Re: [MaxImDL] Re: Question on stacking in Version 5...


Aug 4, 2008

 


----------------------------

#33038 Aug 4, 2008

I'm trying to familiarize myself with Version 5. I just upgraded from

4.51.

The Combine operation seems to be gone. It is replaced with the much

more sophisticated and very powerful Stack operation.

I thought I would reprocess some of my recently acquired images, say,

m22 from about a month ago.

Unfortunately, neither "Classify by Object" nor "Classify by Filter"

is working.

According to the manual these classifications work by looking at the

Object and Filter entries in the fits header.

Unfortunately, my fits headers have neither of these keywords.

Now, missing the Object keyword seems like a minor issue to me since I

only have one object per folder.

But, missing the Filter keyword is a big problem and I don't know why

it is missing.

The reason it is a problem is that now the fancy new Stack command

does not automatically give me

separate R, G, and B stacks. I just get one big stack combining all

of them. Okay, so I can check and

uncheck things and manually get R, G, and B stacks. But, they are not

aligned to each other. So, I have

to go back and align and stack the three R, G, and B frames to get an

L frame. Then I have to go back again

to color combine the synthesized (and perhaps log stretched and

sharpened) L frame with the R, G, and B frames.

All of these manual steps seemed easier to me before. I really like

the features of the new stack command, especially

the ranking and sorting of the images based on quality. But, a simple

combine of R, G, and B into an L and then the

color combine of L, R, G, and B seems much more tedious with this

fancy stack command. It would be great to have

both a stack and the good old combine. Am I missing something?



Anyway, much of my complaint above would disappear if I could figure

out why my R, G, and B frames do not

have a "Filter" entry in their fits headers. The files were created

using Maxim DL 4.51's "Sequence" tab. I don't have

an automatic filter wheel. I change them by hand. But, I do tell the

sequence tool what filter I'm using and the sequence

capture does correctly put the appropriate suffixes on the file names

(e.g. m22-001R.fit). So, why is Filter missing from the

header? Is there some simple automated way to put it in after the

fact (I have over 120 frames, so I'm not going to

do it one by one by hand).



Any help would be greatly appreciated.



--Bob

Robert J. Vanderbei

Princeton University

www.princeton.edu/~rvdb

"If I only had a brane" -- S. Hawking









[Non-text portions of this message have been removed]



----------------------------

#33039 Aug 4, 2008

Okay, I can partially answer my own question.

I fired up the old version (4.51) and had a look at the "Setup

Sequence" window.

It has a column called "Filter". But, for me the option is grayed out

and has written

"No Filter" in every row of that column. So, I consulted the user's

manual and it

says: "If a filter wheel is present and the exposure type is Light or

Flat, the Filter drop-list allows you to select the filter to use."

Well, this explains it. I don't have a filter wheel. I change

filters by hand. It never seemed worth it to me to invest several

hundred dollars in a gizmo that can automatically change my filters

for me (I use a Starlight Express camera, so the filter

wheel is not something that is built in, which until now I thought was

a good thing), especially given that I like to refocus

when I change filters anyway. It would be really great if the Stack

command could be tweaked somehow to be more

useful to those of us who don't use an electronic filter wheel.



Robert J. Vanderbei

Princeton University

www.princeton.edu/~rvdb

"Just what the truth is, I can't say anymore." -- Moody Blues



On Aug 4, 2008, at 10:04 PM, Robert Vanderbei wrote:



> I'm trying to familiarize myself with Version 5. I just upgraded from

> 4.51.

> The Combine operation seems to be gone. It is replaced with the much

> more sophisticated and very powerful Stack operation.

> I thought I would reprocess some of my recently acquired images, say,

> m22 from about a month ago.

> Unfortunately, neither "Classify by Object" nor "Classify by Filter"

> is working.

> According to the manual these classifications work by looking at the

> Object and Filter entries in the fits header.

> Unfortunately, my fits headers have neither of these keywords.

> Now, missing the Object keyword seems like a minor issue to me since I

> only have one object per folder.

> But, missing the Filter keyword is a big problem and I don't know why

> it is missing.

> The reason it is a problem is that now the fancy new Stack command

> does not automatically give me

> separate R, G, and B stacks. I just get one big stack combining all

> of them. Okay, so I can check and

> uncheck things and manually get R, G, and B stacks. But, they are not

> aligned to each other. So, I have

> to go back and align and stack the three R, G, and B frames to get an

> L frame. Then I have to go back again

> to color combine the synthesized (and perhaps log stretched and

> sharpened) L frame with the R, G, and B frames.

> All of these manual steps seemed easier to me before. I really like

> the features of the new stack command, especially

> the ranking and sorting of the images based on quality. But, a simple

> combine of R, G, and B into an L and then the

> color combine of L, R, G, and B seems much more tedious with this

> fancy stack command. It would be great to have

> both a stack and the good old combine. Am I missing something?

>

> Anyway, much of my complaint above would disappear if I could figure

> out why my R, G, and B frames do not

> have a "Filter" entry in their fits headers. The files were created

> using Maxim DL 4.51's "Sequence" tab. I don't have

> an automatic filter wheel. I change them by hand. But, I do tell the

> sequence tool what filter I'm using and the sequence

> capture does correctly put the appropriate suffixes on the file names

> (e.g. m22-001R.fit). So, why is Filter missing from the

> header? Is there some simple automated way to put it in after the

> fact (I have over 120 frames, so I'm not going to

> do it one by one by hand).

>

> Any help would be greatly appreciated.

>

> --Bob

> Robert J. Vanderbei

> Princeton University

> www.princeton.edu/~rvdb

> "If I only had a brane" -- S. Hawking

>

>

>

>

> [Non-text portions of this message have been removed]

>

>

---------------

>

> Yahoo! Groups Links

>

>

>







[Non-text portions of this message have been removed]







----------------------------

#33040 Aug 4, 2008

Hi Robert,



It's not much help with the images you've already taken, but for

future reference there's a "Manual" filter wheel option in Maxim.



Selecting this lets you tell Maxim which filter you're using (it will

prompt you when to change the wheel manually when doing a sequence)

and it will put the appropriate keyword in the fits header.



Cheers,



John



--- In MaxImDL@yahoogroups.com, Robert Vanderbei rvdb@...> wrote:

>

> Okay, I can partially answer my own question.

> I fired up the old version (4.51) and had a look at the "Setup

> Sequence" window.

> It has a column called "Filter". But, for me the option is grayed out

> and has written

> "No Filter" in every row of that column. So, I consulted the user's

> manual and it

> says: "If a filter wheel is present and the exposure type is Light or

> Flat, the Filter drop-list allows you to select the filter to use."

> Well, this explains it. I don't have a filter wheel. I change

> filters by hand. It never seemed worth it to me to invest several

> hundred dollars in a gizmo that can automatically change my filters

> for me (I use a Starlight Express camera, so the filter

> wheel is not something that is built in, which until now I thought was

> a good thing), especially given that I like to refocus

> when I change filters anyway. It would be really great if the Stack

> command could be tweaked somehow to be more

> useful to those of us who don't use an electronic filter wheel.

>

> Robert J. Vanderbei

> Princeton University

> www.princeton.edu/~rvdb

> "Just what the truth is, I can't say anymore." -- Moody Blues

>

>

> On Aug 4, 2008, at 10:04 PM, Robert Vanderbei wrote:

>

> > I'm trying to familiarize myself with Version 5. I just upgraded from

> > 4.51.

> > The Combine operation seems to be gone. It is replaced with the much

> > more sophisticated and very powerful Stack operation.

> > I thought I would reprocess some of my recently acquired images, say,

> > m22 from about a month ago.

> > Unfortunately, neither "Classify by Object" nor "Classify by Filter"

> > is working.

> > According to the manual these classifications work by looking at the

> > Object and Filter entries in the fits header.

> > Unfortunately, my fits headers have neither of these keywords.

> > Now, missing the Object keyword seems like a minor issue to me since I

> > only have one object per folder.

> > But, missing the Filter keyword is a big problem and I don't know why

> > it is missing.

> > The reason it is a problem is that now the fancy new Stack command

> > does not automatically give me

> > separate R, G, and B stacks. I just get one big stack combining all

> > of them. Okay, so I can check and

> > uncheck things and manually get R, G, and B stacks. But, they are not

> > aligned to each other. So, I have

> > to go back and align and stack the three R, G, and B frames to get an

> > L frame. Then I have to go back again

> > to color combine the synthesized (and perhaps log stretched and

> > sharpened) L frame with the R, G, and B frames.

> > All of these manual steps seemed easier to me before. I really like

> > the features of the new stack command, especially

> > the ranking and sorting of the images based on quality. But, a simple

> > combine of R, G, and B into an L and then the

> > color combine of L, R, G, and B seems much more tedious with this

> > fancy stack command. It would be great to have

> > both a stack and the good old combine. Am I missing something?

> >

> > Anyway, much of my complaint above would disappear if I could figure

> > out why my R, G, and B frames do not

> > have a "Filter" entry in their fits headers. The files were created

> > using Maxim DL 4.51's "Sequence" tab. I don't have

> > an automatic filter wheel. I change them by hand. But, I do tell the

> > sequence tool what filter I'm using and the sequence

> > capture does correctly put the appropriate suffixes on the file names

> > (e.g. m22-001R.fit). So, why is Filter missing from the

> > header? Is there some simple automated way to put it in after the

> > fact (I have over 120 frames, so I'm not going to

> > do it one by one by hand).

> >

> > Any help would be greatly appreciated.

> >

> > --Bob

> > Robert J. Vanderbei

> > Princeton University

> > www.princeton.edu/~rvdb

> > "If I only had a brane" -- S. Hawking

> >

> >

> >

> >

> > [Non-text portions of this message have been removed]

> >

> >

---------------

> >

> > Yahoo! Groups Links

> >

> >

> >

>

>

>

> [Non-text portions of this message have been removed]

>



----------------------------

#33043 Aug 4, 2008

I see. Thanks for that. I didn't know about the "manual" setting

before.

Of course, I'm mainly interested in future images as I've already

calibrated

and processed old images. Just thought it would be fun to see if the

automatic ranking/sorting/rejecting of images might quickly and

easily give improvements on some old images. I guess I'll just pass on

that and wait until I have new data to work with.



BTW, it seems that the ranking/sorting/rejecting algorithm works very

well. Also, I like the "remove background color" function. Version

4.51

didn't have this. There are significant gradients in my m22 data

(it's quite

low from my latitude) which I had corrected before using Photoshop.

That's what got me playing around with that data. Version 5 with no

effort does a better job that I was able to achieve using Photoshop.

All in all, the new version 5 looks really great.

Robert J. Vanderbei

Princeton University

www.princeton.edu/~rvdb

"I sleep in the daytime, work in the nighttime, I might not ever get

home" -- Talking Heads



On Aug 4, 2008, at 11:46 PM, John Winfield wrote:



> Hi Robert,

>

> It's not much help with the images you've already taken, but for

> future reference there's a "Manual" filter wheel option in Maxim.

>

> Selecting this lets you tell Maxim which filter you're using (it will

> prompt you when to change the wheel manually when doing a sequence)

> and it will put the appropriate keyword in the fits header.

>

> Cheers,

>

> John

>

>

> --- In MaxImDL@yahoogroups.com, Robert Vanderbei rvdb@...> wrote:

>>

>> Okay, I can partially answer my own question.

>> I fired up the old version (4.51) and had a look at the "Setup

>> Sequence" window.

>> It has a column called "Filter". But, for me the option is grayed

>> out

>> and has written

>> "No Filter" in every row of that column. So, I consulted the user's

>> manual and it

>> says: "If a filter wheel is present and the exposure type is Light

>> or

>> Flat, the Filter drop-list allows you to select the filter to use."

>> Well, this explains it. I don't have a filter wheel. I change

>> filters by hand. It never seemed worth it to me to invest several

>> hundred dollars in a gizmo that can automatically change my filters

>> for me (I use a Starlight Express camera, so the filter

>> wheel is not something that is built in, which until now I thought

>> was

>> a good thing), especially given that I like to refocus

>> when I change filters anyway. It would be really great if the Stack

>> command could be tweaked somehow to be more

>> useful to those of us who don't use an electronic filter wheel.

>>

>> Robert J. Vanderbei

>> Princeton University

>> www.princeton.edu/~rvdb

>> "Just what the truth is, I can't say anymore." -- Moody Blues

>>

>>

>> On Aug 4, 2008, at 10:04 PM, Robert Vanderbei wrote:

>>

>>> I'm trying to familiarize myself with Version 5. I just upgraded

>>> from

>>> 4.51.

>>> The Combine operation seems to be gone. It is replaced with the

>>> much

>>> more sophisticated and very powerful Stack operation.

>>> I thought I would reprocess some of my recently acquired images,

>>> say,

>>> m22 from about a month ago.

>>> Unfortunately, neither "Classify by Object" nor "Classify by Filter"

>>> is working.

>>> According to the manual these classifications work by looking at the

>>> Object and Filter entries in the fits header.

>>> Unfortunately, my fits headers have neither of these keywords.

>>> Now, missing the Object keyword seems like a minor issue to me

>>> since I

>>> only have one object per folder.

>>> But, missing the Filter keyword is a big problem and I don't know

>>> why

>>> it is missing.

>>> The reason it is a problem is that now the fancy new Stack command

>>> does not automatically give me

>>> separate R, G, and B stacks. I just get one big stack combining all

>>> of them. Okay, so I can check and

>>> uncheck things and manually get R, G, and B stacks. But, they are

>>> not

>>> aligned to each other. So, I have

>>> to go back and align and stack the three R, G, and B frames to get

>>> an

>>> L frame. Then I have to go back again

>>> to color combine the synthesized (and perhaps log stretched and

>>> sharpened) L frame with the R, G, and B frames.

>>> All of these manual steps seemed easier to me before. I really

>>> like

>>> the features of the new stack command, especially

>>> the ranking and sorting of the images based on quality. But, a

>>> simple

>>> combine of R, G, and B into an L and then the

>>> color combine of L, R, G, and B seems much more tedious with this

>>> fancy stack command. It would be great to have

>>> both a stack and the good old combine. Am I missing something?

>>>

>>> Anyway, much of my complaint above would disappear if I could figure

>>> out why my R, G, and B frames do not

>>> have a "Filter" entry in their fits headers. The files were created

>>> using Maxim DL 4.51's "Sequence" tab. I don't have

>>> an automatic filter wheel. I change them by hand. But, I do tell

>>> the

>>> sequence tool what filter I'm using and the sequence

>>> capture does correctly put the appropriate suffixes on the file

>>> names

>>> (e.g. m22-001R.fit). So, why is Filter missing from the

>>> header? Is there some simple automated way to put it in after the

>>> fact (I have over 120 frames, so I'm not going to

>>> do it one by one by hand).

>>>

>>> Any help would be greatly appreciated.

>>>

>>> --Bob

>>> Robert J. Vanderbei

>>> Princeton University

>>> www.princeton.edu/~rvdb

>>> "If I only had a brane" -- S. Hawking

>>>

>>>

>>>

>>>

>>> [Non-text portions of this message have been removed]

>>>

>>>

---------------

>>>

>>> Yahoo! Groups Links

>>>

>>>

>>>

>>

>>

>>

>> [Non-text portions of this message have been removed]

>>

>

>

>

---------------

>

> Yahoo! Groups Links

>

>

>







[Non-text portions of this message have been removed]







----------------------------

#33071 Aug 5, 2008

--- In MaxImDL@yahoogroups.com, Robert Vanderbei rvdb@...> wrote: > the fancy new Stack command does not automatically give me

> separate R, G, and B stacks. I just get one big stack combining

> all of them. Okay, so I can check and uncheck things and manually

> get R, G, and B stacks. But, they are not aligned to each other.

> [...] Am I missing something?



Bob,



As you've indicated, changing your procedures so the images are

tagged with the appropriate FILTER (OBJECT is a good idea too!) is

definitely the way to go... but sometimes you have to process data

that isn't tagged the way Stacker wants it. The good news is it's

still possible, albeit with rather more user interaction. Here's one

way:



1) Click on Add Files and select all the luminance images. Shift- or

control-click in the file selector box to do multiple selections.

Also, your naming conventions allow you to list only the images for

the channel of interest in by typing wildcards like "M22-???L.*" into

the Filename box. This makes the entire batch selectable by the

single keystroke Ctrl-A. When done, click Open.



2) Stacker will create a monochrome "Group1 (no OBJECT)" or similar.

Right-click on it and pick "Set Group > L" to label it as luminance.

(You might also want to rename it just to keep things straight.)



3) Now control-click the group to unselect it, causing the

description box to go blank. This is important so the next color

channel's images are put in a new group rather than added to the

existing L group.



4) Click on Add Files and select the red images, as in step 1.



5) Mark the newly created group as Red using "Set Group > R" and

rename it if you care to.



6,7,8) Unselect the R group, Add Files for green, and "Set Group > G".



9,10,11) Ditto for blue.



Now Stacker knows about all the images, but they're in four

independent "top level" groups. To combine them together using a

single reference, you need to put them into a 'supergroup':



12) Right-click anywhere on the tree and select New Group. Stacker

will create a new, empty group at the bottom.



13) Right-click on the new group and select "Set Group > LRGB".



14) Now drag each color channel group onto the newly created LRGB

group: Stacker puts them "inside". You can select all four first and

drag the entire lot at once if you prefer. (While it isn't necessary

in this scenario, you should be aware that you can also drag images

between groups. This is handy if you accidentally add files to the

wrong group.)



You now have the group structure that Stacker would have created

automatically had there been FILTER tags. At this point you will

probably want to pick a reference image from the luminance subgroup,

though Stacker will pick one itself if you don't; it just might not

be the optimal choice. You can Set Object Name on the top-level

group, too, if you want the result to have a meaningful OBJECT tag.

Set the quality filters if desired, alignment method, color model,

combine mode, and click Go. Voila: (semi-)instant gratification!



Hope this helps demystify Stacker operations further. It may seem

complicated and elaborate, but it's really not difficult and I like

to think it's pretty logical.



Hilderic Browne

Diffraction Limited



----------------------------

#33078 Aug 6, 2008

Hilderic,



A million thanks. Your instructions are very helpful. Only one thing

didn't work exactly as

you indicated, but it was just a parenthetical remark so it hardly

counts. :-)

You suggested changing the name of each new group to help keep things

organized.

I did that. But, when I dragged the R, G, and B groups into the RGB

supergroup,

their names were changed automatically back to things like groupxx.

So, I had to change them

again after they were placed within the RGB supergroup.



I do lots of HaO3O3 imaging. That is, I grab Ha and O3 data and then

combine using

R=Ha, G=O3, and B=03. I was able to do this with the new stack

routine by assigning the Ha images to

the R group, the O3 images to be the B group, and making a G group

consisting of one randomly

chosen image. Then in the Color tab I set R->R, B->G, and B->B each

with weight 1.0 and I put zero

weight in the other places. This worked fine and gives a perfect

result. One little nit. I'm not

using the G input frame for anything---all of the G output weights are

0. But, the tool required me

to create a G group (that seems okay to me) and it also required me to

put a file in that group. I would

prefer it not to have to do this last step. It's completely

unnecessary and the image I put there was

completely arbitrary.



One last thing. Since I'm operating in slightly manual mode here, I

was thinking it would be nice

to set up tree structures for my various old image sets and not delete

them. In other words I would over

time have more and more supergroups---one for each object I've

stacked. That way I could come back

weeks or months later and try different quality settings or color

weights or what have you. Seems like

a good idea and seems like it should be trivial. But, there is a

small catch. Groups have to have distinct

names even if they are in different supergroups. In other words, the

following is not allowed...



m22

Red

Green

Blue

ngc6992

Red

Green

Blue



Of course, I can work around this by changing the above to



m22

m22-Red

m22-Green

m22-Blue

ngc6992

ngc6992-Red

ngc6992-Green

ngc6992-Blue



but that's a bit ugly.

Robert J. Vanderbei

Princeton University

www.princeton.edu/~rvdb

"This ain't no party, this ain't no disco, this ain't no foolin

around" -- Talking Heads











[Non-text portions of this message have been removed]







----------------------------

#33080 Aug 6, 2008

Robert J. Vanderbei

Princeton University

www.princeton.edu/~rvdb

"Just what the truth is, I can't say anymore." -- Moody Blues



On Aug 6, 2008, at 1:59 PM, hildericbrowne wrote:



> --- In MaxImDL@yahoogroups.com, Robert Vanderbei rvdb@...> wrote:

>>

>

>> I do lots of HaO3O3 imaging. [...] I grab Ha and O3 data and then

>> combine using R=Ha, G=O3, and B=03. [...] I'm not using the G

>> input frame for anything---all of the G output weights are 0.

>

> Set the supergroup type to "Two Color", then you only need to

> add 'Red' and 'Green' groups. By default, the Blue channel is set to

> R/2+G/2, but you can adjust the weights as desired, including putting

> 100% of the 'Green' input to the Blue instead of the Green output

> channel.



Grok.



>

>> it would be nice to set up tree structures for my various old

>> image sets and not delete them. In other words I would over

>> time have more and more supergroups---one for each object I've

>> stacked. That way I could come back weeks or months later...

>

> Doesn't work that way I'm afraid. The group definitions persist

> within a session but not if you exit and restart. Perhaps down the

> road, though if it "seems like it should be trivial" to you, maybe

> you'd like to come and work here. :-)



:-)



>

>

>> Groups have to have distinct names even if they are in different

>> supergroups.

>

> True, but Stacker's preferred naming convention for internal color

> groups is "{parent group name} {color}", and such a group displays as

> if it were simply called "{color}". So you'll achieve the effect of

>

>> m22

>> .Red

>> .Green

>> .Blue

>

> if you name them

>

> m22

> .m22 Red

> .m22 Green

> .m22 Blue

>

> i.e., use a space as a separator rather than a hyphen.



I see. I'll do it that way. Thanks a ton for your help.



--Bob



[Non-text portions of this message have been removed]



Contact Us
This Site's Privacy Policy
Google's privacy policies

S
e
n
i
o
r
T
u
b
e
.
o
r
g