It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Turning a recorded whisper into something more audible, more like a speaking voice.

page: 1
2
<<   2 >>

log in

join
share:

posted on Nov, 4 2014 @ 07:54 PM
link   
I know there are some pretty clever folks on the ATS science and technology forum so I was wondering if it is possible to turn a whisper into a proper 'harder" more solid voice sound?

I guess that if it was possible, then it would be through the use of software filters but what kind of program would I need to do this? Probably Audacity would not do the business and it would need some extra special software.

Maybe a whisper in audio is like a compressed JPG photograph - if the detail is not there in the original, then it is impossible to recreate it or if it was recreated, then it would be a "best guess" at filling in the missing frequencies.

Any advice gratefully received. Thanks.




posted on Nov, 4 2014 @ 08:16 PM
link   
a reply to: qmantoo

I suppose it could be amplified quite a bit louder....then all the background noise (hiss) removed....and then use an equalizer with separate freq. bands to define it....I think with most software...even some simple ones...you could do this yourself.

1. Amplify it
2. Remove noise (background and hiss)
3. Round out the separate freq's to make it fuller

Good luck! Let us know!



posted on Nov, 4 2014 @ 08:19 PM
link   
www.screamingbee.com...

Maybe this can help you?



posted on Nov, 4 2014 @ 08:21 PM
link   
a reply to: qmantoo

I would say you are thinking correctly about it.
You would need some type of software that can manipulate digital audio. Perhaps standalone programs are available that have the necessary effects.
I use a DAW, or digital audio workstation, which has myriad effects and instruments, with a managing template wherein to assemble them.
In the case of your treating a sample of a whisper, I would do the following, somewhat as you were thinking.
I would use an equalizer to filter out the lowest frequencies entirely, something like below 500 hz. Also one could boost the sort of fuller frequencies of the whisper's waveform. Trim the static highs and "essey" frequencies...

Compression then is what will fill out the sound for you. Basically, the compressor is making the quietest parts of the whisper equally as loud as the parts of the whisper with highest volume.
These are key, but surely there are other processors your whisper could be finally run through.
One is another type of compressor called a limiter, which can equalize the volume of all components of your whisper as loudly as possible before being distorted.
When your little whisper is run through this very simple chain of effects, it is re-recorded: now it is of such audio quality that it could stand out even if mixed in with music.


edit on 4-11-2014 by ecapsretuo because: (no reason given)

edit on 4-11-2014 by ecapsretuo because: (no reason given)



posted on Nov, 4 2014 @ 08:22 PM
link   
I saw a video on how to push audible sound into a frequency that only digital recorders can pick up



posted on Nov, 4 2014 @ 08:40 PM
link   
OK, thanks for all your comments. So, opening up the topic slightly...

The application is EVP or electronic voice phenomena. For some reason the voices captured using this method always seem to me to be whispers and since they are whispers, there are various "interpretations" as to what they are saying.

Whatever is creating these noises whether it is ghosts or spirits (if you believe in them) or whatever appear to be unable to speak normally as we would do.

There are two ways of approacing this making EVP phenomena less "iffy". 1) design and create another way to record these 'voices' and 2) run the whispers through some kind of systems which makes them less ambiguous.

There is no reason at all why phenomena such as this should not be subjected to scientific study and if it were, maybe we would bring the paranormal out from under the bedclothes and more into the realms of science. Of course, it will never be repeatable and on-demand but at least it will possibly be explainable.
edit on 4 Nov 2014 by qmantoo because: (no reason given)



posted on Nov, 4 2014 @ 09:11 PM
link   
I've had EVP tell me "we got your back" and answering a question. And you should get large headphones and play the DNA healing sound YouTube video, place a digital recorder to your right ear between the speaker and ear, speaker angled forward,



posted on Nov, 4 2014 @ 09:29 PM
link   
a reply to: sanitizedinfo

They say that white noise which is randomized sound is supposed to allow these sounds to come through, however, some will say that it is our ears/brain playing tricks with us. A large seashell held close to the ear will sound as if it is white noise.

Maybe there is a way to filter out the random noise and leave the other sounds?



posted on Nov, 4 2014 @ 09:40 PM
link   
I recorded a male voice say my first name using the technique above.



posted on Nov, 5 2014 @ 01:07 AM
link   
 


off-topic post removed to prevent thread-drift


 



posted on Nov, 5 2014 @ 01:51 AM
link   
a reply to: qmantoo

Signal processing is signal processing.

'Special software' is often just taking more than one step and hiding its methods to pretend to be more advanced than other software. It normally isn't. What you're looking to do is a fairly basic function of most audio editors and there are plenty of tutorials. You could go back to basics and hit a program like MatLab up, use your own methods and maths but it probably won't help that much. The only real different between forensic audio clean up and film audio clean up is the forensic is less interested in how good it sounds, more interested in it being accurate and audible to a jury.

If you think of a graph of amplitude, the data is either there or not and you either mathematically predict it or you can't. For example, in an image I might be able to use averaging to remove snow or noise, but this has a strong logical basis. I might be able to predict missing parts of an image with reasonable success such as changing the geometry of a badly angled license plate. In audio you don't have many signifiers without an image to sync it with. What tells you what should come next.

I've never looked at it closely, but I also imagine a lot of the noise removal you would be doing would be near the same floor that is creating the words you're hearing. Any attempt to interpolate the noise is likely just adding convenient syllables. How do you truly know when a signal will go up or down within random noise? What would be convincing is if you could isolate part of the signal that repeatedly demonstrate EVP in a logical and predictable way. People have tried it before.

Messing with noise so that it sounds like something isn't the hard part, it's doing it in a way that makes logical sense. I have many books on signal processing if you would like reading materials, but it would be a large under taking to learn all that to be disappointed.



posted on Nov, 5 2014 @ 03:17 AM
link   
a reply to: qmantoo

you should look up spectral layers its a good program to single out frequency's and than use another program to amplify that's what i do.



posted on Nov, 6 2014 @ 12:10 AM
link   
This is an interesting problem from a computer science point of view. Whispers have a very different phonetic composition so I suspect that there is no easy way of transforming whispers into normal sounding speech. You'd probably have to create a voice recognition system specifically tailored to understand whispers. Then you could just use a text-to-speech program to convert the text output into computerized speech.
edit on 6/11/2014 by ChaoticOrder because: (no reason given)



posted on Nov, 6 2014 @ 12:47 AM
link   
a reply to: ChaoticOrder

I think the problem maybe far more fundemental than that because the EVP whispers are not generally easy to hear and it takes a trained ear to recognise the words. Probably similar to hearing and understanding a foreigner, after a while your ear gets used to their way of speaking and then you can hear what is being said easier than someone else who has just heard them for the first time.

Of course, making this more difficult, the EVPs could also be in a totally different language so the possibilities are endless.

Personally, I feel the way forward on this one is to try to make the EVPs clearer at the recording stage, but how to do that is another matter entirely.



posted on Jan, 25 2016 @ 03:24 AM
link   
OK...so I am relatively new to this (I mean computers, software, and the physical properties of sound,) and this question is a little off topic: Let's say, hypothetically, that I am personally hearing human whispers directed at me and these whispers sound exactly like someone I know, hypothetically of course. If I wanted to mask or completely cancel his whispers, which I realize are two entirely different processes, could one of you give me some ideas on how to do so, respectively, and if not maybe provide me with some helpful resources or links? I will attempt to outline my current plan of attack, or should I say defense...Preferably I would rather cancel than mask (keep in mind that some of my thinking /theory may be wrong). First: I would need to record a whisper that fell within the same frequency range (wondering if tone and intonation play a crucial role? Is a whisper considered a flat, non harmonic tone? ) of about 5 to 50 Hz. Second: either design a radio jammer or use a DAW, thanks Ecapsretuo, to broadcast this signal over the specified frequency range, but inversely (??? to cancel a sine wave I would need to overlap 3-4 of same recording 180 degrees out of phase to each other in a loop, but how do I determine how long the period should be???). If I used a spectrum tuner could I hone in on exact frequency and adjust accordingly? I would like this system to be portable so I am wondering whether a jammer (i.e. oscillators and/or circuits) or a compressed digital recording in mp3 format would be more practical? After all, both of these could be amplified at a later point to mask the original signal/sound/whisper...correct? Raising another question: Modulation. What type of program would be suitable and over how much area do I need to broadcast the carrier signal? In the case of the jammer, would I accomplish this with an external amplifier or by boosting battery power? If I can't determine exactly what this guy's going to say, how can I mimic the EM wave? Maybe masking is the better solution ??? Someone please help!!!!!!!



posted on Jan, 25 2016 @ 03:41 AM
link   

originally posted by: qmantoo

There is no reason at all why phenomena such as this should not be subjected to scientific study and if it were, maybe we would bring the paranormal out from under the bedclothes and more into the realms of science.


It has. It's bunk. Recording background noises so low to the noise floor as to be indistinguishable can fool our brains to hear auditory pareidolia. Even the mere suggestion of what "should" be heard makes people hear specific words, even though it's literally just noise.
edit on 25-1-2016 by GetHyped because: (no reason given)



posted on Jan, 25 2016 @ 03:49 AM
link   

originally posted by: jotclotFirst: I would need to record a whisper that fell within the same frequency range (wondering if tone and intonation play a crucial role?


Yes.


Is a whisper considered a flat, non harmonic tone? )


No.


of about 5 to 50 Hz.


No.


Second: either design a radio jammer or use a DAW, thanks Ecapsretuo, to broadcast this signal over the specified frequency range, but inversely


You're not going to successfully perform noise removal when the signal itself is so low below the noise floor.


(??? to cancel a sine wave I would need to overlap 3-4 of same recording 180 degrees out of phase to each other in a loop, but how do I determine how long the period should be???)


This will not work. In order to perform cancellation in this way, you need to polarity flip the SAME signal you wish to remove (and it has to be already isolated), not just another recording. But noise is, by definition, noise. You won't get the same noise floor each time you record.


If I used a spectrum tuner could I hone in on exact frequency and adjust accordingly?


No.



posted on Dec, 3 2017 @ 09:16 AM
link   
Hi all

Anyone to help me to get whispering more clearer, its just 40 sec long.

I have none so ever any skills in this area.

Anyone??

//Ullabett



posted on Dec, 3 2017 @ 09:31 AM
link   
a reply to: qmantoo

Who are you spying on buddy?

It's an honest question, but as to an answer for your own, there is software that can do as you require, freely available, simply by clicking a button, that it will do the work for you to varying degrees of amplification and clarity dependant on the source of course.

Edit: Maybe not "freely available" but available none the less.
edit on 3-12-2017 by andy06shake because: (no reason given)



posted on Dec, 3 2017 @ 10:56 AM
link   
I make music, dj and used to do a radio show and a decade of having to tidy up audio from interview plus indepth knowlege of sound analysis and wave manipulation as I have a degree in Applied Maths. Basically a geek.

It can be done using sound analysis programs like Izotope, it requires a bit of understanding about wave analysis and the human voice though. Background noise would be similar to white noise so I would run and inverted white noise wave to cancel out as much of that as possible. Then narrow the freqeuncy band to that of a human voice. Once that's done you'll be left with a bright orange/white wave with a blue/purple haze around it. Use a lasso tool to cut out the purple/blue and you'll be left with a clear sound. Obviously turn up the gain to boost volume.

With a bit of practice, following the above will produce broadcast quality sound. I've not got much on tomorrow if you want a hand editing it? Interested in hearing the audio as have spent so much time around mics I've developed an ear for what type of mic was used/what top end quality sounds like and when journalists are trying to bull# by claiming they set up the mic or used a passive.

A bi-directional mic, callibrated to the room/enivronment - plus ideally actively foccused on sound frequency band of observed phenomina would be my recomendation for the best microphone for the job. I'm friends with Joe Fossard and Tony Rigg, ex sound engineers from Ministry of Sound UK and proffesors/sound technician for University of Central Lancashire Sound Engineering course if ATS wants a definitive expert guide.


edit on 3-12-2017 by bastion because: (no reason given)

edit on 3-12-2017 by bastion because: (no reason given)




top topics



 
2
<<   2 >>

log in

join