OK, so some people around here work with LCDs...

We've got a laptop at work with a cracked screen. Around where the crack 
is, the pixels are all black. Which isn't surprising. Slightly more 
interesting is the red vertical lines that seem to extend vertically 
outwards from the cracks. Why red? Why every other pixel? Why vertical? 
I have no idea.

What's weirder still is that close to the cracks, the dead pixels 
exhibit a curiously fractal pattern. It looks like some kind of weird 
pixellated plant growth or something. Does anybody have any idea why?

(I've got photographs, but I'd have to find something that has Bluetooth 
in order to get them off my phone...)

Post a reply to this message

On 8/7/2015 6:05 PM, Orchid Win7 v1 wrote:
> I've got photographs

Unless we see them. It is not true. ;-)


-- 

Regards
     Stephen

Post a reply to this message

Orchid Win7 v1 <voi### [at] devnull> wrote:
> What's weirder still is that close to the cracks, the dead pixels
> exhibit a curiously fractal pattern. It looks like some kind of weird
> pixellated plant growth or something. Does anybody have any idea why?

No idea. But it reminds me of what I saw last year on somebody's water damaged
plasma TV. The unit worked fine, except for a localized black dendritic growth
that looked a lot like a cellular automaton. The guy said it changed and moved
across the screen over time... very strange (but cool).

> (I've got photographs, but I'd have to find something that has Bluetooth
> in order to get them off my phone...)

Can you at least upload them to this site? http://postimage.org/

Post a reply to this message

Am 07.08.2015 um 19:05 schrieb Orchid Win7 v1:

> What's weirder still is that close to the cracks, the dead pixels
> exhibit a curiously fractal pattern. It looks like some kind of weird
> pixellated plant growth or something. Does anybody have any idea why?

Because the display is now a hardware implementation of a cellular 
automaton?

Post a reply to this message

Orchid Win7 v1 <voi### [at] devnull> wrote:
> OK, so some people around here work with LCDs...
>
> We've got a laptop at work with a cracked screen. Around where the crack
> is, the pixels are all black. Which isn't surprising. Slightly more
> interesting is the red vertical lines that seem to extend vertically
> outwards from the cracks. Why red? Why every other pixel? Why vertical?
> I have no idea.
>
> What's weirder still is that close to the cracks, the dead pixels
> exhibit a curiously fractal pattern. It looks like some kind of weird
> pixellated plant growth or something. Does anybody have any idea why?
>
> (I've got photographs, but I'd have to find something that has Bluetooth
> in order to get them off my phone...)

At first, I was thinking it could be a moire effect, but...

"LCD stands for "liquid crystal display" and the liquid crystal inside a screen
is subject to different levels of current to change the amount of the backlight
that passes through. When a screen is cracked, this process breaks down and
incredible patterns emerge, as some sections of the goo is electrified and some
isn't." -- Wired UK

Fractal forums had something about it possibly being a form of Diffusion-Limited
Aggregation.  Something like a Lichtenberg Figure, maybe?

Diffusion-Limited Aggregation is a process whereby, particles undergoing a
random walk via Brownian motion, tend to cluster together.

It would make sense that the crack in the screen disrupts the even distribution
of electrons through the liquid crystal, and the result is that the distribution
becomes chaotic and the electrons go where other electrons have been.

I have no technical expertise with electronics, so this is the best idea I can
come up with.  I have nothing on the red lines.

Regards,
A.D.B.

Post a reply to this message

"Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> Fractal forums had something about it possibly being a form of Diffusion-Limited
> Aggregation.

http://www.fractalforums.com/fractals-in-nature/broken-lcd-screen/

Post a reply to this message

On Fri, 07 Aug 2015 19:05:08 +0200, Orchid Win7 v1 <voi### [at] devnull> wrote:

> OK, so some people around here work with LCDs...
>
> We've got a laptop at work with a cracked screen. Around where the crack  
> is, the pixels are all black. Which isn't surprising. Slightly more  
> interesting is the red vertical lines that seem to extend vertically  
> outwards from the cracks. Why red? Why every other pixel? Why vertical?  
> I have no idea.
>
> What's weirder still is that close to the cracks, the dead pixels  
> exhibit a curiously fractal pattern. It looks like some kind of weird  
> pixellated plant growth or something. Does anybody have any idea why?
>
> (I've got photographs, but I'd have to find something that has Bluetooth  
> in order to get them off my phone...)

Our laptop had vertical multicoloured lines and not much fractal beauty.

-- 
-Nekar Xenos-

Post a reply to this message

On Sat, 08 Aug 2015 05:09:58 +0200, Samuel Benge <stb### [at] hotmailcom>  
wrote:

> fractalforums

After scrolling down, the next two images show these wonderful leaf-like  
fractals.
Does anyone know the algorithm for these? I would really like to be able  
to use it off-off-topic :)

-- 
-Nekar Xenos-

Post a reply to this message

"Nekar Xenos" <nek### [at] gmailcom> wrote:
> On Sat, 08 Aug 2015 05:09:58 +0200, Samuel Benge <stb### [at] hotmailcom>
> wrote:
>
> > fractalforums
>
> After scrolling down, the next two images show these wonderful leaf-like
> fractals.
> Does anyone know the algorithm for these? I would really like to be able
> to use it off-off-topic :)

I know of at least two ways.

The first method has a seed layer (2D or 3D array), and a 1D array of particle
positions. Each particle moves around randomly (usually via Brownian motion)
until it detects a positive value on the seed layer. At this point it dies and a
value is added to the seed layer. Another particle is then introduced from some
convenient location, such as a screen edge. Since a particle rarely slips
between pixels before making contact, branching growth occurs.

The second method uses an additional 2D or 3D layer representing a solution
(instead of an array of particles). Every cell is evaluated at every frame. The
solution layer is 'saturated' at initialization. If the solution layer has
material to expend and a neighboring cell of the other layer has a positive
value, then material is added to that other layer and material is subtracted
from the solution. The solution is blurred each frame, which simulates diffusion
(albeit with some loss).

I hope that made sense :/ Maybe I'll get something going and upload it to
glslsandbox.

Post a reply to this message

"Samuel Benge" <stb### [at] hotmailcom> wrote:
> Maybe I'll get something going and upload it to glslsandbox.

OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0

You'll have to place the mouse in the lower left corner to properly initialize
the layers (goofy mistake). Seeds are automatically placed under the mouse
cursor.

The effect tends to be large-scale compared with other diffusion-limited
aggregation setups, but it can be tweaked by changing the settings.

Post a reply to this message

"Samuel Benge" <stb### [at] hotmailcom> wrote:
> "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > Maybe I'll get something going and upload it to glslsandbox.
>
> OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0
>
> You'll have to place the mouse in the lower left corner to properly initialize
> the layers (goofy mistake). Seeds are automatically placed under the mouse
> cursor.
>
> The effect tends to be large-scale compared with other diffusion-limited
> aggregation setups, but it can be tweaked by changing the settings.

neat. But is there a way to add constructive interference?  It could make a
great simulator for frost, if you could do that.

Regards,
A.D.B.

Post a reply to this message

"Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > > Maybe I'll get something going and upload it to glslsandbox.
> >
> > OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0
> >
> > You'll have to place the mouse in the lower left corner to properly initialize
> > the layers (goofy mistake). Seeds are automatically placed under the mouse
> > cursor.
> >
> > The effect tends to be large-scale compared with other diffusion-limited
> > aggregation setups, but it can be tweaked by changing the settings.
>
> neat. But is there a way to add constructive interference?  It could make a
> great simulator for frost, if you could do that.

I was thinking about another version that adds a water tank/ripples effect to
the solution... is that what you mean? It would be interesting to see how the
solid layer builds faster in areas of the solution where it's concentrated due
to CI. Hm...

Post a reply to this message

"Samuel Benge" <stb### [at] hotmailcom> wrote:
> "Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> > "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > > "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > > > Maybe I'll get something going and upload it to glslsandbox.
> > >
> > > OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0
> > >
> > > You'll have to place the mouse in the lower left corner to properly initialize
> > > the layers (goofy mistake). Seeds are automatically placed under the mouse
> > > cursor.
> > >
> > > The effect tends to be large-scale compared with other diffusion-limited
> > > aggregation setups, but it can be tweaked by changing the settings.
> >
> > neat. But is there a way to add constructive interference?  It could make a
> > great simulator for frost, if you could do that.
>
> I was thinking about another version that adds a water tank/ripples effect to
> the solution... is that what you mean? It would be interesting to see how the
> solid layer builds faster in areas of the solution where it's concentrated due
> to CI. Hm...

mostly, I was thinking that the end result would look less like a voronoi
diagram.

to have it really look like frost, it would have to grow in spikes that get
wider as they get longer.  I have no idea how you'd do that...

Regards,
A.D.B.

Post a reply to this message

On Sun, 09 Aug 2015 00:49:03 +0200, Samuel Benge <stb### [at] hotmailcom>  
wrote:

> "Samuel Benge" <stb### [at] hotmailcom> wrote:
>> Maybe I'll get something going and upload it to glslsandbox.
>
> OK, here's something if anyone's interested:  
> http://glslsandbox.com/e#27058.0
>
> You'll have to place the mouse in the lower left corner to properly  
> initialize
> the layers (goofy mistake). Seeds are automatically placed under the  
> mouse
> cursor.
>
> The effect tends to be large-scale compared with other diffusion-limited
> aggregation setups, but it can be tweaked by changing the settings.
>
>
Beautiful.

It would be nice to have as a Pov-Ray texture.

-- 
-Nekar Xenos-

Post a reply to this message

"Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> "Samuel Benge" <stb### [at] hotmailcom> wrote:
> > I was thinking about another version that adds a water tank/ripples effect to
> > the solution... is that what you mean? It would be interesting to see how the
> > solid layer builds faster in areas of the solution where it's concentrated due
> > to CI. Hm...
>
> mostly, I was thinking that the end result would look less like a voronoi
> diagram.

It could be done, of course. Not yet sure how I'd go about it though...

> to have it really look like frost, it would have to grow in spikes that get
> wider as they get longer.  I have no idea how you'd do that...

Either reduce the amount of solution consumed, increase the growth rate, or do
both.

And for water crystals, it helps if the system uses a neighborhood that tests
cells six directions from center:

..oo
ooo
oo.

The result has a six-fold symmetry but ends up skewed, so to look good the final
image would need to be transformed.

Post a reply to this message

"Nekar Xenos" <nek### [at] gmailcom> wrote:
> On Sun, 09 Aug 2015 00:49:03 +0200, Samuel Benge <stb### [at] hotmailcom>
> wrote:
> >
> > OK, here's something if anyone's interested:
> > http://glslsandbox.com/e#27058.0
> >
> > You'll have to place the mouse in the lower left corner to properly
> > initialize
> > the layers (goofy mistake). Seeds are automatically placed under the
> > mouse
> > cursor.
> >
> > The effect tends to be large-scale compared with other diffusion-limited
> > aggregation setups, but it can be tweaked by changing the settings.
> >
> Beautiful.

Thanks.

> It would be nice to have as a Pov-Ray texture.

You could always take a screenshot, crop out the window frame, and remove the
blue channel.

Alternatively, it can all be done procedurally in POV, if you're prepared to
render many, many frames, or endure an hours-long parse time :P

Post a reply to this message

> OK, so some people around here work with LCDs...
>
> We've got a laptop at work with a cracked screen. Around where the crack
> is, the pixels are all black. Which isn't surprising. Slightly more
> interesting is the red vertical lines that seem to extend vertically
> outwards from the cracks. Why red? Why every other pixel? Why vertical?
> I have no idea.

It's more than likely due to the column drivers malfunctioning. One 
possible column driver scheme is to use only one DAC for each *pixel* 
column (not sub-pixel) and then multiplex the output between RGB 
subpixels. It saves by a factor of 3 the number of DACs needed and the 
number of wires to/on the panel. What is likely happening is that when 
each row is getting refreshed, the driver is reset as data is clocked 
in, red is likely first and it gets screwed (because all the column 
tracks are broken at the crack so the wrong capacitance and resistance 
or possibly even shorted) and it never manages to switch the output to 
the green and blue columns.

Another possibility, if the drivers are on the glass substrate itself, 
or a bare IC bonded to the glass, is that they got damaged when the 
crack was made.

> What's weirder still is that close to the cracks, the dead pixels
> exhibit a curiously fractal pattern. It looks like some kind of weird
> pixellated plant growth or something. Does anybody have any idea why?

The "growth" bit is because gradually moisture is getting in to the LC 
material through the crack. They are normally completed sealed in, I'm 
not a chemist but I know moisture+LC is bad and gives those patterns 
even without a crack. Also does the pattern move when you press around 
the crack? It's likely the cell-gap (distance between the inner two 
faces of glass) has changed due to the crack, that normally needs to be 
very accurately controlled to create a uniform image.

Post a reply to this message

On 08/08/2015 11:49 PM, Samuel Benge wrote:
> OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0

That's just trippy. Not quite as intricate as some of the 
reaction-diffusion systems I've seen, but mental none the less...

Post a reply to this message

Orchid Win7 v1 <voi### [at] devnull> wrote:
> On 08/08/2015 11:49 PM, Samuel Benge wrote:
> > OK, here's something if anyone's interested: http://glslsandbox.com/e#27058.0
>
> That's just trippy. Not quite as intricate as some of the
> reaction-diffusion systems I've seen, but mental none the less...

Freudian slip? It /is/ probably more of a reaction-diffusion system than it is a
traditional DLA ;)

More detail can be added by making the following line changes:

line 28: float solSub = 10.0;
line 92: R = 1.1 * Sol;

It looks much darker, though. Also, the resolution can be changed via a pull
down menu in the upper left (the default for all shaders is 2x).

Post a reply to this message

On 11/08/2015 07:22 PM, Samuel Benge wrote:
> Orchid Win7 v1<voi### [at] devnull>  wrote:
>> That's just trippy. Not quite as intricate as some of the
>> reaction-diffusion systems I've seen, but mental none the less...
>
> Freudian slip? It /is/ probably more of a reaction-diffusion system than it is a
> traditional DLA ;)

It hadn't really registered that it's DLA. It looks like the type of 
organic growth you sometimes see with RD systems.

The difference, of course, is that I understand how DLA works. ;-)

Post a reply to this message