Eric Lerner is presenting his recent paper on a static universe on PhysicsForums

[SelfSim]

This is most regrettable.
Lerner didn't deserve it and hopefully the ban is temporary.

Yeah .. something's gone on in the background on that one, methinks. He didn't deserve being suspended. Interesting that they didn't lock it?

As a matter of interest I downloaded a dithered and drizzled five image combination of M30 with the f814w filter and compared to the single image version I examined previously.

Five image combination FWHM:- 0.110 +/- 0.005 arcsec.
Compared to single image FWHM:- 0.326 +/- 0.058 arcsec.

The five image version is now critically sampled.
This highlights the dangers of doing resolution analysis on single and stacked images which perhaps Lerner didn't take into consideration.

I'm a bit lost on this.
How does the sampling become 'critical'?
Isn't the sampling rate a parameter you control in the AIP4WIN software?
How does the FWHM figure become negative? (Or is that the point you're making about 'the dangers'?)
Cheers

 

[sjastro]

Yeah .. something's gone on in the background on that one, methinks. He didn't deserve being suspended. Interesting that they didn't lock it?

I think the moderators at the very least should explain their motives.

I'm a bit lost on this.
How does the sampling become 'critical'?
Isn't the sampling rate a parameter you control in the AIP4WIN software?

Sampling relates to the radius of the PSF of a star to the size of the pixels.
If the radius is too small relative to the pixel size then the image is said to be undersampled and is characterized by blocky or square looking star shapes.
If the radius is too large relative to the pixel it is said to be oversampled which is not a bad thing since images have a greater latitude for being post processed without the introduction of artefacts such as noise.
Sampling is critical when the radius is neither over nor undersampled and meets the Nyquist requirement.

Pictures tell a thousand words.

The left hand side of the enlarged image is the single exposure of the M30 using the 814W filter.
It is clearly undersampled as the stars are blocky while the background is noisy.
The right hand side is a 5 image stack of M30 performed by Hubble using the same filter and camera.
In this case the individual images were sightly offset and then combined so that the PSF covers more pixels. The PSF was then reconstructed to produce a critically sampled image.

The FHWM measurements I performed on the single image are clearly suspect as the PSF should be round not square and the S/N ratio be reasonably high.

How does the FWHM figure become negative? (Or is that the point you're making about 'the dangers'?)
Cheers

The FWHMs can never be negative.
It reads "FWHM:-" "0.110 +/- 0.005 arcsec."
Not "FWHM:" "- 0.110 +/- 0.005 arcsec."

Sorry for any confusion.
The dangers of using undersampled noisy images particularly the Galex data can result in erroneous radius measurements.
I suspect this is what Lerner has done.

 

[sjastro]

Just to make sure my own FWHM measurements were consistent with Hubble's measurements on critically sampled images, 1417 stars were measured in the f373n data of NGC6302 composed of 8 images which were drizzled and dithered.

The Hubble spec. for the FWHM @ 370nm is 0.1-0.14.
My result was 0.12 +/- 0.03.

To reiterate Lerner needs to show if the image data he used was critically sampled.

 

[SelfSim]

Ok thanks for that .. a lot clearer now.

So, the five image combination FWHM of 0.110 +/- 0.005 arcsec is a far smaller than the single image FWHM of 0.326 +/- 0.058 arcsec, which means its getting closer to say, the Hubble scope performance for that filter (FWHM and PSF), but still quite a long way from it.
So the point you're making is that the performance (FWHM and PSF) of your scope can be significantly improved by using mulit-image and drizzling/dithering, yes?

I'm now trying to see how Lerner's processing produces its intended results (from his last post) .. Not making much headway on that yet ..

 

[SelfSim]

I notice following my post at physFs querying what happened to Lerner, it was deleted and a mod post appeared in its place saying that thread is now 'closed for moderation'.

I don't think I like what goes on at that place very much .. too mysterious, methinks ..

 

[ianw16]

I notice following my post at physFs querying what happened to Lerner, it was deleted and a mod post appeared in its place saying that thread is now 'closed for moderation'.

I don't think I like what goes on at that place very much .. too mysterious, methinks ..

When I saw your post on ISF last night, I checked on PF. I clicked on his username, and checked his activity. His last post was 56 minutes previously. However, it showed that he was 'last seen' only 6 minutes previously in 'converstations' (?).
So, as you say, perhaps things were afoot behind the scenes.

 

[SelfSim]

When I saw your post on ISF last night, I checked on PF. I clicked on his username, and checked his activity. His last post was 56 minutes previously. However, it showed that he was 'last seen' only 6 minutes previously in 'converstations' (?).
So, as you say, perhaps things were afoot behind the scenes.

Hi ianw;
Yep .. mysterious alright ..
There's been a lot of mysterious things happen round these parts lately also ...

 

[sjastro]

Ok thanks for that .. a lot clearer now.

So, the five image combination FWHM of 0.110 +/- 0.005 arcsec is a far smaller than the single image FWHM of 0.326 +/- 0.058 arcsec, which means its getting closer to say, the Hubble scope performance for that filter (FWHM and PSF), but still quite a long way from it.
So the point you're making is that the performance (FWHM and PSF) of your scope can be significantly improved by using mulit-image and drizzling/dithering, yes?

That's correct.
The theoretical value is based on a perfect world...........perfect optics, perfectly flat filters and most importantly perfect guiding of long exposures.
The actual values are still obscenely low.
The FWHM might be have been slightly lower if Hubble stacked more images.
I use Hubble's drizzling and dithering technique when using a modest 300mm camera lens and 15-20 stacked images which is the optimum number.
Here is an example of what the technique can achieve in the hands of a rank amateur.
IC 4628

I'm now trying to see how Lerner's processing produces its intended results (from his last post) .. Not making much headway on that yet ..

You can appreciate how much Lerner has ignored by looking at this link on how surface brightness for galaxies are measured.

 

[SelfSim]

...
I use Hubble's drizzling and dithering technique when using a modest 300mm camera lens and 15-20 stacked images which is the optimum number.
Here is an example of what the technique can achieve in the hands of a rank amateur.
IC 4628

Awesome ...
Some of those stars appear to be in front of the gas clouds .. are they really, or is this a result of mixing point sources with bright surfaces (gas clouds) in the processing?

 

[SelfSim]

... You can appreciate how much Lerner has ignored by looking at this link on how surface brightness for galaxies are measured.

Sheessh!!
And all this is labelled by him as being 'stellar evolution models', which he then wants to deliberately discard?

 

[sjastro]

Awesome ...
Some of those stars appear to be in front of the gas clouds .. are they really, or is this a result of mixing point sources with bright surfaces (gas clouds) in the processing?

What you see are foreground stars and stars in the red emission nebula which is the result of gas being photo-ionized by these embedded stars.
The emission shown here is the H-Alpha (Ha) emission.
The nebula also emits the doubly ionized Oxygen line (OIII) which is a quantum mechanically forbidden emission line that cannot be reproduced on Earth as it is instantly de-excited through collisions with other atoms.
This means the emission nebula is of extremely low density and is therefore optically thin for the embedded stars to be seen in the image.
The black regions are thick gas clouds that not ionized and are opaque.

 

[sjastro]

Sheessh!!
And all this is labelled by him as being 'stellar evolution models', which he then wants to deliberately discard?

It's not surprising that Lerner's model doesn't need to consider a size evolution model for galaxies as all the necessary causes for evolution have been ignored; on top of the resolution issues; and suspect processes as pointed out by Jean Tate.

It makes you wonder how it got through the peer review process.................

 

[SelfSim]

Y'know, they may have shut the thread down because they decided the debate had run its course ... and Lerner was about to go into that mode with which we are only too familiar?

I notice the thread also reached a magic 100 post count, too. Maybe they decided that 100 posts should be sufficient for him to demonstrate if he had a real case?

All just speculation .. and I guess the saying 'flogging a dead horse' also comes to mind, although I hadn't quite reached that conclusion yet .. but then again, some of the threads here have gone on waaayyy too long, and we've unfortunately become accustomed to that?

 

[SelfSim]

... The nebula also emits the doubly ionized Oxygen line (OIII) which is a quantum mechanically forbidden emission line that cannot be reproduced on Earth as it is instantly de-excited through collisions with other atoms.

What?!! ... Not reproducible '*in the lab*'???!!!

This means the emission nebula is of extremely low density and is therefore optically thin for the embedded stars to be seen in the image.
The black regions are thick gas clouds that not ionized and are opaque.

Hmm ...

 

[sjastro]

Y'know, they may have shut the thread down because they decided the debate had run its course ... and Lerner was about to go into that mode with which we are only too familiar?

I notice the thread also reached a magic 100 post count, too. Maybe they decided that 100 posts should be sufficient for him to demonstrate if he had a real case?

All just speculation .. and I guess the saying 'flogging a dead horse' also comes to mind, although I hadn't quite reached that conclusion yet .. but then again, some of the threads here have gone on waaayyy too long, and we've unfortunately become accustomed to that?

I've noticed at PhysicsForum when threads are closed for moderation it doesn't necessarily mean the death knell of the thread.
Hopefully the thread will be reopened with the emphasis on Lerner submitting his data for analysis.

For example analysis of the Galex image for IC3418 reveals the following.
FWHM:- 6.20 +/- 5.43 arcsec.
The ridiculously high standard deviation makes the data useless for analysis.
The reason is quite simple as the statistical noise for the image σ = 2.037 x 10 ̄ ³ is very high.

By comparison the UV Hubble data of IC3418 which is a drizzled and dithered composition of 6 images using the F275W filter is:
FWHM:- 0.09 +/- 0.01 and σ = 8.786 x 10 ̄ ⁶
The noise is over 1000X less than the Galex image.

I'm assuming Lerner is using more "pristine" Galex data but this needs to be analysed.

NB: This what the reviewers at the Monthly Review of the Royal Astronomical Society should have been doing.

 

[SelfSim]

Eric L has risen from the ashes!! .. This time at ISF here (see his post#200)!

 

[sjastro]

So Lerner was suspended at PF for not quoting people directly.
It is moderation out of control but to then close the thread is a slap in the face to everyone participating in the thread.

It's interesting the Astrostatistics Facebook page Lerner refers to, there are calls for the paper to be retracted from the Monthly Review of the Royal Astronomical Society for the errors that have been pointed out in the various discussion threads.

 

[SelfSim]

Yeah ... I don't particularly feel like starting all over again from scratch at the moment ... he seemed to just completely ignore the issues we raised (as well as your test results). Whilst I think he copped the usual over-moderation treatment, (which I agree seemed unfair), I also think he's using the discontinuity to his own advantage.
Its sometimes hard to handle disparate comments but he is clearly seeking the peer-review which somehow, for some suspicious reason, was completely absent from the MRAS reviewers.

The question is: Why did this happen in the first place?

 

[SelfSim]

Soo much drama over such a flawed paper! ... Now the PFs thread has been re-opened and JT is refusing to post on the FB thread!
At PFs, our queries have escaped moderation but JT's was stomped on.

 

[sjastro]

Soo much drama over such a flawed paper! ... Now the PFs thread has been re-opened and JT is refusing to post on the FB thread!
At PFs, our queries have escaped moderation but JT's was stomped on.

The knock out blow against the paper would be to measure the FWHM in the Galex data used by Lerner.
If the standard deviation and statistical noise is of the same magnitude as found in post #55 it would be impossible for Lerner to accurately assign radius values due to the statistical noise.

We can make the request but it would probably fall into the "independent replication" category that Jean Tate fell foul of.
Still I think it is worth the effort............

PS. The moderation at PF is sheer lunacy.