Black hole scientists stitching the data together

[Zyxzevn]

Black hole sun, won't you come..


Here is the chandra x-ray interference image.


It is a great image.
But I notice the contrast adds bright pixels.
Does the same method also remove bright pixels? Like in the centre?
Something that is common with interference images?

And this seems the same image after removing the "foreground" and "background"
and CSI-style zoom+enhance:

[Webbman]

so in the very very very remote chance this wasnt cobbled together/made-up/completely fabricated im glad they have demonstrated that there is a ring there as you might anticipate in an electromagnetic universe fed by a tubular birkeland current.

[Zyxzevn]

"Reality is always more complex"

The images are produced with different systems and telescopes.
The Chandra version is a real image, and it does not show any black hole.
I first thought it was the Milkyway's centre, but it is the M87 galaxy.

This is Chandra's zoomed in image, with no black hole.


The Milkyway centre has its own image of a black hole,
a ball of light. (posted above).
But Astronomers turned this into a black-hole beam, to avoid conflicts with their model.


This "donut" is a processed image, which was produced by
the EHT's Mutlwavelength working group.
They mixed many different signals from different telescopes,
using special algorithms that increased the resolution with interference and
likely assumed that there must be a black hole and nothing else.
The resolution is very small, as you can see from the lack of detail in the picture.

One problem with interference, is that you get black holes at the wrong places,
when you are out of focus. Which is very close to what we see.

Interference

Interference can produce white spots and dark spots in any configuration.
It is easy to see the parallel with the "black hole" image that was produced.

The resolution is another problem.
I compare it to the "face of Mars", where a low resolution image and image-processing showed
something completely different than the better resolution image.

But still a nice effort by the teams.
A nice thing is that they have now collected a lot of data, which might be used for some useful analysis.
And I expect them to produce papers about fast moving plasma and strong magnetic fields.

[Zyxzevn]

[https://www.youtube.com/watch?v=S_GVbuddri8]Veritasium - First images of Black Holes[/url]
Veritasium explains the mainstream view.

According to the video, they also generated an image of the Milky-way centre.
But according to the comments, the image is a simulation, not from the data.

The video explains the problem of the bad resolution a bit.
It does not go into detail into the interference problem.
(If you are out of focus, you can generate any kind of donut.)

Sadly, the video also shows the enormous confirmation bias with this discovery.

----

One more video:
How to take a picture of a black hole | Katie Bouman | TEDx
I was completely correct with the interference problem.
From the data you can create an infinite number of possible images.
They selected the image that they found the "most likely".
The only selected the donut-shaped images, because that is how they think
it would look like.

So the image is 99.9% fake!
It could look like a star, cross, or anything else with the same certainty.

[Webbman]

They have special filters that remove the light that doesnt fit the model.

[Zyxzevn]

They have special filters that remove the light that doesnt fit the model.

Yes. The video shows that clearly.

Fitting the model is the astronomer's best path towards "success".

[D_Archer]

They claim to have observed the "shadow" of a "black hole".... that is all folks..

Regards,
Daniel

[seasmith]

They have special filters that remove the light that doesnt fit the model.

That's a fact.
Just out of curiosity, do you have some semblance of a model, or maybe an idea of what it is they are actually imaging here ?
I value your opinions, (and even wag's are worthwhile).

[Cargo]

I found a paper which describes the method (I think) of how they got this image. And funny enough, when they simulated a black signal to thier super-resolution processor, they got almost the exact same image. Amazing.
Here is quick page where you can see the image from the Simulation
https://www.miz.nao.ac.jp/eht-j/c/news/topic/20140905-1

And here is where you can read the paper.
"Super-resolution imaging with radio interferometer using sparse modeling"
https://arxiv.org/abs/1407.2422

[Zyxzevn]

I think that this image of a black hole has the SAME scientific value:

[Webbman]

our sun has sunspots. sometimes they are really big/high energy.

while there are no black holes it doesnt mean there are no black suns which is just a sun with full sunspot coverage and a massive ring/torus around it. Unexpected in the highest energy place in the galaxy?

easy to explain as we can see the magnetic fields of the sunspots so total coverage might result in one massive magnetic structure (torus/ring). The sun is likely alot smaller than the hole, assuming that there is even a hole there and its not just a scam.

if its anything at all its that.

[Zyxzevn]

Some more images of the same thing.
Found these at http://saidit.net





Does anyone know what these images mean?
Are this these simulated variants?

Haha. It seems a joke:

The 2nd link and 1st are not a black hole, but is exact match to this picture of a so called black hole, exactly what NASA, 'never a straight answer' is showing you. So many people eating up this picture hole from @NASA I guess you can be shown anything on tv and you will believe it. Mass amounts of people would accept the earth being flat if TV told them it was flat. This is how easily conditioned americans are.

This image has one black hole and 3 PET scans
Which is the real hole?

[Webbman]

the saddest part is that you have to take everything they do with a grain of salt or maybe a truckload of salt as they seem more interested in deception than anything else.

Im confident that youll eventually find suns with extreme sunspot activity though. Sunspot activity is related to energy output. All you have to do is scale it up. Im sure one of you guys could make a model or something demonstrating what happens when you keep adding sunspots.

[Zyxzevn]

I think the general approach of the mainstream is good.
Try to get a higher resolution by mixing the data of different telescopes all over earth.

I have worked with the SAR radar, which has an artificial large size using
a "Synthetic Aperture" algorithm. It combines the phase of an array of different antennas,
and combines them to get a very accurate signal.
It is so accurate that it can also be used to measure changes of the ground after an earth quake.
But it also has its drawbacks.
1. You need to focus exactly all antennas. That is why you usually see them in arrays.
This is a hard problem when you have antennas all over earth.
With SAR you have special places on earth that you use for real-world calibration.
2. Small difference in phase can create large differences in the image. The phase is
linked with the position on the image.
3. Noise is hard to identify without the accurate information of all other antennas.
4. You need very high accuracy that is far beyond the telescope's specifications.
They to get the data from exact the same position.
5. The telescopes have different frequency ranges. Some may not overlap
and can give completely different signals.
6. Errors in the all these above points can give false positives.

The solution would be to do it in small steps. Not go straight for the fuzzy Donut.

They can just do 2 times the resolution of the full system with the "black hole" and then
go a level further each time. This way you keep the consistency of the structure.
And you can directly check whether your data is still accurate or is noise or interference.
You can remove noisy or inaccurate images if necessary, but this also reduces your final resolution.
And I don't think they would have made the final needed resolution.

But this way, you can slowly get more data, even construct a 3D structure.
Then repeat it after half a year, from a slightly different perspective.
This can increase our understanding in improve the accuracy each cycle.

But they just went for the smallest thing possible, left out potentially important data.
Introduced systematic errors. And possibly magnified errors, because they appeared
more like the wanted model.
And they have no way of backtracking.

So, whatever produces that plasma-yet,
this is not science.

And my hope is that after a few years we will see a more accurate picture of
what is really there. And I suspect that we will see a repeat of the
"Face on Mars": After getting more detail, the thing will look completely different.

[Infinion]

Here's a link to the paper for anyone to read through https://iopscience.iop.org/article/10.3 ... 13/ab0e85

And here is the 2017 Event Horizon Telescope imaging challenge http://vlbiimaging.csail.mit.edu/
The challenge allowed the teams to calibrate their super-resolution sparse modeling algorithms using samples of ground truth images of accretion disks, as well as without (jets/nebulas/galaxies), then tasked to analyze blind raw data.

In section 5 of the paper, it mentions that the teams were blind to each others' work and prohibited from discussing their imaging results and parameters. It goes on to say that no restrictions were imposed on the procedures or data processing by each team.

It sounds good, but this just means that each team used different sparse modeling algorithms with differently biased parameters that were not shared. They don't need these restrictions because each team's algorithm is trained by the same ground truth images and test data. The purpose of this demonstration was to get each team's image to agree geometrically in size and shape, showing that they could independently generate faithful reconstructions and demonstrate persuasive certainty.


In section 6, they say the purpose of the blind tests was to identify intrinsic from spurious data within the possibilities of a coarse-grained model.

Our surveys are coarse-grained and do not completely explore the choices in the imaging process. Nonetheless, they identify regions of imaging parameter space that consistently produce faithful image reconstructions on synthetic data, and they help us identify which features of our reconstructions are consistent and which features vary with specific parameter choices.

This highlights the fact that bias does indeed still exist to simplify and constrain the possibilities down to the four geometric models that are most likely, as seen in figure 10 of 6.3.2 of a Ring, Crescent, Disk, and Double, along with a bonus GRMHD computer-generated image of an accretion disk.


You could easily substitute a donut, cross, star/sphere, or elephant as the ground truth image and you'd generate a new set of final reconstructions instead.


[Larger scale image from figure 5]

That's what's so dangerous about this interferometer super-resolution sparse modeling technique. When a team tries to collect agreeing reconstructed data from an infinite number of possible interferometer images, even if the teams are blind to each others' work, even if the ranking system minimizes reliance on expert judgment, the assumptions of the model that the explanations are carefully made to respect, remain in the work to influence the result towards a single possibility.

And this persuasive power is treated as confidence without a shadow of a doubt in the public's eyes, as we are seeing in the hundreds of youtube videos, comments, and press, further rooting deep-seated belief in Einstein's theories and big bang confirmations. It's going to be a major tool.