second that. Especially with the new Version 2.5 it has become even easier.
You simply make your voice recording then let melodyne do an analysis. Finally You load a VST synthesizer into melodynes Mixer window and send the midi from the speech track to the synthesizer track.
It's a piece of cake. Yummy!