APOD: Planets of the Solar System (2022 Jun 25)

[jfgout]

How are you so certain that the image of Mercury shown in the APOD is actually a star? I didn't understand your first explanation above. Also, there's a bright object - a star? - very close to Venus and being partially eclipsed by it, that I'm not clear about either.

mercury and venus.JPG

You can use Stellarium or any other planetarium software to see what things were supposed to look like. Doing so will show that in the APOD image, Mercury (or should I say, the star-like feature visible just above the le label Mercurio) is only about half of its true distance to Venus. This can be easily overlooked by someone with limited experience, but anyone who has used an atlas with their telescope (no GoTO) will immediately notice the problem. If I have the time tonight, I'll overlay the Stellarium view to the image to illustrate my point.

Of course, it could be an image of Mercury taken 30 minutes latter and then just incrusted into the image or it could be an image of a random star. The point is that, taking the positions of Venus, Moon, Mars, Jupiter, Uranus, Neptune and Saturn as reference, there should be nothing at the location marked as "Mercury" (or, at best some magnitude ~4 star, which would be invisible since we can't even see the Pleiades)

As for the object close to Venus, this is probably just Venus again from a different layer. With Venus being towards the edge in one layer, deformations due to lenses having more coma/etc near the edges will make it difficult to align perfectly with Venus from a more centered picture. That's fine.

jf

[johnnydeep]

How are you so certain that the image of Mercury shown in the APOD is actually a star? I didn't understand your first explanation above. Also, there's a bright object - a star? - very close to Venus and being partially eclipsed by it, that I'm not clear about either.

mercury and venus.JPG

You can use Stellarium or any other planetarium software to see what things were supposed to look like. Doing so will show that in the APOD image, Mercury (or should I say, the star-like feature visible just above the le label Mercurio) is only about half of its true distance to Venus. This can be easily overlooked by someone with limited experience, but anyone who has used an atlas with their telescope (no GoTO) will immediately notice the problem. If I have the time tonight, I'll overlay the Stellarium view to the image to illustrate my point.

Of course, it could be an image of Mercury taken 30 minutes latter and then just incrusted into the image or it could be an image of a random star. The point is that, taking the positions of Venus, Moon, Mars, Jupiter, Uranus, Neptune and Saturn as reference, there should be nothing at the location marked as "Mercury" (or, at best some magnitude ~4 star, which would be invisible since we can't even see the Pleiades)

As for the object close to Venus, this is probably just Venus again from a different layer. With Venus being towards the edge in one layer, deformations due to lenses having more coma/etc near the edges will make it difficult to align perfectly with Venus from a more centered picture. That's fine.

jf

So, here's what I see with Stellarium set to Ladispoli, Italy on 6/24/2022 compared with the APOD.

[jfgout]

So, here's what I see with Stellarium set to Ladispoli, Italy on 6/24/2022 compared with the APOD.

planet alignment on 6-24 from Ladispoli, Italy.JPG

planet alignment on 6-24 from Ladispoli, Italy APOD.JPG

Thank you for posting the screenshot. In case it's not obvious to everyone, you can compare the distance between Mars and Jupiter to that between Venus and Mercury. On the Stellarium screenshot, you can see that it is roughly the same. However, on the APOD image, the distance between Mercury and Venus is a LOT shorter than that between Mars and Jupiter.

Note that lens distortions and the like can make comparing distances in pictures delicate, but that's not what is happening here.

So, what is this bright point of light labeled as Mercury in the APOD image? Nothing but a star (or planet) taken from a different image and incrusted in the wrong location(!!!!) on this image to make us believe that they had Mercury in their image. In plain English, we call this a lie (confusing Mercury with a bright star would be a honest mistake).

jf

[mar_s]

I totally agree with jfgout. Saturn is even more off than mercury. If you look closer everything is somehow off. In the following image Venus and Jupiter are used as reference points.



I tried many different projection methods, but no chance to fit the view.

mar_s

[mar_s]

A different approach:


I also tried to use some stars (Formalhaut and the stars of Aries) as reference points. And yes, I also took into account that there still could be some distortion or mapping/projection issues. Try all this yourself and you'll see this APOD is utter nonsense.

[jfgout]

I totally agree with jfgout. Saturn is even more off than mercury. If you look closer everything is somehow off. In the following image Venus and Jupiter are used as reference points.



I tried many different projection methods, but no chance to fit the view.

mar_s

Thanks mars_s.

For Saturn it can be a bit tricky because, as you have already noticed, the projection method used can make a big difference. But for Mercury, there is no question and your overlaid images show it nicely. What really annoyed me is that the authors of this image went through the trouble of adding something to make it look like they had captured Mercury, rather than just say something like: "unfortunately, Mercury was too low on the horizon to be captured in this image."

jf

[mar_s]

This kind of reminds me of the fake Milky Way photos that NatGeo published a few years ago:
https://petapixel.com/2019/05/10/scient ... ay-photos/

IMHO, this kind of fakery is ruining APOD's reputation.

[Chris Peterson]

There's nothing wrong with constructing an image like this out of stacked layers.

Using stacked layers can be an acceptable method, but not the way it was done here! Maybe you are still under the impression that they simply used layers to cover a wider field of view than what camera can record (which would be fine).

But no, here, they must have used a layer from a picture of who knows what to add a star at the position where they would have liked to see Mercury. This goes against the ethic's statement from APOD and against simple principles of honesty.

I obviously have no problem with people using layers, such as in the recent analemma APOD.

jf

My assumption is that the layers cover the same FOV (or approximately so), and were taken at different exposures in order to cover the wide dynamic range of the different planets. Or perhaps a combination of exposure time and position, optimized for the targets. Perfectly acceptable, IMO.

[jfgout]

My assumption is that the layers cover the same FOV (or approximately so), and were taken at different exposures in order to cover the wide dynamic range of the different planets. Or perhaps a combination of exposure time and position, optimized for the targets. Perfectly acceptable, IMO.

Chris,

That would have been acceptable indeed (although it is questionable to make Uranus & Neptune appear brighter than magnitude 3 stars...). But that's not even what happened here.

1) The stars around Uranus & Neptune should be visible then, or it should be said explicitly that they were digitally removed (the layers containing Uranus & Neptune should be full of stars down to magnitude ~8)

2) What they did for Mercury is that they simply added something (who knows if it's actually Mercury, a star ?) at a position which is not even where Mercury truly was!!!!

One more thing. Look carefully at the full resolution image and you will see just above the horizon a pattern of sharp stars that looks like Cassiopeia repeated twice in the image. This tells you that they've incorporated the same layer at two different locations in their image, which is very different from what you described in your message.

The fact that most people do not see it shows that (1) they did a pretty good job at making this fake image and (2) most people do not pay attention to what the sky actually looks like. The wrong location of Mercury (way too close to Venus) was immediately visible to me, but I'm one of the lucky few who got to see the planets in the morning with my actual eyes. So, I could tell immediately that something was wrong here. For those who were not that lucky, see the screenshots posted by mars_s in the previous messages.

Again, I truly appreciate the service provided by APOD and I think it is our job as astronomy enthusiasts to help the editors spot this type of fakery.

Best,

jf

[mar_s]

Adjusting the brightnesses in this cherry picking way is misleading. Messing around with positions of planets and stars is not acceptable, IMO.

[Chris Peterson]

My assumption is that the layers cover the same FOV (or approximately so), and were taken at different exposures in order to cover the wide dynamic range of the different planets. Or perhaps a combination of exposure time and position, optimized for the targets. Perfectly acceptable, IMO.

Chris,

That would have been acceptable indeed (although it is questionable to make Uranus & Neptune appear brighter than magnitude 3 stars...). But that's not even what happened here.

1) The stars around Uranus & Neptune should be visible then, or it should be said explicitly that they were digitally removed (the layers containing Uranus & Neptune should be full of stars down to magnitude ~8)

2) What they did for Mercury is that they simply added something (who knows if it's actually Mercury, a star ?) at a position which is not even where Mercury truly was!!!!

One more thing. Look carefully at the full resolution image and you will see just above the horizon a pattern of sharp stars that looks like Cassiopeia repeated twice in the image. This tells you that they've incorporated the same layer at two different locations in their image, which is very different from what you described in your message.

The fact that most people do not see it shows that (1) they did a pretty good job at making this fake image and (2) most people do not pay attention to what the sky actually looks like. The wrong location of Mercury (way too close to Venus) was immediately visible to me, but I'm one of the lucky few who got to see the planets in the morning with my actual eyes. So, I could tell immediately that something was wrong here. For those who were not that lucky, see the screenshots posted by mars_s in the previous messages.

Again, I truly appreciate the service provided by APOD and I think it is our job as astronomy enthusiasts to help the editors spot this type of fakery.

Best,

jf

If I were doing this. I'd draw a tiny circle around the dimmer planets and delete everything outside that before stacking the frames. So I'd have no stars in those images. I haven't looked closely enough to assess the claim about Mercury's position. It is possible that there were processing errors. I think calling it a "fake image" is extreme and not well supported.

[jfgout]

It is possible that there were processing errors. I think calling it a "fake image" is extreme and not well supported.

I did not make these claims lightly. No, these are not processing errors. The problem with Mercury was clearly illustrated by mars_s in some of the previous messages with overlays of Stellarium screenshots.

Now, I'll illustrate the other problem that I was mentioning. If you look at the full resolution image, you will find a mix of stars that come from the "main" image (they tend to be a bit fuzzy, distorted by lens aberrations and a little bit of motion) and small, sharp stars that come from a totally different picture, incrusted in the image twice.

Here is an example of two of these stars next to each other:



I took the time to circle all of these sharp stars in the image. You can see that they form a pattern that is repeated at least twice in the image. To make things really easy to see, I've even numbered some of the stars in the pattern. Make sure you look at the full resolution version here (impossible to see anything in the thumbnail format on the forum): https://photos.smugmug.com/photos/i-4MX ... MX2pwN.jpg



There is no processing error that will do this. This is dishonest digital editing.

I've invited the authors (on APOD's Facebook page) to provide more explanations about their processing, but I'm not holding my breath...

jf

[Chris Peterson]

It is possible that there were processing errors. I think calling it a "fake image" is extreme and not well supported.

I did not make these claims lightly. No, these are not processing errors. The problem with Mercury was clearly illustrated by mars_s in some of the previous messages with overlays of Stellarium screenshots.

Now, I'll illustrate the other problem that I was mentioning. If you look at the full resolution image, you will find a mix of stars that come from the "main" image (they tend to be a bit fuzzy, distorted by lens aberrations and a little bit of motion) and small, sharp stars that come from a totally different picture, incrusted in the image twice.

Here is an example of two of these stars next to each other:



I took the time to circle all of these sharp stars in the image. You can see that they form a pattern that is repeated at least twice in the image. To make things really easy to see, I've even numbered some of the stars in the pattern. Make sure you look at the full resolution version here (impossible to see anything in the thumbnail format on the forum): https://photos.smugmug.com/photos/i-4MX ... MX2pwN.jpg



There is no processing error that will do this. This is dishonest digital editing.

I've invited the authors (on APOD's Facebook page) to provide more explanations about their processing, but I'm not holding my breath...

jf

They look to me like hot pixels, not stars, and I'd expect that sort of repeating pattern.

That said, I've long advocated for APOD to refuse to accept any images that don't have full processing details available. But that hasn't happened, and doesn't seem to be a priority or interest of the editors.

[johnnydeep]

It is possible that there were processing errors. I think calling it a "fake image" is extreme and not well supported.

I did not make these claims lightly. No, these are not processing errors. The problem with Mercury was clearly illustrated by mars_s in some of the previous messages with overlays of Stellarium screenshots.

Now, I'll illustrate the other problem that I was mentioning. If you look at the full resolution image, you will find a mix of stars that come from the "main" image (they tend to be a bit fuzzy, distorted by lens aberrations and a little bit of motion) and small, sharp stars that come from a totally different picture, incrusted in the image twice.

Here is an example of two of these stars next to each other:



I took the time to circle all of these sharp stars in the image. You can see that they form a pattern that is repeated at least twice in the image. To make things really easy to see, I've even numbered some of the stars in the pattern. Make sure you look at the full resolution version here (impossible to see anything in the thumbnail format on the forum): https://photos.smugmug.com/photos/i-4MX ... MX2pwN.jpg



There is no processing error that will do this. This is dishonest digital editing.

I've invited the authors (on APOD's Facebook page) to provide more explanations about their processing, but I'm not holding my breath...

jf

They look to me like hot pixels, not stars, and I'd expect that sort of repeating pattern.

That said, I've long advocated for APOD to refuse to accept any images that don't have full processing details available. But that hasn't happened, and doesn't seem to be a priority or interest of the editors.

Huh - "pot pixel". Another temp I was unfamiliar with. And after searching for some other images of what they might look like, I think that's what they are! For example, see this imaged that was on Reddit (but for which the page link doesn't seem to work any more):

Click to view full size image

Here's a close-up of two of them, that happen to be blue, but that have that same rectangular - almost camera body shaped! - outline as the ones pointed out in this APOD by jfgout:

[ EDIT: ok, here's the link where the pic appeared - https://www.reddit.com/r/photography/co ... _fix_this/. It's extremely slow to respond, or else it's just my browser. ]

[jfgout]

They look to me like hot pixels, not stars, and I'd expect that sort of repeating pattern.

That said, I've long advocated for APOD to refuse to accept any images that don't have full processing details available. But that hasn't happened, and doesn't seem to be a priority or interest of the editors.

Hot pixels are single pixels which will appear as red, green or blue on the image. These are (small) lumps of white pixels, with varying intensity (note the little black edging too, typical of stars that went through some sharpening/wavelet processing).

The only I can think off that hot pixels would look like that is if they are in a layer that has been scaled and then blended with another. So, maybe it could indeed be hot/dead pixels. Does not change the conclusion that the image was manipulated beyond what is acceptable.

jf

[Chris Peterson]

They look to me like hot pixels, not stars, and I'd expect that sort of repeating pattern.

That said, I've long advocated for APOD to refuse to accept any images that don't have full processing details available. But that hasn't happened, and doesn't seem to be a priority or interest of the editors.

Hot pixels are single pixels which will appear as red, green or blue on the image. These are (small) lumps of white pixels, with varying intensity (note the little black edging too, typical of stars that went through some sharpening/wavelet processing).

jf

How hot pixels appear depends very much on the processing and on how one-shot-color cameras handle them internally, even before we get the "raw" images out.

You are reading way too much into your analysis. All you have is a JPEG image, which means the sort of halos and edges you're looking at are normal and unavoidable. There's no way here to reliably distinguish stars from hot pixels and from various JPEG artifacts.

[jfgout]

How hot pixels appear depends very much on the processing and on how one-shot-color cameras handle them internally, even before we get the "raw" images out.

You are reading way too much into your analysis. All you have is a JPEG image, which means the sort of halos and edges you're looking at are normal and unavoidable. There's no way here to reliably distinguish stars from hot pixels and from various JPEG artifacts.

Yes, post-processing/jpeg conversion and the like can cause hot/dead pixels to look differently. Maybe these are dead pixels after all. Still, I would very much like to know why there are true hot pixels just next to them???



Full resolution: https://photos.smugmug.com/photos/i-3DL ... DLghr7.jpg

Sure, they used 4 cameras, so maybe different cameras produce different patterns of hot/dead pixels. Maybe even a single camera produces these two patterns. That's why it would be nice if the authors of the image could provide some actual useful information about how they produce this composite thing. But considering how heavily this image has been manipulated (Uranus & Neptune 50+ times too bright, Mercury not even in the correct location), I doubt we will ever get a honest answer from them.

jf

[Chris Peterson]

How hot pixels appear depends very much on the processing and on how one-shot-color cameras handle them internally, even before we get the "raw" images out.

You are reading way too much into your analysis. All you have is a JPEG image, which means the sort of halos and edges you're looking at are normal and unavoidable. There's no way here to reliably distinguish stars from hot pixels and from various JPEG artifacts.

Yes, post-processing/jpeg conversion and the like can cause hot/dead pixels to look differently. Maybe these are dead pixels after all. Still, I would very much like to know why there are true hot pixels just next to them???



Full resolution: https://photos.smugmug.com/photos/i-3DL ... DLghr7.jpg

Sure, they used 4 cameras, so maybe different cameras produce different patterns of hot/dead pixels. Maybe even a single camera produces these two patterns. That's why it would be nice if the authors of the image could provide some actual useful information about how they produce this composite thing. But considering how heavily this image has been manipulated (Uranus & Neptune 50+ times too bright, Mercury not even in the correct location), I doubt we will ever get a honest answer from them.

jf

I certainly think there is something wrong with the image. The ratios of the distances between the planets in the image is very inconsistent with their actual positions that morning. That's problematic. I just prefer to assume it's a processing problem, as opposed to some kind of deliberate fake (not least because it's very incompetent as a fake... I'd expect better). The issue of brightness does not concern me. The issue of position does.

[jfgout]

I certainly think there is something wrong with the image. The ratios of the distances between the planets in the image is very inconsistent with their actual positions that morning. That's problematic. I just prefer to assume it's a processing problem, as opposed to some kind of deliberate fake (not least because it's very incompetent as a fake... I'd expect better). The issue of brightness does not concern me. The issue of position does.

Fair enough. I 100% agree that the position issue is the biggest problem here. While I don't believe it is the case here, there is indeed a (small) chance that this was all just processing error.

I've invited (on facebook) the authors to come explain their process here. I hope they will.

jf

[GAP]

We are sorry to have created all this havoc, but the result of all this is that first of all 4 different rooms were used. Also, the photo is not a single photo but a composition of 10 panels, so the position of the planets may be slightly distorted.
Finally, the planets were photographed a few hours apart deliberately to get all the planets you saw in the final composition.

We hope we have clarified

The authors

[jfgout]

We are sorry to have created all this havoc, but the result of all this is that first of all 4 different rooms were used. Also, the photo is not a single photo but a composition of 10 panels, so the position of the planets may be slightly distorted.
Finally, the planets were photographed a few hours apart deliberately to get all the planets you saw in the final composition.

We hope we have clarified

The authors

Thank you for the additional information. I appreciate you taking the time to join the conversation.

Still no actual explanation as to how Uranus and Neptune were incrusted in the image, or how Mercury was included, but that's not really a surprise. I guess that saying "we took pictures of Uranus and Neptune one hour before the main image, digitally removed all the stars from these images and inserted them roughly at the position the planets should be in the main image" does not sound nearly as cool as saying that it is a "cooperative astro-panorama".

This being said, thank you also for your honesty, acknowledging that:

Finally, the planets were photographed a few hours apart deliberately to get all the planets you saw in the final composition.

Of course, one might wonder how this squares with the original description:

Simultaneous images from four cameras were combined to construct this atmospheric predawn skyscape.

But I'm sure that, just like Mercury miraculously appearing just above the clouds was just a distortion problem, the simultaneous images that ended up being taken a few hours apart is just a translation error. Only mean people would think that there was some sort of intentional deception going on here.

Best,

jf



jf

[GAP]

Nothing intensional and bad, we wanted to portray the planets in one picture and so we constructed it as explained

Thank you very much

The Authors