Original subtitles

This video is sponsored by Squarespace. Welcome back! My name is Nico Carver and in last week's

episode we covered how to capture M101, the Pinwheel Galaxy, with used, inexpensive gear and

in this week's episode we're going to take that data captured last week and process it with free,

open-source software, specifically these four programs: Siril, GraXpert, Starnet, and the GNU

Image Manipulation Program. I promised I'd show how to install these programs but I've decided to

put that as an addendum at the end of the video so if you do need installation help you can just skip

to the addendum appropriate to your operating system whether it's Mac or Windows. Here's an

overview of what I'll be covering and in what order. I won't bore you by reading this out but

for people that do want a summary overview you can pause the video right here to read through this.

Okay the first step is I've looked at my images in playback mode on the camera and figured out

the ranges of my light frames, my dark frames, my bias, and my flats and I've noted those ranges

down now I'm ready to take out the SD card from my camera and insert it into my laptop here or if

you don't have an SD card slot on your computer or laptop then you'd need an SD card reader for

this step or connect your DSLR with a USB cable in any case once you've put the SD card into your

computer you should see some kind of popup or on Mac it will just appear on your desktop here and

if I double click that and then go into this DCIM folder uh ignore this stuff uh I've used this for

different camera systems but anyways here's all of the pictures I took on the Canon T7 and since

I've already noted down the ranges I can now start selecting these and transferring them to

my computer um however let's talk about computer hard drive space because if I look at my computer

hard drive here I'm just going to click get info you can see that I have it's a 1 TB drive and I

have about 350 GB available now that might be enough but I'm not sure because I took about

500 pictures and the way that serial works the program that we're going to be using it uses way

more space on your hard drive during processing um and then you can delete those pictures that

it creates later but but you need a lot of space on an SSD for serial to work correctly so just to

be safe instead of putting this these pictures just on my internal SSD I'm going to connect an

external SSD which has more free space this one has over 3 terabytes available so that should be

plenty so what I'm going to do is I'm going to open up this external SSD and I'm going to make

a new folder on Mac or Windows you can just do that by right-clicking and choosing new folder

and I'm going to call this M101 that's the uh designator Messier 101 for the pinwheel

galaxy and then inside that folder I'm going to make four other folders

four subfolders one called lights one called darks one called flats and one called biases

and this is the hierarchy that serial expects if you use the scripts for pre-processing which

is what I'm going to be doing and showing to in this tutorial okay then it's just a

matter of going here on the SD card finding the ranges and transferring them to the SSD

and you can either do that by selecting and pressing control-C or command-C on Mac and

then pasting control-V or command-V on Mac or you can drag and drop from one window to the

other okay now with everything all the lights biases darks flats all copied over

um I can go ahead and eject my SD card and so I'm going to start with Siril

you can see I am on the latest version as of recording which is 1.4 beta 2 and we're

going to do two things right at the start here we're going to click on this blue home button

right up here which will change our current working directory you can see currently my

current working directory is my desktop but I want to change it to my external SSD so

I'm going to click on other locations go into the computer go into the volumes and there it

is x10 Pro and then M101 okay so when you get to your working directory you should see these four

folders that we created and then go ahead and click open and then up here at the top of the

interface you'll see that it lists your working directory so for mine it's volumes X10 Pro M101

okay so that's always the first step before you use the script you have to tell it where

your data in those four folders lives then we can go into the scripts and this new

uh script dropdown is a little different than in past versions there's now Python scripts which are

as it says Python scripts things written in Python um within serial and there's only one right now in

there by default the catalog installer but there's many other people currently writing Python scripts

that will work with Siril um so this is a cool new way for any kind of developer just to extend the

functionality of serial through Python scripts and then there is the standard serial scripts all of

these ones except for RGB composition have to do with pre-processing your data so if you don't use

the scripts you can instead go through these tabs right over here calibration registration stacking

but what the scripts do is they take care of all of that for you um so uh but they of course then

make some assumptions about how you want to do the pre-processing okay anyways um there

are some new scripts like this Bayer Drizzle I'm not going to use that here since I didn't

uh guide and dither and things like that that would be required for drizzling and so I'm not

going to go into that right now but I'm just going to use the standard one uh OSC pre-processing

and oneshot color OSC is appropriate for anytime you're using a DSLR or mirrorless anytime you're

using a camera that's not a mono camera you'll use one of these OSC scripts these ones are for

if you use narrowand filters so since we didn't use any filters we used just a normal stock DSLR

we're going to do this one OSC pre-processing so to start it all you have to do is just click on

it and you might get a popup here saying like something about scripts are experimental or

or don't uh trust all scripts or something like that if you get that you can just click okay I've

already dismissed it before so that's why I didn't get it um but as you can see now it's just going

really quickly through all of the pre-processing it creates uh calibration masters then it uh

calibrates your lights using those masters then it debeayers your light frames meaning

um actually interpolates them to make them full color images then it registers your light frames

so you that's why you need those round in-focus stars because it's using star patterns to do the

registration so that they're all aligned then once they're all aligned it can actually stack them so

that's the mathematical averaging and it can throw out outliers like satellite trails and things like

that um but it also of course greatly improves the signal to noise ratio of your images which lets

you stretch them um much brighter and actually see these dim objects that we've captured so I'm

just going to let this go and when it's done we'll catch back up okay and I just finished stacking it

took about 17 minutes to calibrate and register and stack about 500 lights from a DSLR and this

is just on a uh fairly new Mac laptop but nothing special in terms of computer power but I would say

as I've said many times before you really want to be doing this as on an SSD if you're doing it on

a spinning disc drive like a normal hard drive um it takes way way longer like hours compared

to minutes uh the the really thing that speeds it up is working on an SSD once it finishes it'll say

script execution finished successfully but nothing opens up here automatically we do have to click

on open and then here in our working directory you can see there's two new folders masters and

process and then a result um and then the result will be underscore and then the total time of what

it stacked so this stacked 9,900 seconds so that would be 165 minutes so almost 3 hours of data

so I'm going to go ahead and open up that result and this is what it looks like um but actually

while I was talking about that those folders I wanted to mention one thing before we move

on these masters if you ever you know want to calibrate manually without the script you might

want to hang on to like your master bias and uh this is also a good time just to look at your

master calibration frames and see if there if you see anything wrong with them um like banding in

the flat or just weird issues um it can be good to save these but the process folder I almost always

delete it's it's going to be really really huge because it's full of all the intermediary files

that serial creates and just to give you an idea of how big it is my original lights folder was 16

gigabytes of data this process folder ballooned to 320 GB of data so um over 10 times what we

started with even if you include the calibration frames it's over 10 times bigger um and we don't

need it so what I always try to remember to do is at this point delete it so I'm just going to move

it to my trash and then empty my trash to free up that hard drive space so back to the result

um this is being shown in what's called a linear space if you look down here at the bottom of the

interface um it says linear right there and that's uh good there's certain things that we want to

do to the picture while it's still linear but you can't really see the galaxy at all in this linear

mode so we're going to change the view mode to auto stretch just by clicking on the word

linear down here at the bottom and change it to auto stretch and now you can see the galaxy you

also see a lot of noise which don't be alarmed by we're going to manage the noise through processing

but it's actually good to be able to see this noise because uh you can see that I didn't very

carefully keep the galaxy centered the whole time I was tracking it so there's some parts of

the image that are noisier than others you can see there's a big band down here at the bottom and on

the left i mean on the right and at the top where these sections didn't get as much stacking as the

center part so I would definitely want to crop those off um and you can just see any issues that

are maybe here in the data with it stretched this brightly but don't worry too much about its like

color or appearance at this point because we're going to do a lot of different things to it um but

it gives you a a good sort of first sense of what what you've captured so you can see there there's

a lot of um that doesn't the galaxy itself doesn't look very noisy even if we're seeing a lot of

noise out here in the sky the galaxy itself looks like we have a lot of signal to work with which

is great um even in these outer outer arms we can see there's they're starting to come in and in the

the brighter central part it's definitely bright enough now in this auto stretch we're not seeing

a whole lot of color variation in the galaxy it just sort of has this bluish white tone to it but

um we can tease out the color later so don't worry about that either okay anyways uh so that's

the auto stretch. We're going to move actually right into GraXpert there is now in serial 1.4

this ability to use GraXpert right within serial now um when you open that though I want to point

out that it gives you this deprecation notice and says this interface is not what it's going

to be in future versions instead it's going to be a Python script so I'm actually not going to show

that this interface in my video i know it exists but um I'm going to show GraXpert as a standalone

program because I don't want this video to get confusing in the future okay so we're going to

go to GraXpert Standalone but before we do there's one thing I want to measure in this image because

we need this one piece of information in order to accurately do something in GraXpert called

deconvolution which you can think of as basically just undoing blur in your image like there's blur

from the atmosphere and different things and deconvolution takes out some of that blur

so it's it's uh very mathematical but we need to know the average size of the stars in the image

to do deconvolution as correctly as possible so uh there's two things we have to do to get that

number the first is we have to plate solve our image meaning figure out exactly the scale of

uh the image and where it is in the sky so we're going to go to tools astrometry image plate solver

and depending on your gear this might be filled in for for me it's not very filled in so uh I'm going

to go ahead and fill it in so I'm going to type in my object name up here in the little search

box M101 click find it finds the coordinates I'm going to type in my focal length i was using a 300

mm focal length lens and it actually did pull the pixel size correctly pretty much uh for my camera

it's actually more accurately 3.72 and you can just Google that uh pixel size your camera name

okay and then I'll go ahead and click okay and it plate solves the image quite quickly you can

see down here in the console uh it calculated the focal length to be 298 mm focal length and

it gave me um some center coordinates here okay then next we can go to the image processing menu

up here at the top go down to star processing and we don't actually want to click on any of these

four options we want to click on this little gear icon and this will bring up the dynamic PSF window

um and dynamic PSF PSF is point spread function so it's going to calculate uh data about all of

these stars so to do that we're going to go ahead and uh click this detect stars button it finds all

the stars it gives you like their position their exact right ascension and declination meaning like

their uh celestial coordinates and then the one we're interested in is uh this one the full width

half maximum and so we can actually get a nice average of all the PSFs um by clicking this little

uh epsilon character down here okay and that tells us the average star data okay so full width half

maximum x 12.4 arcsec full width half maximum y 9.4 arcsec so let's just say we'll pick a value

in the middle of that uh so let's say 11 so we'll say that our average uh full width half maximum

the average size of our stars in this image is 11 arcseconds all right so you can if you don't want

to see all these little detections of the stars you can click this delete button right here that

will just get rid of this display you can close out of that but actually we can now close out of

serial uh completely all you have to do is just write down that number uh the average full width

half maximum for us it was 11 arcsec so I'm going to go ahead and quit out of serial and open up

graXpert and it warns me that this is a beta release of GraXpert which we know

um it will eventually become an official release but so far uh they haven't gotten to that yet so

this is a release candidate you can see that up here RC2 um but the reason I want to use it

is because it includes deconvolution which we're going to use on this image so go ahead and click

load image and then uh open up your result right here okay and it opens up in graXpert um normally

I would go down here to this display thing and make this brighter but um I think on the Mac there

might be a bug with this because when I tried earlier it uh made the program unresponsive so

I'm just going to leave it at 10% background but I am going to crop so we'll just follow along on

the left hand side here I'm going to open up the next uh thing here which is crop and I'm going to

crop pretty far in on the galaxy of course we saw earlier how there were major parts along the top

bottom and right that were noisier than the rest of the image just because of my imprecise tracking

so you may be able to keep more of your image than me but I don't want that those noisy parts

to interfere with um the other parts of this process so I'm going to get rid of all that

stuff and I'm going to do a fairly widescreen kind of image because that will look good for YouTube

of course so I think that looks good so I'm going to go ahead and click apply crop and of course

the nice thing with cropping is it also makes the galaxy appear bigger on your screen we could crop

in even further but I think there's a limit where it's like if you crop in too far on something

you'll start seeing that you don't really have the detail to support it so I think uh this looks

nice I'm going to click on background extraction i wouldn't say this image really needs background

extraction because I shot it fairly high in the sky and there doesn't seem to be any kind of light

pollution gradient here so we could just skip it but I'm going to go ahead and use just to show

you view I'm going to use the AI model and I'm going to turn up smoothing quite a bit because

I don't want it to overshoot on calculating the background and take away actually signal in the

picture so I'm going to turn the smoothing up to 0.9 if you had a complicated gradient in

your picture you might want to turn the smoothing down and it actually did a little bit of something

there that I think looks pretty good so here is the original there is the gradient corrected i

don't know if you can see that on YouTube there's the background i mean not much going on there but

maybe just a little bit of uh brighter on this side okay let's go on to deconvolution

if you don't yet have these AI models set up I just picked the latest on each one here under

advanced and then the first time you run these it will ask do you want to download the AI model and

you just say yes and that takes a little while the other thing I will note is I believe on

Mac and Windows 10 I would uncheck this enable hardware acceleration because I think there's

issues with it both on Mac and Windows 10 so what this normally would do AI hardware acceleration is

use your GPU for the AI tasks but I think there's a bug with it right now so I would turn it off

so I'm going to go ahead and close that okay and then you have two sliders deconvolution strength

and image full width half maximum in pixels okay and so this gets a little complicated but

um we need to figure out our image scale so we can do that with any number of uh calculators online

I'm just using the one here at Astronomy Tools under their CCD calculators and we know our pixel

size with the Canon T7 is 3.72 and our telescope focal length is 300 so that gives us a resolution

of 2.56 arcseconds per pixel so remember our full width half maximum the average width of the stars

in arcseconds that we calculated using serial was 11 arcsec so now we want to convert that

to a pixel value and we have our resolution here of 2.56 arcsec per pixel so then we can just do

the simple math of 11 / 2.56 to get our average uh full width half maximum in pixels which is

4.3 so now we just plug that into GraXpert right here where it says image full width F maximum in

pixels 4.3 okay and then the other slider here is deconvolution strength i think the default

is 0.5 and so if you want a stronger uh result you make it higher if you want a less strong

result you make this lower I'm going to do just a little bit above the uh the default so I'll do

0.6 and we'll give that a try okay and zoomed out like this I don't know if I

see much difference but usually the the difference is in the details you really have to look at the

uh galaxy up close and I think I can see that it's sharper but then the really good

way to tell is you go up here to where it says deconvolved object only and switching back and

forth between gradient corrected and deconvolved and you should see more details here in the core

if you see weird artifacts uh like like little wormy stuff that you don't expect then you want

to turn down the deconvolution strength or maybe turn up the image full width half maximum so you

can play around with it but every time you run it it does take a while so uh just keep that in

mind trying to get these values somewhat correct right away is probably preferable okay next up we

have denoising and so um you can see there is a fair amount of uh grain and color noise um in all

of the dim parts here of this image especially noticeable in the sky background denoise should

help with that and so you have different option here you can go anywhere from uh 0 to one i think

the default is 0.5 I'm maybe just going to go up a little bit from the default and go up to like

0.6 okay that took quite a long time but it did finish and

um now I can see what the non-denoised this one versus the denoised looks like and it did

make a fairly big difference but I don't know how well it's going to come across on YouTube

because you're seeing a video that already has been compressed but let me zoom in even

further and this is a good idea to zoom in like this anyways just to see okay is the denoising

taking away any detail in the galaxy so let's go back to just deconvolved and then denoised

and I think it did a really nice job um very impressive you can see the before there's like

a lot of like color noise and color model and grain in the background and then just take take

a look at the galaxy too and then here's after the denoise and I think it's smoothed out a lot

of that grain without taking away really much detail that we've uh achieved in the galaxy

and we can also we're also going to sharpen up the galaxy some more later on so I think that

is a good tradeoff and I'm going to keep it like that and then we'll just go right on to saving and

for saving we want to save as a 32bit FITS file because we're going to go back into serial and

that is the correct option so click save selected and then it will ask where do you want to save it

will probably default back to where you opened up the file which is this working directory so

uh I'm going to go ahead and save it there and open back up Siril and open our one that

we were just working on from GraXpert so it'll be called GraXpert Denoised FITS okay and I'm

going to turn back on the auto stretch and don't be alarmed when you turn on the auto stretch in

serial that all of a sudden it looks much noisier again because serial's default auto stretch may be

different than the view you the auto stretch you had up in GraXpert so uh don't worry about that

not an issue at all next thing that I sometimes do when I have I'm working with Canon cameras

is if I see any horizontal banding artifacts from Canon cameras I'll uh see if the banding reduction

filter in serial helps with that so I'm not really seeing enough banding in this to really make this

uh worth it but I just want to show you that um uh you can play around with this and basically

see like apply did that make it better or worse if it made it worse just undo and then try again with

different values here for the highlight protection and the amount and so you can just keep going

back and forth until the banding reduction maybe helps your image rather than hurts it if you see

any horizontal bands uh in this very stretched view that's common with Canon and some other

uh sensors again I with modern Canon cameras like the T7 it's really not too noticeable and I so I

really probably didn't need to do that but I just wanted to show it just in case okay the next thing

I want to do is run the image plate solver again so I know we've already gone over this but when

we brought our image into GraXpert and saved there it's it didn't retain that information so I'm just

going to run this again by typing in M101 click find type in 300 here 3.72 for the pixel size

and click okay okay and again it came up with the same solution which is good and we needed

to do that because the next thing we're going to do is either photometric or spectrophotometric

color calibration um the difference between these is spectrophotometric also takes into account the

your camera sensor like the uh the noise and the filters and things in your actual camera there's

a bunch of options here for when you're working with a spiral galaxy like M101 you can leave the

white point reference as average spiral galaxy um if you're working with another type of galaxy you

might want to pick that if you're just unsure of what to choose for the white point reference

usually a G2V star is always a good choice okay the next thing here is we have to put in our type

of camera so you can see right now it's set to mono here we're working with one shot color so

we'll switch this to the other side and now you can see it says one shot color and then it has

a bunch of different sensors here including the sea stars um it only has a limited range of Canon

cameras so I'll just pick the newest Rebel which is the 600D that's probably going to be close

enough to the 2000D that I'm not worried about the differences uh same thing with the low pass filter

it only has one Canon low pass filter in here the 350D but Canon's low pass filters are generally

very similar between models so I'm just going to leave that uh in there okay okay and then I wasn't

using any additional filter with my Canon camera if you were you could put it in here all the rest

of this stuff you can probably leave alone it's going to consult an online catalog uh called

the Gaia Archive which is from a huge survey of stars it's the biggest star survey uh available

and over 30 million stars so it's going to use that information from our plate solve to know the

uh color temperature of all of these different stars in the image and then create a white balance

reference based on the star information as well as our camera information that we put in here okay so

I'm going to click okay and what it's doing right now is it's downloading information I'll just move

this to the side here so you can see you can see it's downloading information about uh from the

Gaia uh catalog it pulled down a particular uh CSV for this particular area of sky and now it's

done and it plotted lots of different um stars and then it gives you these plots uh which is

a summary of what it did and then there's a trend line and that's how it figured out the

white balance reference for this image and then it gives you some information about that the white

balance factors that it applied down here okay I'm going to go go ahead and click close and just to

show you the difference here was before a lot more green in the image and then here's after we color

calibrated it and you might think well it took out color but don't worry about that because we're

going to saturate the image later and there's plenty of color information in here there's no

worry about that so um even though this is just a very bright picture where you might think oh well

it turned it white um it didn't really um it's just that the auto stretch is very aggressive

okay next up now that we've color calibrated I want to remove the stars from the galaxy so I

can work on them stretch them separately bring them both into GNU image manipulation program

and continue working on them separately in there so I'm going to do that with Starnet Star removal

which is under star processing and the defaults uh that are checked pre-stretch and generate

star mask are exactly what I want sometimes I'll turn on this one right here but for now we're not

going to do that because I actually don't want to make just one finished image at this point

i actually want both the starless and the star mask separate if this execute button isn't blue

for you it means you don't have Starnet installed and configured correctly in serial i will have an

addendum that covers that for your system uh but at the end of the video anyways I'm going to click

execute and this runs pretty quickly on Macs if you uh have that new version and a newer Mac

and it did a good job here you can see it left um this this this this this and this 1 2 3 4 5

six maybe seven there's like six or seven other small galaxies around M101 um these big sort of

blemishes are star halos which you may or may not have depending on the quality of your telescope um

but don't worry about them because when you add the stars back in they're not a big deal they're

also quite dim um and we're going to actually make this sky pretty dark so I'm not worried

about that it looks good uh it did a good job so I'm going to go ahead and stretch this do an

initial stretch on this result here so to do that I'm first going to make a box over just an empty

area of sky make a selection so I'm just going to click and drag on an empty area of sky in the

image then I'm going to turn off the autostretch I'm going to go to linear mode and then I'm

going to open up image processing stretches generalized hyperbolic stretch transformations

and move this to the side a little bit and I'm going to sample this uh box right here with the

symmetry point by just hitting this little eye dropper and that does an averaging of all the

pixel values in that box and then this is uh the average number if you normalized the pixel values

from 0 to one so if you normalized them we'd be at 0.03 so it's very very dark um then I'm going to

turn up this local stretch intensity that sort of uh determines the shape of this curve that we're

going to do and then the stretch factor determines how far you bring that curve up and so what I'm

trying to do right now is just basically bring this histogram peak which is represents all of

the data all the pixel values up and out away from this left hand side then I'm going to go ahead and

apply that and then I want to basically bring this whole peak back down to the left so to do that I'm

going to go up here to type of stretch change it to linear stretch and move my black point to the

right until this histogram peak is back over to the left apply that and then repeat go back to

GHS sample the symmetry point one more time bring the local stretch intensity way over to the right

and now when you do this the galaxy for me is popping out but I don't want to go all the way

to there because I think that we're going to start losing detail in the core so I'm going to

do something more like this where you can see that you can still really see the spiral arms even in

the core of the galaxy so I'm not stretching too much at this point maybe just a little bit more

and then I'm going to reset my black point one more

time for this initial stretch I don't like to stretch too much and the fact that you know

we just have uh that most of the data is way over here on the left hand side is normal with an image

like this one where most of the image is black um and we just have this bright galaxy right here

um which is represented probably by just a tiny bit of data um in terms of this histogram so I'm

not worried about exactly what the histogram looks like I'm mostly just doing this on what

the image looks like and I want to make sure that um I'm not blowing out any detail in the galaxy

core when I'm doing this initial stretch I'm also not concerned at this point that um we don't have

a lot of saturation we're going to add that later it's a lot of fun but don't worry right now if uh

the photo looks sort of bleached out or something like that and I'll go ahead and save this and

I'll save it as just starless and you can if you are moving to the GNU image manipulation program

save it as a Fitz file that's fine sometimes I in the past I have recommended switching to TIFF at

this point to avoid issues but let's go ahead and try fits and see how that well that works

and I'll just leave it on 32bit okay now let's go ahead and open up the stars and stretch those so

that's the star mask result here and for this I'm just going to use just a normal histogram

transformation sometimes I'll use an arc sign but it's warning about clipping pixel values so let's

just rescale values that are under zero so yeah anyway sometimes I'll use an arc sign for this but

um with a lens like the tear I don't want these stars to be like super colorful anyway so I'm

just going to use a histogram transformation and just take this middle slider and move it over to

the left two times applying it both times and that looks good we can always stretch the stars more

once we move to the GNU image manipulation program and then I'll go ahead and save this as just

stars.fit and save and we can put these in 32bit too okay that's it in Siril. I'm going to go ahead

and quit out of serial and open up the GNU image manipulation program I'll go up here to the file

menu and choose open as layers go to my uh folder which is in other locations volumes x10 pro yes

allow it to see that go into m101 okay and it was stars.fit and starless.fit so I'm just going to

hold down command on Mac or it would be control on Windows so I can click and click and select both

of those to open them both as layers and click open and it's asking replacement for undefined

pixels do I want to turn them black or white i want to turn them black i don't know why there's

undefined pixels but let's see what happens okay I think this looks good and let's work on this

starless image first and I like to go to the view menu and turn off rulers and turn off the

layer boundary which is that yellow dotted line just because I find them distracting I'm going

to start by making a selection because I want to start saturating the galaxy but I don't want

to saturate the background because I don't want to bring out any color noise in the background

so to do that I'm going to choose my select by color tool up here which is the fourth tool the

one on the far right in the top row of tools and I'm just going to click in the background and by

doing that it basically selects the background as you'd imagine and I guess a few hot pixels

but also the galaxies um if yours doesn't look like this try changing the threshold and then I'm

just going to go ahead and invert this selection so I'm going to go to select invert so now it's

just selecting the galaxies and a few hot pixels if I was worried about saturating those hot pixels

um I could go to select toggle quick mask and paint with a brush whoops

in black over them and then toggle the quick mask back off and you can see I did an okay

job but didn't get all of them but I don't really care i I think this is good enough um if it picks

up a few pixels and saturates them it's probably not that big a deal cuz our background already

looked mostly black now I'm going to with this selection I'm going to apply something from the

colors menu and it will add it as a layer effect to the starless layer but with only this selection

active so it will saturate but just the galaxy and not the background so I'm going to just start

with just a basic saturation adjustment and let's bring that way up basically to max and hit okay

okay now what's really cool about the new GNU image manipulation program is if I

uh zoom in with just the plus button and I go into this little effects menu on the starless

layer there's a little eye icon next to the saturation adjustment now and I can turn that off

and on and see exactly what it did and you can see before we have a fairly ghostly unsaturated galaxy

and after we have a nicely um fairly subtle subtly saturated galaxy um but I can see already that the

colors that it's bringing out the reds the blues and the yellows are accurate because we did that

color calibration earlier okay um but this first go at saturation wasn't enough for me i want to

go a little bit further so I'm going to go back to colors keep my selection active and go to hue

dash saturation and what's cool about this one is you can bring up of course the master saturation

that's the default but you can also go into the different colors and bring up yellows and bring up

uh magentas and bring up cyans right and then it's only affecting the saturation on the cyan

or the yellow or the blue and I often like to do this with galaxy work just to sort of give it my

final artistic touch of what I think the galaxy should look like i know that now we're sort of

wrecking our accurate uh color calibration so if you are very concerned about like this

being like super color accurate don't do this just use master saturation boost but if you

uh want to just sort of try to like make the image look as pretty as possible I sometimes like going

in and messing around with the saturation of the different colors uh independently and um I

should point out uh that another cool thing with the new GNU image manipulation program in 3.0

O version is you can always uh bring this layer effect menu back up and double click and you still

can dynamically edit any of this so I'm right now dynamically editing the saturation of the yellows

in the in this selection but I can go back to the blues and dynamic dynamically edit those as well

so that's really cool the other thing I love about this is just like it keeps the selection

active so now I can just keep adding more filters so we went through a couple in the colors menu we

can also go here to the filters menu i'm going to do that go down to enhance go to sharpen

uh parentheses unsharp mask and play around with this a little bit so the radius is the

size of the kernel that it's going to sharpen so you can see you can really overdo it that's

with a radius of five pixels but if I try something more like 2.2 pixels I think that

looks pretty good and then this is the amount of sharpening again you can easily way overdo

this by applying a huge amount but I just want just a little bit of um sharpening here so I'm

going to apply it at actually the initial was pretty good 0.5 maybe just a little bit less

0.4 okay and then I'm going to go ahead and turn that off and

on and see if I liked what it did i do i mean it maybe doesn't look

uh super great this far zoomed in but let me back off just a little

bit okay if I back off a little bit I think that

that sharpening does help define the galaxy a little bit better okay I'm

going to keep the selection active and work on the stars a little bit

here so let me zoom back

out okay I'm going to put the stars on top

and let's go ahead and turn their blending mode from normal so you see right here up

at the top of the layers panel it says mode normal let's change it to screen oo I think

this is already looking pretty good but I think maybe the galaxy should just be

a little bit brighter so I'm going to go back here to my starless layer and

with that selection outline still selected I'm going to add a curves

adjustment and I'm just going to try just adding a little bit of an S-curve so we'll

just add a little bit more global contrast to the galaxy itself and maybe make it just

a bit more the star of the show okay so I'm just going to add a very subtle S-curve there

and now we're finally done with the selection so I'm going to get go to select none oh wow

so I think this looks excellent i might not do anything else to it but let me just see

if there's anything else I want to do i re I mean really the stars I don't I just got

that the stars are just what they are uh right out of the right out of the serial i didn't do

anything with them and I think they look pretty good with just that histogram transformation

i don't think they need to be stretched any

further let me just zoom out and zoom back in and make sure I like it

yeah so I mean if if you if you got to this step and your stars look sort of weak and like too dim

you could always just click on your stars layer and go up here to colors and curves and you know

continue stretching your stars if you need to or you could do a reverse curve and and make them

um smaller and dimmer so uh that's up to you but um I really like actually how the stars

look right now so I think I am done if you think that you'll return to this image in

the GNU image manipulation program you can do file save and save it as an XCF

file but then when you're ready to share it with the world you can do file export and

uh go down here to select by file type to see all the different types of files

it's probably the most common uh file type for sharing is the JPEG so I'm just going

to choose JPEG image click export and it will give me some JPEG uh options here you

know like quality so quality is how much compression it will apply so if you want

um the least amount of compression set quality to 100 but that will make a bigger file so but

let's try it let's set quality to 100 and let's see how big the file

is okay with quality at 100 it's still under 10 megabytes so that's that's fine it'll save

it'll send on any email attachment and uh I think that's that's a good compromise so let's go ahead

and open this up and see what our final image looks like okay and that's it so for me what

really makes an image I think is just getting the technique down uh on the field making sure you're

in focus and dark skies make a huge difference for galaxy shots like this um it's not so much having

you know the perfect equipment or the perfect software dark skies just matter so much um for

getting a nice uh picture of a galaxy where you have nice color um and and nice detail so

hopefully this uh walkthrough of the free software was was helpful and if you're watching this and

you're into processing your astrophotography the next question is what do you actually do with your

astrophotography and what I'd recommend is setting up your own personal website where you have

complete control over all aspects of its design and Squarespace the sponsor for today's video

makes that super easy um not only is the website easy to set up but you can also have your own

domain name like this is my Squarespace site at nicocarver.com my name and for me it's just about

having you know a very professional landing page and online portfolio of my photography so mine is

fairly simple but I think elegant and it's really easy to change if I ever want to add a photo of

course that's incredibly easy to do if I ever want to change the design Squarespace has an excellent

design system called Squarespace Blueprint that makes that super easy to do i also like that

there's so much built in to Squarespace including a spam-free contact form that I really appreciate

and tons of selling tools i haven't turned this on myself but it's nice to know that if

I ever want to sell prints of my work or services through my website that's really easy to turn on

and set up and then customers can pay however is most convenient to them whether it's PayPal

credit card or Afterpay so if you have a need for a website and want to try this out head to

squarespace.com/nebulaphotos for a free trial when ready to make a first-time purchase of hosting or

a domain use code nebulaphotos at checkout for 10% off okay here we go with installing the software

on Windows so we're going to start by just downloading the packages so I have the different

websites up here starting with serial this is the GitHub for GraXpert where you get the new beta

versions so just scroll down here until you see assets and twirl that open and then download the

Windows package MSI file starnet you're going to go to download uh scroll past all this Pix

Insight stuff till you get to the section called command line tool and then download the zip file

for Windows and the GNU image manipulation program just click on this big uh orange download button

and then down here click on download directly okay I have my downloads going directly to my

desktop so they'll all show up right here and then we'll just go ahead and install them it's pretty

straightforward on Windows but I'll show you some things that we're going to do to configure serial

after we have everything installed so that you can work with Starnet and GraXpert directly inside

serial okay let's start installing things so I'm just going to double click on each setup package

okay when you open up GraXpert for the first time or actually every time since

this is a beta release it'll just give you a little warning that this is a beta

release you can just say okay um I'm going to go ahead and load an image

this can be any image doesn't matter what it is so I just want to show you

something if you go down here to background extraction you can see the default is AI and

if you go over here to the advanced menu on the right hand side you can see there's some

different options here for AI models for background extraction deconvolution and

denoising and I just usually pick the latest for each of these but the first time you do any

of these things like let's say if I go in here into deconvolution and click deconvolve image

it will say do you want to install the AI model should I download it now and you can say yes and

I'm just pointing that out because um you need a fairly good internet connection these are fairly

uh slow downloads so just so you know that's going to take some time and so maybe if you want to do

it all ahead of time before you start processing you can just load in any image and uh get the

AI models this way by first selecting them over here in the advanced tab and then just clicking

deconvolve denoise whatever and the first thing it'll do is actually download uh the correct model

that you've picked over here for Starnet come zipped and you can just rightclick and say extract

all and then just move this folder somewhere where you're going to keep it permanently so

I just put it in my documents but you can also put it in your uh programs wherever you want

to keep it is fine okay then finally let's uh install click a new image manipulation program

here okay last step is we're going to go back to serial here and we're going to

configure it so that Starnet and GraXpert work well within serial to do this go over

here to the hamburger menu in the upper right the three horizontal lines click

on preferences go down to miscellaneous and click where it says software location

click on this first folder where and then next to it where it says must point to a valid Starnet

executable and like I said I put my folder in my documents but wherever you put your Starnet folder

is fine all you have to do is just find that folder and double click it and then just find the

uh file within there called Starnet++ this is the actual executable and when you click

open you should say you should see right in this box right here it says starnet++.exe you don't

need to do anything with this second box okay and then this one where the GraXpert executable is is

probably the weirdest thing uh for Windows it's a little it's not in your programs folder um go down

to this PC click on Windows go down to users and then find like your user account mine is called

nebul for some reason double click it and then it doesn't show there's a hidden folder in here

called app data so if you just go up here to the path and do slash AppData like that capital A pp

capital D ata and press enter it will then bring you into this app data folder go into local okay

then once you're in local you might be thinking oh I'll go into the graXpert folder but nope go

to programs somewhere here there it is programs then there's GraXpert and then within here find

the GraXpert icon right there and click on it and click open and then you should see GraXpert.exe

right there and then click apply okay then go ahead and open up an image doesn't matter what it

is and just go to the image processing menu go to star processing starnet star removal and make sure

that this execute button is blue that indicates that starnet star removal is working then go

back to that image processing menu and go down to graXpert interface it will tell you this is

about to be deprecated you can just say okay um but just make sure that you can access this and

for now you can use this graXpert interface uh but eventually they're going to move GraXpert

into the Python scripts all right and that is it for Windows installation so I'm here on my

Macintosh computer and I'm going to go ahead and download the software that we need so there's

two different versions there's the Apple Silicon version and there's the Intel version if you're

on an Apple computer that is an M chip then you have Apple Silicon to figure out if you are on

one of those you can click the Apple up here and click about this Mac and if your chip says Apple

something Apple M something then you're on Apple Silicon if it says Intel right here then you're

on Intel so anyways I'm going to click on Apple Silicon it downloads the latest version which

is 1.4.0 beta 2 as of May 2025 and then they do have some options for uh donations right

here so I encourage you if you enjoy serial to go ahead and donate to the team so they can continue

putting out these awesome updates let's go ahead and just download all the software while we're at

it here while we're in the browser so I'll go on to the next tab this is GraXpert. GraXpert does

have a website but they also have a GitHub and I would recommend these um newer release candidate

versions on GitHub because they have some cool stuff like deconvolution which we'll try out so to

download the latest release candidate we go down here on this release page and right here where it

says assets just twirl that open and then here's all the different versions for Windows Mac and

Linux so I am on uh Apple silicon which is also uh the the architecture of the chip is called ARM so

um if you were on Intel you would choose AMD if you're on Mac uh Apple silicon you'll choose

ARM right here all right so we have that one now let's go ahead and move on to Starnet and

uh click into the download page and there's a bunch of different versions here sort of a lot

of information if you are on uh an Intel based Mac you'd scroll down here to where it says command

line tool and click on that if you are on a Apple silicon uh ARM based Mac then you'll right here

on the downloads page in this gray box right here you'll click on the word here where it says see

new experimental CLI distribution for Mac OS and this is the one you want so go ahead and click

download there and it will be called Starnet 2T Mac OS.zip that downloads okay while that's

downloading let's go ahead and move on to the GNU image manipulation program uh this one it's pretty

obvious where to download it's this big orange-ish reddish button right here at the top and then it

should just um automatically know which operating system you're on so since I'm on Mac right now it

just defaults to that but you can see there's also Linux and Windows here and then uh it

also detected that I was on Apple Silicon but if you needed the Intel version you could see it's

right there anyways I'm going to not do it via Bit Torrent I'm going to do it directly so I'm going

to use this orange button on the right and there we go so that's downloading as well now I have my

browser set up to download things to my desktop so they're all right here on my desktop if you

have your browser set to download things to your downloads folder you would just navigate to your

downloads folder and they'll all be right there now we're ready to install these and I'll do them

in the order that I downloaded them so I'll start here with serial at the top so I'll just double

click this DMG file and then this is a standard way one of the standard ways of installing

applications on the Mac it's just asking you to drag the serial app to your applications folder so

I'll just go ahead and do that I'll drag and drop and if you are updating your serial application

you would want to delete it from your applications folder before doing this and you can delete it

just by right-clicking and say move to trash or you can drag it to the trash bin down here

okay once you've done that you've then installed serial so just to make sure that it's actually

working um I can just go to my magnifying glass up here near the clock and type in

serial and it says serial is an app downloaded from the internet are you sure you want to open

it i do I'll click open and there we go so I can see that it is working and we can delete both the

DMG just by dragging it to the trash and eject this temporary drive that it created by dragging

that to the trash too okay now let's install GraXpert so I'll double click the DMG and same

thing just drag this to the applications folder with that done we can drag these both to the

trash and just double check that GraXpert installed by typing in grax and hit enter

okay and it says this software uh cannot be checked by Apple blah blah blah okay okay so

what I'm going to do is I'm going to go up here to the Apple and click on system settings and go down

here to security privacy and security and then scroll down and you can see right here GraXpert

was blocked from use because it's not from an identified developer just click open anyway and

then I'll just use my thumbprint reader or you could use your password and click open and then

once you've done that once GraXpert will open just fine every time after that okay and you can see

that I'm on the release candidate version because there's this tab here on the side on the left

called deconvolution and it has object only stars only strength and so forth okay so we can quit out

of GraXpert now that that's installed next up we have Starnet i just double clicked the zip file

to unzip it and then we can delete the zip and the way that this one works is you're going to

want to put this folder somewhere permanently um so I keep mine in my documents folder you could

also keep it in your applications folder wherever you want to put it and this holds actually the

uh application now there are a couple things that we want to do in addition to install it

okay if you look at the README right here it does have installation instructions um there's just a

couple little things that um you might run into problems with even following these instructions

so I'm just going to go through it with you but I'll keep the instructions open so we can sort

of try to follow them so we can click on the magnifying glass up here type in terminal it

says open terminal and then run chmod plus x install.sh okay what it left out there is we

need to actually be in the starnet folder so let me go ahead and put it where I want which is in

the documents folder so I'll just drag it in there okay so you can see it's now here in my documents

folder so I have to actually go there in terminal so I'm going to type in cd for change directory

space and then I'm going to drag in this folder into terminal just like this and press enter and

so you can see now I'm actually in that folder so when I run this command this change uh chmod

command which changes the permissions of that file it will actually know what file I'm talking

about okay chmod plus x starnet 2 okay and then we're going to also do that chmod plus x on the

install.sh file okay and then the README says okay just run the install.sh so we're going to do dot

slash install.sh give it our password and I got this

error uh this is a copy error saying user local lib is not a directory so I'm not

sure what that meant but I knew how to find directories on my Mac so what I did was I

went to Finder here went to the go menu went down to go to folder and typed in

/usr/local double clicked and this is what I found i found that under usr local there

was no lib directory so we can just make one a directory is the same thing as a folder so we can

just right click and choose new folder give it our password again and type in lib enter okay so now

there's a lib folder for it to copy some things into so we can now type in that command again

/install.sh enter and this time it worked and I know it worked because look at all

of these files that it copied into that folder okay with Starnet installed now we'll move on

to the last one gnu image manipulation program just double click the DMG drag the application

icon this uh character is named Wilbur to your applications folder

and then you can drag both of these to the trash okay and just to make sure that worked let's try

opening up the GNU image manipulation program click open when it asks and yes it is working

okay and the last step here is we're going to configure serial so that it works natively with

Starnet and GraXpert and to do this we go up here to the preferences menu which is under this little

hamburger menu in the upper right so click on the little three horizontal lines that's called the

hamburger menu then click on preferences and then there's a bunch of different preferences here but

the one we're concerned with is miscellaneous which is the final option on the left and then

go down here to where it says software location and if you've never done this before these three

options here will be empty mine are filled in because I've done this before but I'll show

you how to do it so for the first one Starnet must point to a valid Starnet executable you're

going to click the little folder icon and go to where you've stored your Starnet folder

for some reason in the latest version of serial uh this file picker is in Dutch rather than in

English but it's still pretty easy to figure out so documents is documenting and um then here's my

Starnet folder and then I'll just click on Starnet 2 here and when you click on it you can see this

is an executable I'll click open and then you can see it says Starnet 2 right there same thing with

this next one down you just want to click on the weights file it ends in pt click open and then

GraXpert is a bit trickier uh for some reason you can see right now it says graxpert.app but

that's actually not I don't know why it defaults to that because what you have to actually do it's

not going to let you if you just pick graxpert.app and click open i know that looks the exact same

but it doesn't actually work what you have to do is you click on this folder and then open

up finder in Finder go to your applications and then right click on GraXpert click show package

contents go into contents then click on MacOS and then there's the actual executable and so just

um drag this from your finder to this file picker window here and then it will actually bring that

up in the file picker and then click open here and then you can see now it says just graxpert right

there that's the executable now after you hit apply and then open this back up you can see it

switches to graxpert.app i don't know why it does that but anyways you have to do it the way I just

showed or it doesn't actually work um but now just to double check it let's go ahead and open up an

image and if we go to star processing star net star removal you can see the

uh blue execute button or the execute button is blue so that means it's working and if we go to

graxpert interface you can see this comes up and the apply button is blue indicating

this is working as well and since this video is over 20 minutes long you're now seeing the names

of everyone who already supports this channel on patreon.com/nebulaphotos and I can't thank

everyone you're seeing here enough for their support because between sponsors and ads and

the support on Patreon especially I'm able to run Nebula Photos full-time and that has been

such a blast and if you're are able to join over on Patreon there are a number of perks that I've

split into different tiers that you can check out but everyone who supports me at $1 a month and up

gets direct messaging support with me a monthly Zoom call access to the Patreon channels on my

Discord and much more so I would definitely uh appreciate it if you could take a look if you are

a supporter of this channel it would really mean the world to me if you could join over on Patreon

well till next time this has been Nico Carver and I'm wishing everyone watching clear skies!