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

a reply to: EvillerBob

I answered your question and you should realize that I have been debating on ATS for awhile and the people making these posts are the same people I have debated before and in most cases they don't understand what they're talking about.

It's not my fault that you and others can't understand that the sound is being recreated by visual data. In order to disrupt this technology you would have to distort the visual data. This is why they talk about several things.

Reconstructing audio from video requires that the frequency of the video samples — the number of frames of video captured per second — be higher than the frequency of the audio signal. In some of their experiments, the researchers used a high-speed camera that captured 2,000 to 6,000 frames per second. That’s much faster than the 60 frames per second possible with some smartphones, but well below the frame rates of the best commercial high-speed cameras, which can top 100,000 frames per second.

Forget it, I'm not going through all this again.

I keep backing up what I'm saying based on the actual technology and you guys keep on avoiding what they're saying.

You guys have talked about everything from magical to invisible waves, waving a bag of chips and other things that have nothing to do with the technology.

It's harder to mask the audio because the sound is being recreated through images that are captured in frame. You have to move fast enough to distort the visual that's used to recreate the sound.

It's really that simple.

originally posted by: neoholographic
a reply to: EvillerBob

It's harder to mask the audio because the sound is being recreated through images that are captured in frame. You have to move fast enough to distort the visual that's used to recreate the sound.

And what exactly is "the audio"? Explain how the waveform of the human voice is any more or less difficult to mask using this method compared to any other waveform.

Edit:

At any position you wish to record, no matter what the recording method is, you are recording the SUM of all audible audio sources from that position because, you know, physics. If the amplitude of the audio sources you do not wish to record ("noise") is sufficiently greater than the audio source you wish to record ("signal"), the masking of the signal will be too great to get any useful data from it.

I've posted wiki articles. I've posted videos. At least 3 or 4 other posters have reiterated exactly the same thing, yet you claim with hubris that it is everyone else who is misunderstanding this here, yet it is clear to anyone with even half a grasp of the principles at play here that is is you who has the misunderstanding.

Not only that, but none of your claims regarding masking are supported by the vid in your own OP.

It's not my fault that you and others can't understand that the sound is being recreated by visual data. In order to disrupt this technology you would have to distort the visual data. This is why they talk about several things.

"Reconstructing audio from video requires that the frequency of the video samples — the number of frames of video captured per second — be higher than the frequency of the audio signal. In some of their experiments, the researchers used a high-speed camera that captured 2,000 to 6,000 frames per second. That’s much faster than the 60 frames per second possible with some smartphones, but well below the frame rates of the best commercial"

We get it. The amplitude and frequency of the vibrations of the plant or bag of chips are being read visually, and that visual data is then translated by the software, then the software recreates electrical impulses to emulate the vibrations that were visually read.

...Am I right so far? I think I am. All that is great and the people who did this are very clever. We can all agree with that -- right?

HOWEVER, where this thread is breaking down is the idea that you have that somehow the bag of chips or the plant are somehow immune to picking up unwanted vibrations. Those unwanted vibrations could came from playing a stereo loudly in the room I am in (the room with the plant or bag of chips). If the stereo was playing while someone was speaking "Mary had a little lamb", this video camera would pick up BOTH the vibrations caused by the music and by the voice speaking the poem....

The video camera and software would not inherently know the difference between the signal (the poem) and the noise (the music on the stereo) -- at least not simply by the visual data. That's what we mean by signal-to-noise ratio. If the noise is loud enough, it would mask the signal.

The bottom line is this:
The 100th of a pixel vibrations cause by the noise (music) adds to the 100th of a pixel vibrations cause by the signal (poem), thus the vibration would be a jumble of music over top of an unintelligible voice.

a reply to: GetHyped

Simple,

“When sound hits an object, it causes the object to vibrate,” says Abe Davis, a graduate student in electrical engineering and computer science at MIT and first author on the new paper. “The motion of this vibration creates a very subtle visual signal that’s usually invisible to the naked eye. People didn’t realize that this information was there.”

These vibrations are a visual signal. So the sound is being recreated through visual data and this is different then audio recorded by a bug in a room.

This is why they showed you how this technology works through soundproof glass.

It's really quite simple.

a reply to: Soylent Green Is People

Or, an entirely new sound all together.

There is more than one building that 'sings' due to this phenomenon.

a reply to: Soylent Green Is People

You said:

HOWEVER, where this thread is breaking down is the idea that you have that somehow the bag of chips or the plant are somehow immune to picking up unwanted vibrations.

What are unwanted vibrations? Show me in the article on this technology where they talked about unwanted vibrations.

Again, you don't understand what you're talking about.

All of these vibrations are WANTED because these small vibrations are being used by the algorithm to recreate the sound from the room. So what are unwanted vibrations? How are these unwanted vibrations different from the vibrations being used to recreate the sound from the room?

Ahhhh...big deal. Want to impress me? Show me technology to recover video from audio

a reply to: neoholographic

There is nothing in that quote that supports your claim.

"The sound" will be the SUM of all audible sources at that position.

originally posted by: neoholographic
What are unwanted vibrations? Show me in the article on this technology where they talked about unwanted vibrations.

Unwanted vibrations will be all of the non-voice caused vibrations that will exist out side of the controlled environment.

Everything from people moving to the changes in the building envelope as the sun moves through the sky.

It is one thing to produce audio from a visual signal in a controlled environment, it is a completely different animal to do the same in an open environment, and another completely different animal to do it in a secure environment.

originally posted by: neoholographic


What are unwanted vibrations? Show me in the article on this technology where they talked about unwanted vibrations.

Neo, we have ALL told you MULTIPLE times EXACTLY what these unwanted vibrations are. SERIOUSLY?

The unwanted vibrations are AIR CONDITIONERS, MUSIC PLAYING ON THE RADIO, YOU WALKING AROUND IN THE ROOM, THE HOUSE SETTLING, AN AIRPLANE GOING OVER HEAD, A CAR DRIVING BY OUT FRONT, THE GARBAGE MAN PICKING UP, YOUR FRONT DOORBELL RINGING, YOUR KEYBOARD CLICKING AWAY, A NEST OF BEES IN YOUR OVERHANG.

In a REAL WORLD situation there are TONS of UNWANTED vibrations that will mask the vibrations made by your voice.

The article does not mention unwanted vibrations BECAUSE THEY REMOVED THEM. The "sample" is INSIDE A SOUNDPROOF BOX WITH NO OTHER SOURCE OF SOUND INSIDE. They did this to eliminate all unwanted vibration that could interfere with their experiment, because they know that "atmospheric" sounds will create vibrations that make this method impossible, or at least very hard.

Now I know it's a troll, you are asking questions WE ANSWERED FOR YOU TWO PAGES AGO.

a reply to: neoholographic

What are unwanted vibrations? Show me in the article on this technology where they talked about unwanted vibrations.

they don't because they eliminated them by doing their experiments indoors behind glass doors. They did not discuss any of that because it wasn't part of what they were showing. They showed exactly what they said they could do. The rest is in your head

a reply to: peck420

This is just nonsense.

Unwanted vibrations will be all of the non-voice caused vibrations that will exist out side of the controlled environment.

All of these vibrations are the same and are used to recreate the sound in the room. There are no unwanted vibrations. That's just kooky talk. These vibrations are less than 100th of a pixel and these vibrations are used to recreate sound through visual data.

Sow me in the article where they talk about unwanted vibrations. Again, that's just silly talk.

Again, just show me in the video or article where they talk about unwanted vibrations.

They don't.

First you guys talk about magic waves now you have skipped to the kooky talk of unwanted vibrations.

originally posted by: neoholographic
a reply to: Soylent Green Is People

You said:

HOWEVER, where this thread is breaking down is the idea that you have that somehow the bag of chips or the plant are somehow immune to picking up unwanted vibrations.

What are unwanted vibrations? Show me in the article on this technology where they talked about unwanted vibrations.

Again, you don't understand what you're talking about.

All of these vibrations are WANTED because these small vibrations are being used by the algorithm to recreate the sound from the room. So what are unwanted vibrations? How are these unwanted vibrations different from the vibrations being used to recreate the sound from the room?

Wait. I thought you implied somewhere back a few pages that [paraphrasing what you implied] this works as a method of spying because it is somehow immune to conversation masking (e.g., by turning on a stereo). Am I wrong about that? If I am wrong about that, then what point exactly are you making. We all have already said that what these people did was very clever...

...BUT we also said that the limitation of this is that unwanted noise will be picked up by the plant or bag, and that unwanted noise will be visually read by the video camera and software -- and that video camera and software would not be able to separate the part of the vibration that is the intended target conversation from the part of the vibration that is the unwanted noise.

I know that the article did not mention unwanted noise -- and that is the issue some of us have with this. We want to know how these people intend to separate the part of the vibration that is unwanted from the part of the vibration that is wanted.

It's not like voice soundwaves vibrate the plant/bag at a higher frequency than other soundwaves. All of those soundwaves will combine into one jumble vibration that will be picked up visually by the camera.

a reply to: James1982

You said:

The "sample" is INSIDE A SOUNDPROOF BOX WITH NO OTHER SOURCE OF SOUND INSIDE.

AGAIN, MORE NONSENSE!!

They taped the video through soundproof glass in one of the experiments. What in the world are you talking about?

I'll be back later and maybe we will get a quote from the article or from the video talking about unwanted vibrations LOL.

originally posted by: neoholographic
Again, just show me in the video or article where they talk about unwanted vibrations.

They don't.

First you guys talk about magic waves now you have skipped to the kooky talk of unwanted vibrations.

This is because it's basic physics/acoustics. There is nothing in the video about some magical anti-masking effect. This is a fabrication of your own misunderstanding.

originally posted by: neoholographic


I'll be back later and maybe we will get a quote from the article or from the video talking about unwanted vibrations LOL.

a reply to: ZetaRediculian

/end thread

Edit:

The object is decoupled from the surface to minimize unwanted vibrations.

The speaker is decoupled from the surface to minimize unwanted vibrations.

This is evidently an artificial setup to demonstrate the technology here which has nothing to do with OP's fabrications about being some fancy spy technology that is immune to masking. Throw in some additional audio sources and additional structural vibrations in a real-world scenario and you've got MORE unwanted vibrations.

It's pretty damn clear that, unsurprisingly, the recording can be subject to noise just like any other recording technology. But that's fine, that's not what this technology is about, only inside OP's head is it about this. What they have achieved is cool and impressive enough without having to make up spurious claims about it.

a reply to: neoholographic
You really should have read the actual report MIT released...not just the hyped up blog about it...

From the MIT report:

On our project web page we show another example where we recover music well enough for some listeners to recognize the song, though the lyrics themselves are unintelligible in the recovered sound.

Even in highly controlled environments, they can not distinguish between music and voice.