Cont: Why James Webb Telescope rewrites/doesn't the laws of Physics/Redshifts (3)

[Mike Helland]

Wait, what? Red shift moves the spectrum to larger wavelengths, not smaller wavelength.

Doh. Should have used frequency. (THz instead of nm)

 

[Mike Helland]

Suppose an instrument that counted incoming photons and put them into energy bins. And then graph the count of each bin. That would be a spectrum.

When we do that for the sun, or a nearby galaxy, how many bins are used?

And how many are used for the CMB? From this image it looks like 40:

https://astro.ucla.edu/~wright/CMBspect.gif

If the sun's spectrum was measured using 40 bins, it'd probably smooth out a bunch.

I'm not saying the CMB is made of star light or galaxy light. And I'm not saying the CMB's near black body spectrum is unimportant or unremarkable.

It could be a little over-hyped though.

 

[Ziggurat]

When we do that for the sun, or a nearby galaxy, how many bins are used?

That often depends on the instrument. Instruments always have some limit to their resolution, so there's no point to binning much smaller than this resolution. In addition, the smaller the bin size, the fewer counts per bin. If the signal is low, then statistical noise in the number of counts for any one bin can be significant, and larger bins can help smooth out this statistical noise. You can always choose to reduce the number of bins by just combining existing bins.

If the sun's spectrum was measured using 40 bins, it'd probably smooth out a bunch.

It would definitely be smoother than a spectrum with, say, 4,000 bins. But even smoothed out, at 40 bins it would still noticeably deviate from a blackbody spectrum.

 

[The Man]

Suppose an instrument that counted incoming photons and put them into energy bins. And then graph the count of each bin. That would be a spectrum.

It would be a discrete colletection of 40 energy bins, a discrete spectrum. Photons are already "energy bins", they are discrete packets (quanta) of energy.

When we do that for the sun, or a nearby galaxy, how many bins are used?

And how many are used for the CMB? From this image it looks like 40:

https://astro.ucla.edu/~wright/CMBspect.gif

If the sun's spectrum was measured using 40 bins, it'd probably smooth out a bunch.

I'm not saying the CMB is made of star light or galaxy light. And I'm not saying the CMB's near black body spectrum is unimportant or unremarkable.

It could be a little over-hyped though.

Who is even 'hyping' it, let alone over hyping it? It simply is what is observed. The basis of Planck's theory of blackbody radiation is the consideration of the source as a collection of quantized harmonic oscillators that can only emit or absorb energy in discrete packets.

 

[IGMisBestM]

I think they rely on a thermal equilibrium to get there.

They don't "get there". They don't calculate spectra, they don't even claim it gives the right spectrum, they have not attempted to check. They acknowledge in the text that doing a properly would involve radiative transfer.

Would we still see the same line just as clearly in the compressed data when it's 0.45 nm?

Redshift doesn't change the detectability. Whether it can be detected depends on the strength of the line relative to the continuum and the sensitivity of the instrument. COBE FIRAS and Planck can measure lines, they detected galactic emission lines.

Note that a solar-like spectrum even with the absorption lines removed results in a skewed spectrum which violates the FIRAS CMB spectrum constraints by a factor of about 1000, the error bars on the plot are greatly exaggerated. And the SED from a dusty star forming galaxy would do even worse.

https://www.astro.ucla.edu/~wright/stars_vs_cmb.html

It seems to be motivated by what conditions must have led to the high-redshift galaxies we're seeing in the state we observe them.

It is not. None of the galaxies found with JWST are anything like that. Some early quenched galaxies have been confirmed at lower redshifts. But even the most extreme formed later than they propose, and they are all a factor of 10 less massive than assumed in their calculations. The galaxies they are inventing do not exist. The number density of those passive galaxies at z=3-5 is 20 times lower than assumed, even ignoring the huge difference in mass.

This is not motivated by observations, it's just a sad attempt to revive some silly MOND inspired and attack the CMB as a cosmological probe. There is a reason the paper doesn't claim they can explain the CMB, because they can't. And here they are literally post hoc fitting whatever they want, without regard for physics or observational data. When the CMB was first detected there were lots of similar after-the-fact hypotheses of alternative explanations. But big bang cosmology predicted uniquely the precise blackbody spectrum and it's detailed angular fluctuations. There are exactly zero real alternative models who can explain either, even with all the post hoc assumptions. As it stands the CMB is still direct eveidence that the early universe was very dense and hot.

As an aisde, if ALMA has been operating for 10+ years, how come it didn't find those Little Red Dots? Not in its frequency range or something?

In general ALMA is not efficient at finding galaxies blindly because of it's small field of view, and narrow instantaneous frequency settings. You need to know the redshift pretty accurately to get a line, and only massive dusty galaxies can be detected in continuum. But with LRDs specifically there wasn't much to miss, most are undetected in ALMA follow-up. There is a tentative line detection in one low redshift bright object. Some of the LRD galaxies were previously detected by different facilities, but they aren't particularly unusual in those observations.

 

[theprestige]

Wait, is Mike saying everything is actually a black body, if scientists actually take a proper look?

 

[Ziggurat]

Wait, is Mike saying everything is actually a black body, if scientists actually take a proper look?

No, he's suggesting the CMB only looks like a black body because we're not taking a proper look.

This is not an unreasonable question, but the answer is that this is not why.

 

[Mike Helland]

Photons are already "energy bins", they are discrete packets (quanta) of energy.

Well, yes to the second part. No to the first part.

A photon's energy is E=hf or E=hc/w.

The wavelength could be 100 nm or 101 nm, or literally anywhere in between. Light comes in chunks, but those chunks can be of any energy level.

 

[Mike Helland]

The number density of those passive galaxies at z=3-5 is 20 times lower than assumed, even ignoring the huge difference in mass.

So the big bang predicted how many passive galaxies between z=3 and z=5 there were, and it's predictions were 20 times too high?

This is not motivated by observations, it's just a sad attempt to revive some silly MOND inspired and attack the CMB as a cosmological probe.

Do you have a beef with the authors or something?

Seems like that came out of left field.

Edit:
I see Kroupa has a history of claiming that there are failures in the current cosmological model, even co-authoring with Milgrom.
P. Kroupa, M. Pawlowski, M. Milgrom, "The Failures of the Standard Model of Cosmology Require a New Paradigm", IJMPD 21, 1230003 (2012).
Obvious heretics deserve obvious derision, ammirite? ;-)

 

[IGMisBestM]

So the big bang predicted how many passive galaxies between z=3 and z=5 there were, and it's predictions were 20 times too high?

No, this is Gjergo and Kroupa's model.

And yes, I have a problem with people publishing nonsense. The data to test their claims already existed, but no attempt to ask if what they propose was really physical. The data already exists to refute it. And basic questions were ignored. Kroupa has a strong bias. He claims in the paper that this model was validated by JWST, but actually it can predict anything. Because it's not coupled to any cosmological simulation it cannot predict number densities. The paper is one thing, but the comment in the article is just totally fallacious and deeply ignorant.

 

[Mike Helland]

No, this is Gjergo and Kroupa's model.

Ah, thanks. That makes more sense.

And yes, I have a problem with people publishing nonsense. The data to test their claims already existed, but no attempt to ask if what they propose was really physical. The data already exists to refute it. And basic questions were ignored. Kroupa has a strong bias. He claims in the paper that this model was validated by JWST, but actually it can predict anything. Because it's not coupled to any cosmological simulation it cannot predict number densities. The paper is one thing, but the comment in the article is just totally fallacious and deeply ignorant.

Fair enough. Since I'm skeptical the big bang even happened, it's not like I agree with their most basic premise, that z>10 galaxies had to form so rapidly.

 

[The Man]

Well, yes to the second part. No to the first part.

The first part is dictated by the second part.

A photon's energy is E=hf or E=hc/w.

The wavelength could be 100 nm or 101 nm, or literally anywhere in between. Light comes in chunks, but those chunks can be of any energy level.

You're missing the point. 100 nm would be an energy of 12,398 meV while 101 nm would be 12,276 meV. If you are measuring energy and your resolution > 122 meV they both end up in the same energy reading bin. Further if you are measuring energy you can only determine frequency (1/t) to ΔE * Δt ≥ ħ/4π. Photons are produced in discrete packets by physical processes and measured in discrete packets by often differing physical processes. They are produced in bins by said physical processes and received into likely different bins by different physical processes with resolution and uncertainty thrown in on top of that. So can you stack the receiving bins to say load the deck to get results more to one's liking? Sure, but it is meaningless as just the measurement process forces some such bias aspects on the data. The trick to try to control for and even move away from some particular bias and then see if the data still support the hypothesis.

 

[The Man]

Ah, thanks. That makes more sense.


Fair enough. Since I'm skeptical the big bang even happened, it's not like I agree with their most basic premise, that z>10 galaxies had to form so rapidly.

Do you agree with the basic premise that the early universe was very dense and hot? Perhaps even back to some point where photons could not travel freely through space?

To try put it more succinctly, your claimed skepticism, of the big bang, appears to be functionally misdirected at the CMB.

 

[Mike Helland]

100 nm would be an energy of 12,398 meV while 101 nm would be 12,276 meV. If you are measuring energy and your resolution > 122 meV they both end up in the same energy reading bin.

Unless there was a boundary between bins at 12,300.

They are produced in bins...

Doesn't a black body have a continuous spectrum?

 

[Ziggurat]

Unless there was a boundary between bins at 12,300.

I have no idea what sort of boundary you are imagining.

Doesn't a black body have a continuous spectrum?

A perfect black body does, yes.

 

[Mike Helland]

I have no idea what sort of boundary you are imagining.

A perfect black body does, yes.

The observer defines their bins. They could between 1 and 2 and 3, etc, or 1.05 and 1.55. Nothing says they have to be nice round numbers.

 

[Ziggurat]

The observer defines their bins. They could between 1 and 2 and 3, etc, or 1.05 and 1.55. Nothing says they have to be nice round numbers.

Sure. But I'm not sure how that changes The Man's point in regards to instrumental resolution. If your instrument's resolution is larger than the difference between two photons, you cannot effectively put them in separate bins.

 

[Mike Helland]

If your instrument's resolution is larger than the difference between two photons, you cannot effectively put them in separate bins.

Sure. You previously said "Instruments always have some limit to their resolution, so there's no point to binning much smaller than this resolution." I think we're all on the same page there.

As for the specific examples of photon energies 12.398 or 12.276, and a resolution of > 0.122, doesn't that sort of depend on the instrument though?

If we assumed there was an instrument that could give us the following readings:

12.000,
12.125,
12.250,
12.375,
12.500,

and you could control the energy of the photons sent to the instrument so they were either 12.398 or 12.276, it seems intuitive to me that, your readings would be mostly 12.250 and 12.375.

You could decide that, since those are the smallest increments of readings you get, that you should use larger bins. Presumably, you could combine two bins into one.

So you could chose:

12.000,
12.250,
12.500,

or you could choose:

11.875,
12.125,
12.375,
12.625

Which bins you choose will affect your results. Of course, with enough data, statistically that all washes out.

In any case, the idea that quantum means the energy a photon can have comes from a discrete set of values is a mistake I've made in the past, and it's very possible you were the one that corrected me on that point back then.

 

[Ziggurat]

Which bins you choose will affect your results. Of course, with enough data, statistically that all washes out.

Correct.

In regards to the CMB, we have enough statistics to show it's an incredible fit to a black body spectrum, much better than any stellar body. The number of bins isn't actually much of a limit on this.

In any case, the idea that quantum means the energy a photon can have comes from a discrete set of values is a mistake I've made in the past, and it's very possible you were the one that corrected me on that point back then.

It's possible, I don't specifically remember.

 

[Mike Helland]

In regards to the CMB, we have enough statistics to show it's an incredible fit to a black body spectrum, much better than any stellar body.

No argument there, but comparing a 3 K black body spectrum to a 3000 K black body spectrum seems a bit extreme.

A quick Google says the lowest temperatures we've gotten in labs in 38 picokelvins. So, something like 2.5 K sounds like it would be easier.

How does a 2.5 K artificial black body in the lab's spectrum compare to the CMB's ~2.7 K spectrum in faithfulness to an ideal black body?

My guess is much much better than a star. Maybe even better than the CMB's.

 

[Ziggurat]

No, it would be worse.

 

[Mike Helland]

No, it would be worse.

Huh. Well, that I'd like to see.

Comparing a pin-drop to an earthquake doesn't make as much sense.

I'm not really finding any examples of such a comparison though. Recreating a black body spectrum of a similar temperature to the ~2.7 K CMB in a lab seems like something I'd expect to happen to on a somewhat regular basis. An exercise many aspiring experimental physicists would do as a student. Or even just showing high school students a comparison between a lab-based 2.7 K black body and the CMB. Maybe that's just plain naive. But maybe I'm just not googling right.

It is said that the CMB is the closest spectrum to a black body in nature. Including or excluding contemporary lab results? I also thought that this statement referred to the CMB, way back when it was 3300 K, not as it's observed today.

I thought that, even in theory, the CMB would not reach us as a perfect black body for reasons involving electrons and such.

https://astro.ucla.edu/~wright/CMBspect.gif

"The three curves in the bottom correspond to three fairly likely non-blackbody spectra: the grey curve shows a body with a reflectivity of 100 parts per million instead of zero, and the red and blue curves show the effect of hot electrons adding an excess 60 parts per million of energy to the CMB either before(blue) or after (red) 1000 years after the Big Bang. These curves show the maximum distortions allowed by the FIRAS data."

https://astro.ucla.edu/~wright/CMB.html

Anyways, got any tips for finding papers of low temperature (e.g., less than 50 K) black body spectra produces in labs?

 

[Ziggurat]

Huh. Well, that I'd like to see.

If it were easy to make something black in the long wavelength of the spectrum, stealth aircraft would be common. As it is, not only are they not common, they aren’t even black, only dark grey in those wavelengths. You don’t get anything even close to just 100 ppm reflectivity.

 

[Mike Helland]

These look interesting:

The Fermilab Large Cold Blackbody Test Stand for CMB R&D

https://lss.fnal.gov/archive/2011/conf/fermilab-conf-11-641-ppd.pdf

Design and validation of a cold load for characterization of CMB-S4 detectors

https://lss.fnal.gov/archive/2024/pub/fermilab-pub-24-1008-ppd.pdf

 

[Mike Helland]

Couple videos that came out yesterday.

Dr. Becky Smethurst "JWST's 3 Big mysteries of the early universe"

PBS Space Time's "Is Our Model of Dark Energy Wrong?"

 

[Mike Helland]

Cool JWST stuff:

Viewer – COSMOS-Web DR1

cosmos2025.iap.fr

COSMOS-Web Public Data Release 1​

This page hosts the version 1.0 public data release for COSMOS-Web, including:​

• Reduced NIRCam and MIRI mosaics
• Aperture and model-based photometric catalogs
• Photometric redshifts & physical properties for ~780k galaxies
• Interactive map viewer over the full field

 

[Mike Helland]

COSMOS-Web

cosmos2025.iap.fr

MIRI:

COSMOS-Web

cosmos2025.iap.fr

How many galaxies are pictured in the above image? Dozens? Or thousands?

The story we all know is that they pointed the Hubble Space Telescope at an empty patch of sky for a while, and found all these things.

Then they pointed JWST at the same patch of sky and saw so much more.

And look at all the weird noise between all those things. Is that all just noise?

Would it really be that much of a surprise if the next generation of telescopes, or even just super-extended exposures with JWST reveal that some, or even a lot of that noise is actually more galaxies?

Another way to put it is, if you simply had to guess, how many of the galaxies that are theoretically observable do you think have been observed?

We can simplify the question by saying not over the whole sky, but in the field depicted by COSMOS-Web DR1

95%?

75%?

10%?

1%?

 

[Darat]

No idea - but we do know it's a finite number...

 

[IGMisBestM]

And look at all the weird noise between all those things. Is that all just noise?

Would it really be that much of a surprise if the next generation of telescopes, or even just super-extended exposures with JWST reveal that some, or even a lot of that noise is actually more galaxies?

Another way to put it is, if you simply had to guess, how many of the galaxies that are theoretically observable do you think have been observed?

We can simplify the question by saying not over the whole sky, but in the field depicted by COSMOS-Web DR1

The answer about how many objects are detected in the image, you can over-plot the catalog for one estimate. The answer is probably dozens. Some of the blobs of noise just below the detection threshold would probably turn out to be real galaxies in deeper data, but many wouldn't. The noise in images like this dominated by noise in the background, which has been subtracted but there are still fluctuations. Most of the noise in the MIRI data for example is smaller than the point spread function, so cannot be real galaxies. There are some datasets which are confusion limited, this isn't one of them.

If would be strange if a bigger telescope didn't find more galaxies. The faint end of the luminosity function of galaxies is constrained from fields with lensing.

There is no objective answer to how many galaxies there are in total, because what counts as a galaxy is not well defined. Galaxies have a Schechter function luminosity distribution, which continues rising for lower and lower masses galaxies. If you integrate it it does not converge in number, unless you impose a limit. You can estimate the number down to some observational flux, or above some limiting mass, but not all of them. Simulations also have a finite resolution. Scientifically it's not an interesting number, people work in luminosity functions instead. Dwarf galaxies dominate the number but most of the mass in stars is in Milky way and above mass galaxies.

 

[Mike Helland]

The answer is probably dozens.

Isn't that in the "we'll take your word for it" territory?

When the HST was pointed at a dark patch... there was probably nothing there, they said.

When JWST was pointed at it... we'll see the things we've already seen, but better.

We've always been wrong. There was way more than expected. And we've always been told, we're right this time.

Doesn't skepticism seem reasonable?

Doesn't it seem possible we won't see just a little more with a better telescope, but a lot more? Like we always do?

Most of the noise in the MIRI data for example is smaller than the point spread function, so cannot be real galaxies.

Can you explain why that is?

 

[Mike Helland]

Most of the noise in the MIRI data for example is smaller than the point spread function, so cannot be real galaxies.

COSMOS-Web

cosmos2025.iap.fr

Here's some galaxies in NIRCam RGB view, the red dot is a z=8.8 galaxy, the one below it is z=5.2, and the other ones are around z~2.

And here's the same things in the MIRI view:

It seems to me a lot of those unclassified blobs are larger than some of the ones that are classified.

Anyways. I'm not saying we've only identified a small fraction of what's there so far. But if anyone's taking bets, I'd put some money on the "tip of the iceberg" side.

 

[IGMisBestM]

Isn't that in the "we'll take your word for it" territory?

No. There is deeper data of many fields. COSMOSWeb is quite shallow in comparison. How many galaxies are detected is a statement about the current data, relative to the current noise.

In the case of COSMOSWeb the objects are detected in the NIRCam data. And so an object in the catalog doesn't necessarily have to be detected in the MIRI image. Seeing a blob of pixels doesn't mean it's above the noise threshold.

Doesn't it seem possible we won't see just a little more with a better telescope, but a lot more? Like we always do?

Yes, that is what I said.

Can you explain why that is?

Even a point source like a star shows up on the image with some minium size set by the limiting angular resolution of the telescope. Any real object cannot be smaller than that. Individual noisy pixels are not real sources.

 

[Mike Helland]

Even a point source like a star shows up on the image with some minium size set by the limiting angular resolution of the telescope. Any real object cannot be smaller than that. Individual noisy pixels are not real sources.

Ok. Does that require a constant stream of photons?

If we look at NIRCam and MIRI's abilities, compared with the visible frequencies of light and various amounts of redshift, I get something like this:


We see by z=12, all the visible light has redshifted past NIRCam's wavelength range.

If we can consider this to be a standard looking spectrum for a galaxy:

Messier 31 – thecuriousastronomer

Posts about Messier 31 written by RhEvans

thecuriousastronomer.wordpress.com

Then at z>12 the only stuff NIRCam would see is that <4000 Å range.

Compared to the strength and (relative) consistency of the visible spectrum, anything there would appear very noisy.

That's just me jumping to conclusions as a layman. Is that unreasonable / inaccurate?

 

[Mike Helland]

If it were easy to make something black in the long wavelength of the spectrum, stealth aircraft would be common. As it is, not only are they not common, they aren’t even black, only dark grey in those wavelengths. You don’t get anything even close to just 100 ppm reflectivity.

What about ice?

Not as material for a stealth air craft. But ice does emit microwaves, does it not?

If there are billions of pieces of ice surrounding our solar system, that should add up to something?

I think snow is an even better mimic of a black body? Snow is just air and ice.

A somewhat recent development in our understanding of the solar system is that it's surrounded by a "hydrogen wall":

NASA Spotted a Vast, Glowing 'Hydrogen Wall' at the Edge of Our Solar System

NASA Spotted a Vast, Glowing 'Hydrogen Wall' at the Edge of Our Solar System

There's a vast "hydrogen wall" at the edge of our solar system, and NASA scientists think their New Horizons spacecraft can see it.

www.livescience.com

If there was a wall of hydrogen around the solar system, that also happened to contain billions or trillions of ice crystals, and it's been heated to ~3 K based on starlight from the rest of the galaxy (as Eddington calculated in the early 1900's), what would that look like to us?

 

[IGMisBestM]

Ok. Does that require a constant stream of photons?

No. If you don't have enough photons to see the spatial extent, then it's probably not significantly detected above the noise anyway.

Then at z>12 the only stuff NIRCam would see is that <4000 Å range.

Compared to the strength and (relative) consistency of the visible spectrum, anything there would appear very noisy.

That's just me jumping to conclusions as a layman. Is that unreasonable / inaccurate?

That is not a good standard spectrum. Most of the very high redshift galaxies being observed are Lyman break galaxies, where NIRCam and NIRSpec cover the rest-frame UV. LBGs have steep blue UV spectra, unlike the rest spectrum of M31. Meaning they are brighter in bluer wavelengths. The LBGs are highly star-forming low-mass galaxies, compared to the massive almost-quenched M31.

https://jades-survey.github.io/assets/images/GN-z11_allspec_1D_2D.png

Secondly, you cannot say a spectrum or image will be noisy just from seeing the spectrum is red. That can only be said knowing the sensitivity of the spectrum or image and the flux of the target.

In any case most of the sources the image will be z<5, meaning MIRI doesn't probe rest frame visible light at all. What wavelength the light started with doesn't matter to whether or not a source is noisy.

 

[Ziggurat]

If there was a wall of hydrogen around the solar system, that also happened to contain billions or trillions of ice crystals, and it's been heated to ~3 K based on starlight from the rest of the galaxy (as Eddington calculated in the early 1900's), what would that look like to us?

If the source of the CMB was local (as in, around our solar system), then regardless of what it was made of, it would block microwave signals from beyond our solar system. That's the nature of blackbodies: in order to emit a blackbody spectrum, they also have to absorb everything.

But we see more distant microwave sources. They aren't being blocked. So the source of the CMB cannot be local.

 

[Mike Helland]

If the source of the CMB was local (as in, around our solar system), then regardless of what it was made of, it would block microwave signals from beyond our solar system. That's the nature of blackbodies: in order to emit a blackbody spectrum, they also have to absorb everything.

But we see more distant microwave sources. They aren't being blocked. So the source of the CMB cannot be local.

That would require something like a solid shell of ice around the solar system to block all of that out though, wouldn't it?

Even with a trillion ice crystals surrounding the solar system, there would be plenty of room in between them for other light and microwaves to get through.

 

[Ziggurat]

That would require something like a solid shell of ice around the solar system to block all of that out though, wouldn't it?

Not solid, but optically opaque, yes. As in, no line of sight without going through whatever that is.

But that's required to produce a blackbody spectrum. If our solar system was surrounded by black body objects that only covered half the line of sight, then it would produce a spectrum only half the intensity of a black body. Because it's not just the shape of the curve, the height of the curve matters too.

Even with a trillion ice crystals surrounding the solar system, there would be plenty of room in between them for other light and microwaves to get through.

Then it cannot produce a black body spectrum averaged over the sky even if each object is a perfect black body. The intensity would be too low. It would look like a grey body.

This is why we know that the CMB cannot be local, regardless of what it's made out of. Even a hypothetical material that's transparent in every wavelength except microwave, and it's a perfect blackbody in microwave, still wouldn't match observations.

 

[Mike Helland]

No. If you don't have enough photons to see the spatial extent, then it's probably not significantly detected above the noise anyway.
That is not a good standard spectrum. Most of the very high redshift galaxies being observed are Lyman break galaxies, where NIRCam and NIRSpec cover the rest-frame UV. LBGs have steep blue UV spectra, unlike the rest spectrum of M31. Meaning they are brighter in bluer wavelengths. The LBGs are highly star-forming low-mass galaxies, compared to the massive almost-quenched M31.

Ok. But if there were quenched galaxies out there similar to M31, and if the noise threshold is set to a relative intensity of 0.4 and above, their signal will be regarded as just noise. If we assume no such galaxies exist at those distance, I guess it's not worth giving much thought to.

Also, if UV light is 10 nm - 400 nm at z=0, then at z=9, it would be 100 nm to 4000 nm. That's a much bigger range. Wouldn't stretching it out like that also contribute to noisyness?

At the highest observable redshifts at any point over the last 100 years, we've always seen the bright star forming galaxies first. And then we find there are mature galaxies among them, they're just harder to see.

It seems to me a good bet there's the whole rest of the ice berg out there, and we're just seeing the tip. Some of the noise is just noise, sure. But some of it, a lot of it could be faint signals. It could even be x-rays from z=70 galaxies (if my math is right).

Based on how many galaxies we've observed at say z < 0.1, and extrapolating that number to larger and larger volumes, it seems like 99.99999% of the observable universe has still yet to be charted.

 

[Mike Helland]

Not solid, but optically opaque, yes. As in, no line of sight without going through whatever that is.

The CMB frequencies are no where near optical light, so that seems to be superfluous stipulation.

But that's required to produce a blackbody spectrum. If our solar system was surrounded by black body objects that only covered half the line of sight, then it would produce a spectrum only half the intensity of a black body. Because it's not just the shape of the curve, the height of the curve matters too.

Then it cannot produce a black body spectrum averaged over the sky even if each object is a perfect black body. The intensity would be too low. It would look like a grey body.

This is why we know that the CMB cannot be local, regardless of what it's made out of. Even a hypothetical material that's transparent in every wavelength except microwave, and it's a perfect blackbody in microwave, still wouldn't match observations.

I think I can see what you're saying, if we assume our map of the CMB has more pixels than are microwave sources.

If there are about 80 million pixels in the Planck CMB map, and trillions of ice crystals emitting microwaves, there isn't a separate pixel for each source. What happens then?