APOD: When Roses Aren't Red (2018 Feb 22)

[APOD Robot]

When Roses Aren't Red

Explanation: Not all roses are red of course, but they can still be very pretty. Likewise, the beautiful Rosette Nebula and other star forming regions are often shown in astronomical images with a predominately red hue, in part because the dominant emission in the nebula is from hydrogen atoms. Hydrogen's strongest optical emission line, known as H-alpha, is in the red region of the spectrum, but the beauty of an emission nebula need not be appreciated in red light alone. Other atoms in the nebula are also excited by energetic starlight and produce narrow emission lines as well. In this gorgeous view of the Rosette Nebula, narrowband images are combined to show emission from sulfur atoms in red, hydrogen in blue, and oxygen in green. In fact, the scheme of mapping these narrow atomic emission lines into broader colors is adopted in many Hubble images of stellar nurseries. The image spans about 100 light-years in the constellation Monoceros, at the 3,000 light-year estimated distance of the Rosette Nebula. To make the Rosette red, just follow this link or slide your cursor over the image.

<< Previous APOD

This Day in APOD

Next APOD >>

[/b]

[heehaw]

Alice through the looking glass!

[Ann]

Alice through the looking glass!

Everything is probably mapped color down the rabbit hole!

Ann

[Fred the Cat]

When roses aren’t red
And violets aren’t blue
Finally, it’s said
The second line’s true

With oxygen’s green
Plus blue hy-dro-gen
Make Earth look from space
A Balmer cyan

[Chris Peterson]

This very nicely demonstrates how much more information we can extract about many objects by using narrowband filters and choosing the appropriate color mapping. In the "true color" version of this object the intense emission from hydrogen hides a huge amount of information. With a realistic image, we miss most of the picture.

[coyote@alum.mit.edu]

Blue roses, aka "pleurosis," (a once-fashionable old-time term for pleurisy) was a subject in the Tennessee Williams play, Glass Menagerie....

[neufer]

Blue roses, aka "pleurosis," (a once-fashionable old-time term for pleurisy) was a subject in the Tennessee Williams play, Glass Menagerie....

<<The Glass Menagerie is a memory play by Tennessee Williams (March 26, 1911 – February 25, 1983) that premiered in 1944 and catapulted Williams from obscurity to fame. The characters and story mimic Williams's own life more closely than any of his other works. Williams (whose real name was Thomas) closely resembles Tom; his mother, Amanda. His sickly and mentally unstable older sister Rose provides the basis for the fragile Laura (whose nickname in the play is "Blue Roses", a result of a bout of pleurosis as a high school student), though it has also been suggested that Laura may incorporate aspects of Williams himself, referencing his introverted nature and obsessive focus on just one aspect of life (writing for Williams and glass animals in Laura's case). Williams, who was close to Rose growing up, learned to his horror that in 1943 in his absence his sister had been subjected to a botched lobotomy. Rose was left incapacitated (and institutionalized) for the rest of her life. Eventually Williams was to leave the bulk of his estate to ensure Rose's continuing care. On February 25, 1983, Williams was found dead in his suite at the Hotel Elysée in New York at age 71. Williams had choked to death from inhaling the plastic cap of a bottle of the type that might contain a nasal spray or eye solution. Rose died in 1996.>>

[Ann]

This very nicely demonstrates how much more information we can extract about many objects by using narrowband filters and choosing the appropriate color mapping. In the "true color" version of this object the intense emission from hydrogen hides a huge amount of information. With a realistic image, we miss most of the picture.

It irritates me that hydrogen is mapped as blue, oxygen as green and sulfur as red. There is no logic there. Clearly the Hubble palette, where oxygen is blue, hydrogen is green and sulfur is red is a lot more logical.

Or does the hydrogen-mapped-as-blue refer to Hβ, hydrogen beta, which really is a tad bluer than OIII?

Ann

[Chris Peterson]

This very nicely demonstrates how much more information we can extract about many objects by using narrowband filters and choosing the appropriate color mapping. In the "true color" version of this object the intense emission from hydrogen hides a huge amount of information. With a realistic image, we miss most of the picture.

It irritates me that hydrogen is mapped as blue, oxygen as green and sulfur as red. There is no logic there. Clearly the Hubble palette, where oxygen is blue, hydrogen is green and sulfur is red is a lot more logical.

Why? It's not like our eyes naturally make any connection between color and the concept of wavelength. The point of the narrowband imaging is to isolate details of elemental composition. What difference does it make whether we align the assigned colors in the same wavelength order as the source?

Here's the same image with the blue and green channels swapped so it matches the conventional Hubble palette. Is it any better? Do we see more or less? Is it more intuitive what we're seeing. I don't think so.

[Ann]

I don't see any "blue OIII" in your picture, Chris, and I don't see any "green OIII" in the APOD.

Click to view full size image

The Rosette Nebula in the Hubble palette.
Photo: HST.

Does that mean that there is no OIII in the Rosette Nebula? I find that too improbable to even consider.

The "classic Hubble palette image" at left does show the presence of OIII in the Rosette. It is very much harder to spot any OIII in today's APOD.

Ann

[Chris Peterson]

I don't see any "blue OIII" in your picture, Chris, and I don't see any "green OIII" in the APOD.

Well, you wouldn't necessarily, since our eyes don't record wavelength. Colors are the product of how wavelengths mix. The only way you'd see pure red, green, or blue would be if you had a region with no other emissions.

UPDATE: If you want to approach a scientific understanding, you need to separate the channels. Based on Eric's later comment that this actually is the Hubble palette and the caption is wrong, we can look at the individual elemental components as shown below (keeping in mind that a luminance channel was synthesized for largely aesthetic purposes, and it slightly contaminates the individual color channels by mixing them a little bit. Still, this is probably a pretty good approximation of the actual composition. You can see a central dense ring of O[III], but it overlies a fairly dense Ha region, which is why the central zone appears cyan, rather than green or blue, in the color composite. In the outer regions, the more diffuse S[II] overlaps Ha, which pushes the red towards orange.

[geckzilla]

The whole reason the issue arises is because and H-alpha emissions occupy virtually the same place visually for humans. has a strong spectral line at 673 nm, while H-alpha sits around 656 nm. If you really want to complicate things, throw in an [N II] filter at around 658 nm (also great for H-alpha in galaxies redshifted just a tad)

Anyway, it's up to the image processor to decide how to work the data together. When the wavelengths don't necessarily make sense to assign exactly to their corresponding visual color, sure, there's room for interpretation, and quite often things end up looking one way or another because the processor decided it looked better.

[geckzilla]

Examining the image further, however, I think that the description may be wrongly identifying what filters were assigned to each channel.

[Chris Peterson]

Examining the image further, however, I think that the description may be wrongly identifying what filters were assigned to each channel.

I was suspicious, as well. First, because the image where I swapped the green and blue channels looks nothing like the Hubble palette, but the original image does. Second, because the image caption specifically references the Hubble palette in a link.

[ta152h0]

Nice era of discovery to be living in. One hour of " How the Universe Works " an Mercury hit the Earth and Thea did the same. What is next ? A red convertible in space ?

[geckzilla]

Examining the image further, however, I think that the description may be wrongly identifying what filters were assigned to each channel.

I was suspicious, as well. First, because the image where I swapped the green and blue channels looks nothing like the Hubble palette, but the original image does. Second, because the image caption specifically references the Hubble palette in a link.

I'm going to email the editors, and suggest that they double check.

[coles44]

I am reading some of the comments on today's APOD. Let me clarify a couple of items.
1. In the Hubble Palette image the NB filters are assigned as follow. H=green, S=red and O=blue.
2. The Hubble Palette image is a starless tone map with a luminosity layer produced from the H image with some contribution from S and O starless images.
3. The HRGB image is an RGB image where most of the detail comes from the H image using high-pass filtering.
4. Doing it that way makes the detail in both images equivalent, even though the color palettes are totally different.
If you have any additional questions about the image. Please let me know.
Eric Coles

[MarkBour]

It irritates me that hydrogen is mapped as blue, oxygen as green and sulfur as red. There is no logic there. Clearly the Hubble palette, where oxygen is blue, hydrogen is green and sulfur is red is a lot more logical. . . .
Ann

Ah. The suggested pallette does not suit thy palate?

Who seeks for better of thee, sauce his palate
With thy most operant poison! What is here?
Gold? yellow, glittering, precious gold? No, gods,
I am no idle votarist: roots, you clear heavens!

― William Shakespeare, Timon of Athens

[neufer]

It irritates me that hydrogen is mapped as blue, oxygen as green and sulfur as red. There is no logic there. Clearly the Hubble palette, where oxygen is blue, hydrogen is green and sulfur is red is a lot more logical. . . .

Ah. The suggested pallette does not suit thy palate?

Who seeks for better of thee, sauce his palate
With thy most operant poison! What is here?
Gold? yellow, glittering, precious gold? No, gods,
I am no idle votarist: roots, you clear heavens!

― William Shakespeare, Timon of Athens

palette (n.) 1620s, "flat thin tablet used by an artist to lay and mix colors," from French palette, from Old French palete "small shovel, blade" (13c.) diminutive of pale "shovel, blade," from Latin pala "spade, shoulder blade." Transferred sense of "colors used by a particular artist" is from 1882.

palate (n.) late 14c., "roof of the mouth," from Old French palat and directly from Latin palatum "roof of the mouth," perhaps of Etruscan origin [Klein]. Popularly considered the seat of taste, hence transferred meaning "sense of taste" (late 14c.), which also was in classical Latin. Related: Palatal; palatalize.

[Ann]

It irritates me that hydrogen is mapped as blue, oxygen as green and sulfur as red. There is no logic there. Clearly the Hubble palette, where oxygen is blue, hydrogen is green and sulfur is red is a lot more logical. . . .
Ann

Ah. The suggested pallette does not suit thy palate?

Who seeks for better of thee, sauce his palate
With thy most operant poison! What is here?
Gold? yellow, glittering, precious gold? No, gods,
I am no idle votarist: roots, you clear heavens!

― William Shakespeare, Timon of Athens

Click to view full size image

Rosette Nebula in infrared. NASA/JPL-Caltech/WISE Team

I want mapped color to be logical! Imagine an infrared picture where blue is mid-infrared, green is near-infrared and red is far-infrared. Or imagine that while blue is near-infrared, green is far-infrared and red is mid-infrared. That would be highly confusing.

Consider the picture at left of the Rosette Nebula in infrared. In this picture, blue represents 3.4 and 4.6 microns, green 12 microns and red 22 microns. That's logical. Surprisingly (to me), mid-infrared emission dominates at the outer parts of the Rosette, while far-infrared dominates closer to the central "hole". Thanks to the logical mapped color in this image, I can immediately spot this surprising infrared temperature distribution in the Rosette Nebula and ask myself what causes it. It would seem that the red areas are fairly "empty" (of dust), while dust piles up in the green parts. Clearly that has something to do with it, although I couldn't say exactly how it works.

In mapped color images, I just want blue to represent the shortest wavelengths and red the longest ones!

Ann

[Chris Peterson]

Consider the picture at left of the Rosette Nebula in infrared. In this picture, blue represents 3.4 and 4.6 microns, green 12 microns and red 22 microns. That's logical. Surprisingly (to me), mid-infrared emission dominates at the outer parts of the Rosette, while far-infrared dominates closer to the central "hole". Thanks to the logical mapped color in this image, I can immediately spot this surprising infrared temperature distribution in the Rosette Nebula and ask myself what causes it. It would seem that the red areas are fairly "empty" (of dust), while dust piles up in the green parts. Clearly that has something to do with it, although I couldn't say exactly how it works.

In mapped color images, I just want blue to represent the shortest wavelengths and red the longest ones!

That makes some sense in your example, which consists of multiple, overlapping, wideband channels. That's basically a color image that you're just shifting up into a different range.

But that isn't the case for a narrowband image like today's APOD. In that case it isn't a color image at all. The actual colors of the filters are meaningless. They don't represent colors, but elements. We don't care that O[III] emits light at a shorter wavelength than S[II]. All we care about is the amount and location of oxygen and sulfur. So the choice of colors we assign to them is arbitrary, subject only to the combination that seems to allow us to see the most detail.

[Ann]

I don't see any "blue OIII" in your picture, Chris, and I don't see any "green OIII" in the APOD.

Well, you wouldn't necessarily, since our eyes don't record wavelength. Colors are the product of how wavelengths mix. The only way you'd see pure red, green, or blue would be if you had a region with no other emissions.

UPDATE: If you want to approach a scientific understanding, you need to separate the channels. Based on Eric's later comment that this actually is the Hubble palette and the caption is wrong, we can look at the individual elemental components as shown below (keeping in mind that a luminance channel was synthesized for largely aesthetic purposes, and it slightly contaminates the individual color channels by mixing them a little bit. Still, this is probably a pretty good approximation of the actual composition. You can see a central dense ring of O[III], but it overlies a fairly dense Ha region, which is why the central zone appears cyan, rather than green or blue, in the color composite. In the outer regions, the more diffuse S[II] overlaps Ha, which pushes the red towards orange.

RosetteNebulaNBHColesHelmRGB.jpg

Thanks, Chris (and Eric)! That really helps a lot!

Ann

[geckzilla]

There have been rare instances in my work where I have veered from chromatic ordering (what Ann calls "common sense") and it usually has to do with combining very different wavelengths. I have come to understand that the electromagnetic spectrum has many repetitive patterns in it, and that there are times when it makes sense to overlap them instead of putting them all in "single file" sequence where the patterns begin to repeat. Kind of interesting.