WOW! M.I.T. Researchers can recreate sound from objects in the room

a reply to: neoholographic

You keep repeating this whole "micro vibrations" line when it's clear you don't understand the basic physics behind it all. It doesn't matter how you record the signal: any noise source of sufficient level occupying the frequency band you wish to record will mask the signal. Period. Again, there is nothing magical about the spectral band the human voice occupies.

It makes "no sense" to you because it would appear you do not understand the principles at play here.

a reply to: GetHyped

Again,

The algorithm recreates the sound as a whole through visual data. What you're not explaining is how you can mask sound waves that's being recreated as a WHOLE.

The researchers developed an algorithm that combines the output of the filters to infer the motions of an object as a whole when it’s struck by sound waves. Different edges of the object may be moving in different directions, so the algorithm first aligns all the measurements so that they won’t cancel each other out. And it gives greater weight to measurements made at very distinct edges — clear boundaries between different color values.

What motions are they talking about? Vibrations that are less than 100th of a pixel. The algorithm is recreating the sound as a whole. So each sound in the room creates it's own small vibration this is why DIFFERENT EDGES OF THE OBJECT MAY BE MOVING IN DIFFERENT DIRECTIONS.

How are you going to mask something this small by waving around a bag of chips? Do you understand what the algorithm is doing?

a reply to: neoholographic

slaps forehead

Assuming you're not trolling here, read this wiki article:

en.wikipedia.org...

a reply to: GetHyped

Sadly, you don't understand the algorithm.

If you're dealing with a bug that's capturing audio from the room then you can talk about auditory masking.

This is exactly why in the video they demonstrated this technology through soundproof glass. How can auditory masking affect the recreation of sound through visual data when you're talking about vibrations less than 100th of a pixel?

So you can slap your forehead all day but you haven't explained why a bug that depends on picking up audio from the room is the same as recreating sound through vibrations that are less than 100th of a pixel.

This is why I asked you do you understand the algorithm.

originally posted by: GetHyped
a reply to: neoholographic

slaps forehead

Assuming you're not trolling here, read this wiki article:

en.wikipedia.org...

It is guaranteed to be trolling by this stage. I'm considering stepping back in purely for the amusement value. There is certainly no other value left in this debate.

a reply to: neoholographic

I went to Abe's web site and sent an email and linked this thread. The source code will be available soon according to his site so we can test your magical claims.

a reply to: ZetaRediculian

What magical claims?

The fact that you keep saying this shows how weak your argument is.

What's magical about frames per second and running those frames through an algorithm? You're the one that looks childish because you made the claim but haven't been able to articulate why you support it. I think you just read evil Bob's post then became a Johnny come lately and you have no clue as to what you're talking about.

originally posted by: neoholographic

I think you just read evil Bob's post then became a Johnny come lately and you have no clue as to what you're talking about.

Well there certainly seems to be some confusion in this thread - and as we've all already read the articles, clearly the articles aren't stating the matter in a way that everyone can understand consistently.

I'd invited you previously to help us understand this from your perspective but you ignored that invitation. I'm willing to extend it again.

If I offered some questions, would you be prepared to explain the answers to us, so that we can understand why we seem to be reaching these differing views?

a reply to: neoholographic

What magical claims?

the ones coming out your butt. Do you think the guys from MIT can come up with an algorithm to decipher your butt vibrations

originally posted by: EvillerBob
It is guaranteed to be trolling by this stage. I'm considering stepping back in purely for the amusement value. There is certainly no other value left in this debate.

Well, here is Abe's web site. people.csail.mit.edu... His was the first on the list of names so it seems like he is the lead guy. Its really a shame that such good topics have to be dominated by such idiocy.
Here is Super Sexy himself

originally posted by: ZetaRediculian

That's actually quite mindblowing. I'm observing the invisible microvibrations of his lips, and it's almost like I can actually hear the music. That algorithm really is something amazing. Thank God he didn't have a packet of crisps, it would have ruined the entire experience

a reply to: neoholographic

The best way to get around this technology is to write out what you're doing if you think you're being watched then burn whatever you wrote down after the meeting is over.

Um, there is a giant camera outside your soundproof glass door. You could use the paper to cover up the lens.

Pretty sure Neo is trolling, but on the off chance he isn't:

I'll make it short and easy to follow:

You ask "how can you make vibrations so small, etc etc" well quite easily. That's how this technology works. SOUND makes tiny vibrations in a material (chip bag) then the program compares what the vibrations look like, to what it should sound like, and recreates the sound. If you watch a video of someone clapping, with no sound, you could know "hey, there is a clapping sound" and you could play a clapping sound while watching the video. That's exactly what this technology is doing, just a far more complex version, instead of creating whole sounds like clapping it creates frequencies which combine together to make sounds we recognize.

Now, we know the human voice is capable of causing these tiny vibrations, correct? I mean, if we could NOT make these vibrations with our voices, then the technology wouldn't be able to pick them up. So tell me, what makes our voices special? Why do our voices cause tiny vibrations in the chip bag, but a radio does not? An air conditioner does not? A TV does not? A washer and dryer does not? A dishwasher does not? Why is the human voice so special that it makes these tiny vibrations but nothing else out there can make the same vibrations in the bag?

Because that's what your saying. That the human voice is something special that causes unique vibrations in the bag. That's the ONLY way you couldn't beat this thing, if the human voice made UNIQUE vibrations. Otherwise, other things could make the same vibrations, and there would be no way to tell which vibrations were coming from a voice, and which were coming from something else.

Unless of course, you think the human voice induces unique vibrations into a chip bag that are not like any other vibrations, which is just false and easily proven to be so. The human voice is simply made up of different frequencies, ALL of which can be reproduced via other means.

a reply to: James1982

But...but the frames per second and the algorithm and the sound proof glass door!

a reply to: neoholographic

What's magical about frames per second and running those frames through an algorithm? You're the one that looks childish because you made the claim but haven't been able to articulate why you support it. I think you just read evil Bob's post then became a Johnny come lately and you have no clue as to what you're talking about.

let me explain this in terms of how the real world works. If my boss asks me how I solved something with code I developed, I can provide him with every technical explanation I can muster. It doesn't matter if it doesn't work. Thats it. I'm done if it doesn't work. What you are talking about was not demonstrated. Every demonstration of this is of it working when something is completely still. No technical explanation matters if you can't demonstrate this doing what you are claiming. I'm claiming it does EXACTLY what is shown. Partial recovery of sound from objects in a video that are completely still. You are claiming it does more. Obviously nobody is going to convince you otherwise. So when the source code is provided, you can demonstrate how wrong we all are. Thats it. End of discussion.

originally posted by: ZetaRediculian

Every demonstration of this is of it working when something is completely still.

While I agree in principle with everything you say, I think it's worth expanding this statement a little bit, because it might be at the core of the disagreement.

When we say "still", we mean that there are no other forces present to make it move - in other words, no breezes, no people walking past, no other background/unexpected audio, etc. The only thing affecting the object is the sound waves from the test-specific audio, so the movement detected in the object can only be attributed to that audio.

a reply to: James1982

Unfortunately for you, you're making the same silly mistake as the other 2 guys. You're trying to debate against a claim that was never made.

This clearly shows that you guys have no clue as to what you're talking about.

Whenever a person has to try and debate against things that were never said it shows how sad and ignorant their position is. It's just like evil bob and the other person trying to debate magic or invisible waves when nobody mentioned anything about magic waves or these things being invisible.

So you guys haven't answered any question, explained any position and you spend these SILLY POSTS DEBATING AGAINST THINGS THAT WERE NEVER SAID.

You said:

Now, we know the human voice is capable of causing these tiny vibrations, correct? I mean, if we could NOT make these vibrations with our voices, then the technology wouldn't be able to pick them up. So tell me, what makes our voices special? Why do our voices cause tiny vibrations in the chip bag, but a radio does not? An air conditioner does not? A TV does not? A washer and dryer does not? A dishwasher does not? Why is the human voice so special that it makes these tiny vibrations but nothing else out there can make the same vibrations in the bag?

Because that's what your saying. That the human voice is something special that causes unique vibrations in the bag. That's the ONLY way you couldn't beat this thing, if the human voice made UNIQUE vibrations. Otherwise, other things could make the same vibrations, and there would be no way to tell which vibrations were coming from a voice, and which were coming from something else.

WHAT???????

When did I make the claim that the human voice was unique??????

WHERE DID I SAY YOU CAN'T PICK UP OTHER SOUNDS IN THE ROOM???

It's just disgusting that you guys keep trying to debate against things that I never said!! You spent almost the entire idiotic post debating against a claim that I never made.

On this thread I have constantly talked about how the sound from the room is recreated as a WHOLE!! Do you know what as a whole means??? I have emphasised this and even quoted from the original article where the same thing was discussed.

The researchers developed an algorithm that combines the output of the filters to infer the motions of an object as a whole when it’s struck by sound waves. Different edges of the object may be moving in different directions, so the algorithm first aligns all the measurements so that they won’t cancel each other out. And it gives greater weight to measurements made at very distinct edges — clear boundaries between different color values.

NOT ONCE, have I said the human voice is unique and you can't pick up other sounds. That's simply something MADE UP BY YOU just like other posters made long posts debating against magic and invisible waves when this claim was never made. You do this because YOU CAN'T DEBATE AGAINST WHAT I ACTUALLY SAID.

I said it will be harder to mask the audio because you're dealing with vibrations that are less than 100th of a pixel that gets better as you increase the frames per second. So it will not be like picking up sound with a bug that's in the room that depends strictly on audio data. This technology depends on visual data as well and that's why they did the experiment through soundproof glass. This is because they're recreating sound through visual data.

THIS IS WHAT I SAID!!

I HAVEN'T SAID ANYTHING ABOUT THE HUMAN VOICE BEING UNIQUE OR THAT YOU CAN'T PICK UP OTHER SOUNDS IN THE ROOM.

You guys show how weak your argument is when you have to spend these silly posts debating against claims that were never made.

a reply to: EvillerBob

While I agree in principle with everything you say, I think it's worth expanding this statement a little bit, because it might be at the core of the disagreement.

When we say "still", we mean that there are no other forces present to make it move - in other words, no breezes, no people walking past, no other background/unexpected audio, etc. The only thing affecting the object is the sound waves from the test-specific audio, so the movement detected in the object can only be attributed to that audio.

I believe so. They didn't specify in the documentation if they isolated the vibration from other movements like breezes. Not that I saw. I think that there will always be some noise. And I think that there will always be some filtering and isolation of the signal. If you follow the links, you will see some more technical documentation. Going by the video, everything is relatively undisturbed. That is consistent with the other project that this is related to. That is the video magnification one that works by amplifying the pixels in the video which allows you to see stuff you wouldn't normally see. Like a baby breathing, pulses, etc. For that to work, the subject needs to be still. The movements that are being detected here are even smaller and even more easily disrupted so I have a hard time understanding how this would not be an issue.
another consideration is isolating the pixels of the object from the background. If the background is a similar color, there could be noise there too.

So I believe what we are seeing in the videos are the best conditions in order for this to work. That it works at all is truly remarkable. Also in the documentation they talk about how they were sometimes not able to recover any intelligible information. They did many more experiments then what they showed and were only able to hear garbled voices in some cases. Still impressive.

really would like to get a hold of the source code and really dig into it. First I want to get my baby monitor working with the other project. Of course I could be wrong thats why us dumb guys need to play with the source code.

a reply to: neoholographic

Honestly your histrionics are meaningless, though entertaining.

there really is no debate. If it works, technical babble is meaningless. Show that it works the way you say it does and be done with it. Thats all.

originally posted by: ZetaRediculian
a reply to: neoholographic

What's magical about frames per second and running those frames through an algorithm? You're the one that looks childish because you made the claim but haven't been able to articulate why you support it. I think you just read evil Bob's post then became a Johnny come lately and you have no clue as to what you're talking about.

let me explain this in terms of how the real world works. If my boss asks me how I solved something with code I developed, I can provide him with every technical explanation I can muster. It doesn't matter if it doesn't work. Thats it. I'm done if it doesn't work. What you are talking about was not demonstrated. Every demonstration of this is of it working when something is completely still. No technical explanation matters if you can't demonstrate this doing what you are claiming. I'm claiming it does EXACTLY what is shown. Partial recovery of sound from objects in a video that are completely still. You are claiming it does more. Obviously nobody is going to convince you otherwise. So when the source code is provided, you can demonstrate how wrong we all are. Thats it. End of discussion.

Just an ignorant statement.

The fact is you don't understand what you're talking about.

IT'S NOT COMPLETELY STILL AND THAT'S THE POINT.

You simply debate an issue when you're ignorant about the issue and you make these meaningless points.

The fact is there are VIBRATIONS that can't be seen by the naked eye because they're so small. VIBRATIONS DOESN'T MEAN COMPLETELY STILL.

Based on these small vibrations sound is recreated from the room. The fact that you think the bag is completely still shows that you don't have a clue.

So a bag being waved around can still be caught in frame unless you're moving you're hand at Superman speeds.

WOW!! COMPLETELY STILL LOL!!!