Three-dimensional volumetric model of nebulaes

Dear members of sci.astro.research,

I'm familiar with one of the outcome of Mr. C.R.O'Dell
and Wen Zhang on the Orion nebula, a stunning rendering
and fly-through of this beautiful object.

Somehow though, while I found plenty of information on how
the visualization has been done, I've been wondering how
the three-dimensional source data has been retrived in the
first place.

I did search around, and there's plenty of bibliographic
references to O'Dell and Zhang paper published on the Astrophysic
Journal, but I couldn't actually found any pubblicly browsable
full-text version of it.

So, unless somebody is kind enough to provide a link to
such document, could anybody explain me how such 3d
reconstruction is made? Which basically is all about:
how do you measure the distance/distribution/
tickness of the volume "behind" each pixel composing a
picture of a nebula?

Is the parallax method used to determine star distance
capable of dealing with fuzzy object like the features
of a nebula? Is spectrography involved instead?

One more question: has any other nebula been mapped
and visualized with similar techniques?

Thanks for your help.

Manu

Emanuele D'Arrigo wrote:
Dear members of sci.astro.research,

I'm familiar with one of the outcome of Mr. C.R.O'Dell
and Wen Zhang on the Orion nebula, a stunning rendering
and fly-through of this beautiful object.

Somehow though, while I found plenty of information on how
the visualization has been done, I've been wondering how
the three-dimensional source data has been retrived in the
first place.

I did search around, and there's plenty of bibliographic
references to O'Dell and Zhang paper published on the Astrophysic
Journal, but I couldn't actually found any pubblicly browsable
full-text version of it.

So, unless somebody is kind enough to provide a link to
such document, could anybody explain me how such 3d
reconstruction is made? Which basically is all about:
how do you measure the distance/distribution/
tickness of the volume "behind" each pixel composing a
picture of a nebula?

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud (not much like the Stromgren
spheres we used to learn about...). This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star, most other things such as density
cancelling when we see the integration through the layer's
thickness. It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

Bill Keel

Emanuele D'Arrigo wrote:
Dear members of sci.astro.research,

I'm familiar with one of the outcome of Mr. C.R.O'Dell
and Wen Zhang on the Orion nebula, a stunning rendering
and fly-through of this beautiful object.

Somehow though, while I found plenty of information on how
the visualization has been done, I've been wondering how
the three-dimensional source data has been retrived in the
first place.

I did search around, and there's plenty of bibliographic
references to O'Dell and Zhang paper published on the Astrophysic
Journal, but I couldn't actually found any pubblicly browsable
full-text version of it.

So, unless somebody is kind enough to provide a link to
such document, could anybody explain me how such 3d
reconstruction is made? Which basically is all about:
how do you measure the distance/distribution/
tickness of the volume "behind" each pixel composing a
picture of a nebula?

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud (not much like the Stromgren
spheres we used to learn about...). This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star, most other things such as density
cancelling when we see the integration through the layer's
thickness. It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

Bill Keel

Bill, thanks for your reply.

"William C. Keel" wrote in message
...

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud

So, from what I understand what we see not only
in the animation but also in pictures of other nebulaes,
aren't actually volumes of space, but more like boundaries,
thin shells where the solar wind of the central star(s)
is colliding and pushing the surrounding gas. Is my
interpretation of your words close enough?

This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star,

....mmm... this doesn't seem enough to determine the three-
dimentional structure of the object though. Geometrically
speaking, you suggest that the brightness of a single pixel
in a picture gives us a distance measurement from the star,
which can be rapresented as a sphere around the star.
We then have a line connecting the observer to such pixel.
The intersection of a line and a sphere can have three results:
the line is completely out of the sphere (which we can exclude
because we know that that patch of the nebula is on that sphere),
the line is tangent to the sphere (which is the only unambiguous
case), the line pass trough the sphere, intersecting the sphere
in two points, one closer, one further from the observer.
My (possibly wrong) guess is that the third case is the most
recurring, but another measurement is necessary to discriminate
between the two options. What am I missing?

most other things such as density
cancelling when we see the integration through the layer's
thickness.

Could you please explain this a bit more? I understand
from your words that density can be disregarded but I
don't quite understand how the thickness is measured.
Or how two or more layers can be separated. Does it
have to do with the width or the multiple occurrence
of the same H II emission lines?

It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

I'd like to offer bit of "hope" challenging your belief about this
from somebody who's pretty deep in the 3d business:
for what I understand of how the animation
has been realized (and that's the part I understand fairly well),
I believe that currently available commercial softwares and computers
are well capable of achieving the same visual quality achieved by the
team who did it. Indeed the team used a super-computing center
for the rendering, but that had to do with the media used to
deliver the animation the audience, a large screen and a level of
detail about 20 times the level of detail avalaible on a tv screen.

Given the proper 3d data, a skilled 3d artist with the help of 3d
softwares like Maya (www.alias.com) or Softimage XSI (www.softimage.com),
and a professional workstation (each well under 5000 USD and in reach
of even the average professional in the field) can relatively easily
produce results that are as accurate ass the data is and as visually
stunning as the Orion Nebula animation.
Of course the result generated by a single computer in a reasonable amount
of time would be good enough only for DVDs and VHS, possibly for High
Definition TV, but I doubt most people would want to produce an output
good enough for the Hayden Planetarium, anyway. And if the need arise,
it takes only relatively few more computers and a reasonable additional
amount of work on the 3d artist's side to adjust the 3d scene to the new
requirements.

I'm honestly hoping I'm sowing some seeds he nebulaes are too
beautiful to leave them as flat objects. I wish we'll see more and
more 3d visualization of them.

Bill, thanks again for your reply.

Manu

Bill, thanks for your reply.

"William C. Keel" wrote in message
...

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud

So, from what I understand what we see not only
in the animation but also in pictures of other nebulaes,
aren't actually volumes of space, but more like boundaries,
thin shells where the solar wind of the central star(s)
is colliding and pushing the surrounding gas. Is my
interpretation of your words close enough?

This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star,

....mmm... this doesn't seem enough to determine the three-
dimentional structure of the object though. Geometrically
speaking, you suggest that the brightness of a single pixel
in a picture gives us a distance measurement from the star,
which can be rapresented as a sphere around the star.
We then have a line connecting the observer to such pixel.
The intersection of a line and a sphere can have three results:
the line is completely out of the sphere (which we can exclude
because we know that that patch of the nebula is on that sphere),
the line is tangent to the sphere (which is the only unambiguous
case), the line pass trough the sphere, intersecting the sphere
in two points, one closer, one further from the observer.
My (possibly wrong) guess is that the third case is the most
recurring, but another measurement is necessary to discriminate
between the two options. What am I missing?

most other things such as density
cancelling when we see the integration through the layer's
thickness.

Could you please explain this a bit more? I understand
from your words that density can be disregarded but I
don't quite understand how the thickness is measured.
Or how two or more layers can be separated. Does it
have to do with the width or the multiple occurrence
of the same H II emission lines?

It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

I'd like to offer bit of "hope" challenging your belief about this
from somebody who's pretty deep in the 3d business:
for what I understand of how the animation
has been realized (and that's the part I understand fairly well),
I believe that currently available commercial softwares and computers
are well capable of achieving the same visual quality achieved by the
team who did it. Indeed the team used a super-computing center
for the rendering, but that had to do with the media used to
deliver the animation the audience, a large screen and a level of
detail about 20 times the level of detail avalaible on a tv screen.

Given the proper 3d data, a skilled 3d artist with the help of 3d
softwares like Maya (www.alias.com) or Softimage XSI (www.softimage.com),
and a professional workstation (each well under 5000 USD and in reach
of even the average professional in the field) can relatively easily
produce results that are as accurate ass the data is and as visually
stunning as the Orion Nebula animation.
Of course the result generated by a single computer in a reasonable amount
of time would be good enough only for DVDs and VHS, possibly for High
Definition TV, but I doubt most people would want to produce an output
good enough for the Hayden Planetarium, anyway. And if the need arise,
it takes only relatively few more computers and a reasonable additional
amount of work on the 3d artist's side to adjust the 3d scene to the new
requirements.

I'm honestly hoping I'm sowing some seeds he nebulaes are too
beautiful to leave them as flat objects. I wish we'll see more and
more 3d visualization of them.

Bill, thanks again for your reply.

Manu

Emanuele D'Arrigo wrote:
Bill, thanks for your reply.

"William C. Keel" wrote in message
...

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud

So, from what I understand what we see not only
in the animation but also in pictures of other nebulaes,
aren't actually volumes of space, but more like boundaries,
thin shells where the solar wind of the central star(s)
is colliding and pushing the surrounding gas. Is my
interpretation of your words close enough?

Pretty much. With the usual weasel words like "most", "dominated by",
it seems we can consider many H II regions (not planetary nebulae,
for example) as emitting from the thin layer where the UV radiation
from the ionizing star is just now ionizing the dense molecular
cloud that formed the star. This is a more dynamic situation than
we used to consider, one of a statis balance between ionization and
recombination throughout the volume of a spherical nebula...

This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star,

...mmm... this doesn't seem enough to determine the three-
dimentional structure of the object though. Geometrically
speaking, you suggest that the brightness of a single pixel
in a picture gives us a distance measurement from the star,
which can be rapresented as a sphere around the star.
We then have a line connecting the observer to such pixel.
The intersection of a line and a sphere can have three results:
the line is completely out of the sphere (which we can exclude
because we know that that patch of the nebula is on that sphere),
the line is tangent to the sphere (which is the only unambiguous
case), the line pass trough the sphere, intersecting the sphere
in two points, one closer, one further from the observer.
My (possibly wrong) guess is that the third case is the most
recurring, but another measurement is necessary to discriminate
between the two options. What am I missing?

We already have two components of the 3D separation between a given spot
of nebula and the star, knowing that the spot lies along the vector between
us and its apparent position, so the distance from the star specifies
which of two positions it must occupy. For usual Milky Way clouds,
the star and nebula are optically obscured unless they lie on the
front side of the cloud. For the M42 animation, they could actually
resolve a thin foreground veil from Doppler shifts of absorption
lines.

most other things such as density
cancelling when we see the integration through the layer's
thickness.

Could you please explain this a bit more? I understand
from your words that density can be disregarded but I
don't quite understand how the thickness is measured.
Or how two or more layers can be separated. Does it
have to do with the width or the multiple occurrence
of the same H II emission lines?

The first-order point is that in getting the distance from the star, the
thickness of the emitting layer is small enough not to matter. At a
guess, they may have taken a representative calculated value and used
that to "fuzz" the layer.

It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

I'd like to offer bit of "hope" challenging your belief about this
from somebody who's pretty deep in the 3d business:
for what I understand of how the animation
has been realized (and that's the part I understand fairly well),
I believe that currently available commercial softwares and computers
are well capable of achieving the same visual quality achieved by the
team who did it. Indeed the team used a super-computing center
for the rendering, but that had to do with the media used to
deliver the animation the audience, a large screen and a level of
detail about 20 times the level of detail avalaible on a tv screen.

Given the proper 3d data, a skilled 3d artist with the help of 3d
softwares like Maya (www.alias.com) or Softimage XSI (www.softimage.com),
and a professional workstation (each well under 5000 USD and in reach
of even the average professional in the field) can relatively easily
produce results that are as accurate ass the data is and as visually
stunning as the Orion Nebula animation.
Of course the result generated by a single computer in a reasonable amount
of time would be good enough only for DVDs and VHS, possibly for High
Definition TV, but I doubt most people would want to produce an output
good enough for the Hayden Planetarium, anyway. And if the need arise,
it takes only relatively few more computers and a reasonable additional
amount of work on the 3d artist's side to adjust the 3d scene to the new
requirements.

I'm honestly hoping I'm sowing some seeds he nebulaes are too
beautiful to leave them as flat objects. I wish we'll see more and
more 3d visualization of them.

Hmm, $5000 software isn't routinely in the reach of a professional in
_my_ field - we gripe about IDL for rather less. Still, this is good
news, not just for easthetic reasons but because there are interesting
bits of interpretation that sometimes jump out when we can actually
rotate 3D visualizations. I was oding this with low-resolution
versions of flat galaxy models for comparison with data some years
ago, and that was helpful enough already.

Bill Keel

Emanuele D'Arrigo wrote:
Bill, thanks for your reply.

"William C. Keel" wrote in message
...

From O"Dell's talk at the last AAS meeting, a key point is the
recognition that much (most?) of the hydrogen line emission
arises in a thin skin where the ionization front is eating into
the surrounding molecular cloud

So, from what I understand what we see not only
in the animation but also in pictures of other nebulaes,
aren't actually volumes of space, but more like boundaries,
thin shells where the solar wind of the central star(s)
is colliding and pushing the surrounding gas. Is my
interpretation of your words close enough?

Pretty much. With the usual weasel words like "most", "dominated by",
it seems we can consider many H II regions (not planetary nebulae,
for example) as emitting from the thin layer where the UV radiation
from the ionizing star is just now ionizing the dense molecular
cloud that formed the star. This is a more dynamic situation than
we used to consider, one of a statis balance between ionization and
recombination throughout the volume of a spherical nebula...

This implies that the
surface brightness of a piece of this layer depends on its
distance from the ionizing star,

...mmm... this doesn't seem enough to determine the three-
dimentional structure of the object though. Geometrically
speaking, you suggest that the brightness of a single pixel
in a picture gives us a distance measurement from the star,
which can be rapresented as a sphere around the star.
We then have a line connecting the observer to such pixel.
The intersection of a line and a sphere can have three results:
the line is completely out of the sphere (which we can exclude
because we know that that patch of the nebula is on that sphere),
the line is tangent to the sphere (which is the only unambiguous
case), the line pass trough the sphere, intersecting the sphere
in two points, one closer, one further from the observer.
My (possibly wrong) guess is that the third case is the most
recurring, but another measurement is necessary to discriminate
between the two options. What am I missing?

We already have two components of the 3D separation between a given spot
of nebula and the star, knowing that the spot lies along the vector between
us and its apparent position, so the distance from the star specifies
which of two positions it must occupy. For usual Milky Way clouds,
the star and nebula are optically obscured unless they lie on the
front side of the cloud. For the M42 animation, they could actually
resolve a thin foreground veil from Doppler shifts of absorption
lines.

most other things such as density
cancelling when we see the integration through the layer's
thickness.

Could you please explain this a bit more? I understand
from your words that density can be disregarded but I
don't quite understand how the thickness is measured.
Or how two or more layers can be separated. Does it
have to do with the width or the multiple occurrence
of the same H II emission lines?

The first-order point is that in getting the distance from the star, the
thickness of the emitting layer is small enough not to matter. At a
guess, they may have taken a representative calculated value and used
that to "fuzz" the layer.

It is this relation that allowed them to do a local
3D reconstruction. This is possible for many other H II regions
with similar data, although many more people can do the reconstruction
than the high-visual-quality animations at this point.

I'd like to offer bit of "hope" challenging your belief about this
from somebody who's pretty deep in the 3d business:
for what I understand of how the animation
has been realized (and that's the part I understand fairly well),
I believe that currently available commercial softwares and computers
are well capable of achieving the same visual quality achieved by the
team who did it. Indeed the team used a super-computing center
for the rendering, but that had to do with the media used to
deliver the animation the audience, a large screen and a level of
detail about 20 times the level of detail avalaible on a tv screen.

Given the proper 3d data, a skilled 3d artist with the help of 3d
softwares like Maya (www.alias.com) or Softimage XSI (www.softimage.com),
and a professional workstation (each well under 5000 USD and in reach
of even the average professional in the field) can relatively easily
produce results that are as accurate ass the data is and as visually
stunning as the Orion Nebula animation.
Of course the result generated by a single computer in a reasonable amount
of time would be good enough only for DVDs and VHS, possibly for High
Definition TV, but I doubt most people would want to produce an output
good enough for the Hayden Planetarium, anyway. And if the need arise,
it takes only relatively few more computers and a reasonable additional
amount of work on the 3d artist's side to adjust the 3d scene to the new
requirements.

I'm honestly hoping I'm sowing some seeds he nebulaes are too
beautiful to leave them as flat objects. I wish we'll see more and
more 3d visualization of them.

Hmm, $5000 software isn't routinely in the reach of a professional in
_my_ field - we gripe about IDL for rather less. Still, this is good
news, not just for easthetic reasons but because there are interesting
bits of interpretation that sometimes jump out when we can actually
rotate 3D visualizations. I was oding this with low-resolution
versions of flat galaxy models for comparison with data some years
ago, and that was helpful enough already.

Bill Keel