My Globular Cluster Package for Testing

[chris]

but I really don't see sprites looking like squares on your or my screenshots. The central agglomeration of sprites in your screenshot may accidentally look a bit square, but that's not generic.

Fridger,

The square-sprites I see are located at the periphery on my screenshot:

This is the same thing I mentioned, and which Fridger has fixed already. When generating the star sprite textures, it's essential that the border pixels are completely black. I've had similar troubles a few times when generating for stars or particle systems.

--Chris

[t00fri]

Thanks guys, you obviously got much better monitors than me . Despite Vincent's red arrows, I could not make out squares there. But I could in Chris' image higher up, and as he writes, this trivial oversight has been fixed in seconds .


@Chris,
have a look at sime nice GREEN stars from Hubble a couple posts up

Fridger

[chris]

Chris,

as you see, I am responding step-by-step to your comments:

Like this one:

The star colors look odd, particularly the green.

Well I don't think they are looking "odd" since the original 1777x1864 true-color TIF of M 80 has them, too:

Here are a few green stars from the Hubble photo that have precisely the same RGB composition as my green stars:



I have quite a bit of practice in transfering color profiles from photos /quantitatively/ to my code....Just compare the green color in Vincents screenshot above.

But how was the photo produced? What filters were used, and how were the filtered images combined to produce an RGB image? Here's an image of M13 where no green stars are visible:

http://en.wikipedia.org/wiki/Image:Mess ... ikiSky.jpg

More importantly, we know that the stars in a globular cluster have (roughly) blackbody spectra, and that no blackbody spectrum will appear green to the human eye. To keep the color of stars in globulars consistent with the color of other stars, you could use the temperature->RGB tables in starcolor.cpp.

--Chris

[t00fri]

Chris,

as you see, I am responding step-by-step to your comments:

Like this one:

The star colors look odd, particularly the green.

Well I don't think they are looking "odd" since the original 1777x1864 true-color TIF of M 80 has them, too:

Here are a few green stars from the Hubble photo that have precisely the same RGB composition as my green stars:



I have quite a bit of practice in transfering color profiles from photos /quantitatively/ to my code....Just compare the green color in Vincents screenshot above.

But how was the photo produced? What filters were used, and how were the filtered images combined to produce an RGB image? Here's an image of M13 where no green stars are visible:

http://en.wikipedia.org/wiki/Image:Mess ... ikiSky.jpg

More importantly, we know that the stars in a globular cluster have (roughly) blackbody spectra, and that no blackbody spectrum will appear green to the human eye. To keep the color of stars in globulars consistent with the color of other stars, you could use the temperature->RGB tables in starcolor.cpp.

--Chris

Unfortunately, they don't reveal the "secrets of their image manipulations" to the public, I am afraid. All I did at this point was to transfer the spectrum from Hubble's M80 legacy photo into my code.

We know that M13 and some others are dominated by bluish stars. We also know that apart from that Hubble image, the color rendition of MOST globular photos is simply HORRIBLE. I have a VERY big collection with many amazing contradictions.

Certainly the color pattern of globulars differes a lot! That's why I want to proceed asap to include INDIVIDUAL color spectra as inferred from scientific catalogs.

Right now ( = principle of small steps ) all my globulars have the SAME, generic color profile as abstracted from Hubble. Since Hubble shows green stars, I show green stars

Why don't you write a little mail to Hubble's legacy administrators that their "green M 80 stars look odd" ?

Fridger

PS: I would appreciate a solid citation about your statement that globulars exhibit black body spectra with large accuracy. I know quite different stories...

[chris]

But how was the photo produced? What filters were used, and how were the filtered images combined to produce an RGB image? Here's an image of M13 where no green stars are visible:

http://en.wikipedia.org/wiki/Image:Mess ... ikiSky.jpg

More importantly, we know that the stars in a globular cluster have (roughly) blackbody spectra, and that no blackbody spectrum will appear green to the human eye. To keep the color of stars in globulars consistent with the color of other stars, you could use the temperature->RGB tables in starcolor.cpp.

--Chris

Unfortunately, they don't reveal the "secrets of their image manipulations" to the public, I am afraid. All I did at this point was to transfer the spectrum from Hubble's M80 legacy photo into my code.

We know that M13 and some others are dominated by bluish stars. We also know that apart from that Hubble image, the color rendition of MOST globular photos is simply HORRIBLE. I have a VERY big collection with many amazing contradictions.

Certainly the color pattern of globulars differes a lot! That's why I want to proceed asap to include INDIVIDUAL color spectra as inferred from scientific catalogs.

Right now ( = principle of small steps ) all my globulars have the SAME, generic color profile as abstracted from Hubble. Since Hubble shows green stars, I show green stars

I agree that using a single color profile for all globulars is a perfectly fine place to start.

Why don't you write a little mail to Hubble's legacy administrators that their "green M 80 stars look odd" ?

I don't think that would be productive. Perhaps a good thing to do is to use your image-derived color profile approach but start from the original FITS data. Would they contain enough image about the filters used for us to derive B-V values for the stars? From those, we could get colors. But I don't think that this is necessary for the first version; the globulars are looking very nice right now, so by all means check in the code if you're ready.

PS: I would appreciate a solid citation about your statement that globulars exhibit black body spectra with large accuracy. I know quite different stories...

My statement was that the individual stars in the globular cluster have blackbody colors. And while I know that star spectra deviate from perfect blackbody spectra because of absorption and emission lines, I'm unaware of any case where the star spectrum is modified to such a degree that it would appear green to the eye.

[t00fri]

But how was the photo produced? What filters were used, and how were the filtered images combined to produce an RGB image? Here's an image of M13 where no green stars are visible:

http://en.wikipedia.org/wiki/Image:Mess ... ikiSky.jpg

More importantly, we know that the stars in a globular cluster have (roughly) blackbody spectra, and that no blackbody spectrum will appear green to the human eye. To keep the color of stars in globulars consistent with the color of other stars, you could use the temperature->RGB tables in starcolor.cpp.

--Chris

Unfortunately, they don't reveal the "secrets of their image manipulations" to the public, I am afraid. All I did at this point was to transfer the spectrum from Hubble's M80 legacy photo into my code.

We know that M13 and some others are dominated by bluish stars. We also know that apart from that Hubble image, the color rendition of MOST globular photos is simply HORRIBLE. I have a VERY big collection with many amazing contradictions.

Certainly the color pattern of globulars differes a lot! That's why I want to proceed asap to include INDIVIDUAL color spectra as inferred from scientific catalogs.

Right now ( = principle of small steps ) all my globulars have the SAME, generic color profile as abstracted from Hubble. Since Hubble shows green stars, I show green stars

I agree that using a single color profile for all globulars is a perfectly fine place to start.

Why don't you write a little mail to Hubble's legacy administrators that their "green M 80 stars look odd" ?

I don't think that would be productive. Perhaps a good thing to do is to use your image-derived color profile approach but start from the original FITS data. Would they contain enough image about the filters used for us to derive B-V values for the stars? From those, we could get colors. But I don't think that this is necessary for the first version; the globulars are looking very nice right now, so by all means check in the code if you're ready.

PS: I would appreciate a solid citation about your statement that globulars exhibit black body spectra with large accuracy. I know quite different stories...

My statement was that the individual stars in the globular cluster have blackbody colors. And while I know that star spectra deviate from perfect blackbody spectra because of absorption and emission lines, I'm unaware of any case where the star spectrum is modified to such a degree that it would appear green to the eye.

Chris,

before checking in, could you have a look at my above "screenFrac - based" patch
viewtopic.php?f=4&t=12989&start=10
for improving the central part of a cluster at short distance, when travelling right through? Did you notice the corresponding (nice) video I made (you must download it first, I am afraid for reasons of speed)
viewtopic.php?f=4&t=12989&start=12

Apart from that, I also think we could be more productive, once that globular stuff is checked in. So people could start integrating the globulars into the GUI's etc. If there are any discrete improvements, that's easy to do also later via SVN.

Fridger

[ElChristou]

...Apart from that, I also think we could be more productive, once that globular stuff is checked in. So people could start integrating the globulars into the GUI's etc. If there are any discrete improvements, that's easy to do also later via SVN...

I agree, DW is waiting for the code to update the osX UI with the last new features...

[chris]

Chris,

before checking in, could you have a look at my above "screenfraction-based" patch
viewtopic.php?f=4&t=12989&start=10
for improving the central part of a cluster at short distance, when travelling right through? Did you notice the corresponding (nice) video I made (you must download it first, I am afraid for reasons of speed)
viewtopic.php?f=4&t=12989&start=12

Luckily, shatters.net is sitting under my desk, so the download was quick The patch definitely gives a big improvement in the appearance of globulars at close distances; please incorporate it in your checkin.

Apart from that, I also think we could be more productive, once that globular stuff is checked in. So people could start integrating the globulars into the GUI's etc. If there are any discrete improvements, that's easy to do also later via SVN.

Agreed.

(The globular changes and my significantly less exciting NormalizeMesh patch should be the last major code changes for 1.6.0.)

--Chris

[t00fri]

Chris,

of course we BOTH know why green stars cannot be seen within a VISUAL spectral sensitivity although lots of them are there ?

However, stars emit radiation over a broad range of wavelengths, and the human eye is most sensitive to yellow and green radiation. When a star is green, it is pretty much right in the middle of the visible spectrum. It is radiating strongly at all visible wavelengths, with most of the radiation right in the middle. When we look at the star, then, all these colors are mixed and the result is the color white. So you won't ever see a green-looking star through a telescope.

Fridger

[t00fri]

Chris,

another one, before checking in:

1) For systematic reasons, I suggest to rename my deepsky.dsc data file for the galaxies into galaxies.dsc. Any problems with that?

2) What about my two (trivial) celx scripts that only are "place holders"?
We should definitely assign 2 keys (r,R?) for rendering and labels as well as slots in the GUI's for the globulars. Then we should activate their rendering by default, just like the galaxies.

It would be good to straighten these simple issues out before I commit the stuff.

Fridger

[chris]

Chris,

another one, before checking in:

1) For systematic reasons, I suggest to rename my deepsky.dsc data file for the galaxies into galaxies.dsc. Any problems with that?

No problems--this is a good idea.

2) What about my two (trivial) celx scripts that only are "place holders"?
We should definitely assign 2 keys (r,R?) for rendering and labels as well as slots in the GUI's for the globulars. Then we should activate their rendering by default, just like the galaxies.

It would be good to straighten these simple issues out before I commit the stuff.

I think that the scripts can be omitted from SVN, since their function will quickly be replaced by keyboard bindings and GUI elements. As for the key bindings, I think I prefer your original assignments of U and E--not exactly mnemonic, but the fact that they parallel the galaxy key bindings is nice. A script that would be nice for the standard package is one that marks all globular clusters.

--Chris

[t00fri]

I think that the scripts can be omitted from SVN, since their function will quickly be replaced by keyboard bindings and GUI elements. As for the key bindings, I think I prefer your original assignments of U and E--not exactly mnemonic, but the fact that they parallel the galaxy key bindings is nice. A script that would be nice for the standard package is one that marks all globular clusters.

--Chris

Sure, I also think the scripts were only for testing purposes. Since (dwarf) galaxies and globulars are so closely related and we are short of keys, OK, perhaps assigning both kinds to U, E is the best compromise for now.

Fridger

[chris]

I think that the scripts can be omitted from SVN, since their function will quickly be replaced by keyboard bindings and GUI elements. As for the key bindings, I think I prefer your original assignments of U and E--not exactly mnemonic, but the fact that they parallel the galaxy key bindings is nice. A script that would be nice for the standard package is one that marks all globular clusters.

--Chris

Sure, I also think the scripts were only for testing purposes. Since (dwarf) galaxies and globulars are so closely related and we are short of keys, OK, perhaps assigning both kinds to U, E is the best compromise for now.

I thought you were using u and e for galaxies, and Shift+U and Shift+E for globulars?

--Chris

[t00fri]

I think that the scripts can be omitted from SVN, since their function will quickly be replaced by keyboard bindings and GUI elements. As for the key bindings, I think I prefer your original assignments of U and E--not exactly mnemonic, but the fact that they parallel the galaxy key bindings is nice. A script that would be nice for the standard package is one that marks all globular clusters.

--Chris

Sure, I also think the scripts were only for testing purposes. Since (dwarf) galaxies and globulars are so closely related and we are short of keys, OK, perhaps assigning both kinds to U, E is the best compromise for now.

I thought you were using u and e for galaxies, and Shift+U and Shift+E for globulars?

--Chris

No I was "simple minded" and just switched galaxies and globulars ON|OFF simulataneously to get started...

But we may certainly go for u,e and SHIFT+u, SHIFT+e. Should be simple enough to code it. Let's see.

Fridger

[ANDREA]

And, being so close to both 1.6 and to the globular clusters release, would it be possible to add to .cel and .celx commands the ones regarding such objects, I mean
renderflags set/clear "globularclusters" and
labels set/clear "globularclusterss" ?

Vincent?
Please!
Bye and thanks you.

Andrea

[ANDREA]

Excellent, ANDREA.
So did you have to download the video first? Shatters is a bit slow right now (downloads?) .
Fridger

First I tried to open through both Explorer and Firefox, but in both cases after a short time it stopped running, dunno why.
So I downloaded the video, but fortunately I have a high-speed ADSL connection, so it didn't take long to download.
Bye

Andrea

[t00fri]

And, being so close to both 1.6 and to the globular clusters release, would it be possible to add to .cel and .celx commands the ones regarding such objects, I mean
renderflags set/clear "globularclusters" and
labels set/clear "globularclusterss" ?

Vincent?
Please!
Bye and thanks you.

Andrea

Andrea,

the celx part of what you want is already part of my code. So once it's committed to SVN (tomorrow, presumably) you can just activate the rendering and labels of the globulars with these two

celestia:show("globulars")
celestia:showlabel("globulars")

in start.celx, for example.

This is also mentioned in my introductory post at the beginning of this thread.

Fridger

[t00fri]

Excellent, ANDREA.
So did you have to download the video first? Shatters is a bit slow right now (downloads?) .
Fridger

First I tried to open through both Explorer and Firefox, but in both cases after a short time it stopped running, dunno why.
So I downloaded the video, but fortunately I have a high-speed ADSL connection, so it didn't take long to download.
Bye

Andrea

Same with me. It's a highly compressed MPEG-4 format in fairly big size 720x480, so this takes time and/or plenty of cache I suppose.

Fridger

[ANDREA]

And, being so close to both 1.6 and to the globular clusters release, would it be possible to add to .cel and .celx commands the ones regarding such objects, I mean
renderflags set/clear "globularclusters" and
labels set/clear "globularclusterss" ?
Vincent? Please! Bye and thanks you. Andrea

Andrea,
the celx part of what you want is already part of my code. So once it's committed to SVN (tomorrow, presumably) you can just activate the rendering and labels of the globulars with these two
celestia:show("globulars")
celestia:showlabel("globulars")
in start.celx, for example.
This is also mentioned in my introductory post at the beginning of this thread.
Fridger

Fridger, I know this, I used your .celx scripts to obtain the images I showed in this thread.
Mine was just a joking provocation to Vincent to try to obtain the same commands for cel scripting, because as you well know I'm absolutely unable to understand and use celx (but I'm in good company here), even if I tried it a lot of times.
Sorry.
Bye

Andrea

[ElChristou]

Is 4449 final to build with globulars?