Easy OSC Hubble Narrowband Guide! - PixInsight Image Processing/Editing

hi there everyone so today i'd like to show you a tutorial on how to bring your one shot color narrowband images uh through filters such as my own lx stream or l enhanced filters does many of these available uh how to bring those particular images from looking kind of just like salmon pink and reds uh into a hubble style palette and it should be quite easy to follow along with i'm going to be using recently taken data through my 2600 mc pro and optilong lx stream filter it's not particularly great data because it's got very short integration time it's only 17 five minute shots but hopefully it should be good enough to show you what i'm talking about and you can pause this video and try and follow along with your own data perhaps so there's a couple of prerequisites that you'll need to follow along with this uh tutorial and that's going to be that you will need the easy processing suite installed um and you'll also need to have starnet plus plus installed and working with the opixin site so with that said uh let's get started so the first thing you're going to want to do is we're going to crop this image so as usual i'll apply an sdf so i can actually see what the data looks like while leaving it in a linear state and as usual it's going to come up quite green to begin with but anyway so we're going to use a dynamic crop tool and just bring the edges in just a touch this is just to get rid of any stacking artifacts that may be present and that should do we don't need to be too conservative with this because it's just a tutorial okay that's done now before we pass this tool on into the easy processing suites easy d noise script it does require that you provide it with color calibrated data so we're gonna run through a full color calibration just using the automatic settings though uh before we pass it on to that tool so the first thing we want to go to do is color calibration and background neutralization now you'll know i'm just going to leave everything totally standard and just drag and drop that over and each time i will apply a new sdf just to show you the difference that it's made you don't have to do this you can just trust that the tool is going to be working but for the sake of just showing it visually i'll keep doing that on this video so the next thing after the background neutralization is going to be a color calibration itself again all standard settings just drag and drop worth mentioning i guess at times uh some of these processes in pixie site even with quite a fast computer are gonna take quite a long time so i'm obviously not gonna expect you guys to sit through that i'll just fast forward through uh so if you see a progress bar skip along then you know that's what's happened and i'm not actually processing on a super computer uh okay the next thing so we're gonna need to go to noise reduction and get rid of all this green using s c and r subtractive chrominance noise reduction again everything's standard just drag and drop we don't want to over complicate this so now before we move on and pass this data on to the next tool we can see that there's a gradient hopefully you can see that on your screen that it's lighter on the left hand side and darker on the right now i'm going to try and deal with that using dynamic background extraction i'm going to leave everything standard the correction that we're going to want to apply would be a subtraction and for the purposes of this i'm going to discard the background model you may want to leave that unticked so you can actually see perhaps any other problems that occur you can sometimes see that by seeing the background model that it generates if you're having trouble extracting out any gradients but for now i'm just going to go ahead and place some sample points you don't have to be massively careful with this just as long as you're not selecting bright stars uh in particular bright parts of nebula which could confuse it into thinking that that is also part of the background which is clearly not uh i suppose it's worth mentioning the effect that that could have as well and that's that it'd apply an overcorrection in those areas which would be an erroneous one as well okay so i think that's about okay for that last quadrant maybe a few more sample points and yeah i'm gonna say that that's fine for the purpose of this i'm gonna go ahead drag and drop we'll apply an stf to this new image and i'd say that largely that's got rid of the gradient and i'm happy to continue on with processing now at that so i'll close the dynamic background extraction and i'm actually going to close this first initial image because we're not going to be using that any longer for the for this all right so the next stage of the tutorial it's going to require you to open up your scripts go to the easy processing suite and select easy denoise it's going to warn you that it's under heavy development it could cause your pc to crash and it certainly could but i've had no problems thus far um so you're gonna wanna in this top drop down the first thing is like the image that we're working on i've kept everything neat now so there's only one image to select which is this one now without getting too uh in depth with this basically if you just hit run easy denoise in most cases it's gonna work just fine so i'm gonna go ahead and do that and while that's processing because this one actually does take quite a while i'll just skip ahead for you guys at home okay so we're above 10 minutes later now and that process is finally finished um one thing i'll do now to prevent that taking as long in the future will be to actually resample this image before moving on with this tutorial i wouldn't recommend you probably do this with your own images at home because you're going to want the absolute maximum quality maybe you could resample right before saving but just to make the rest of the processes go faster i'm going to do it now so that's going to be geometry and resample for anybody who wishes to know about this we obviously want these sorry that's it the correct image selected i'm just going to resample to 50 for this which should make processes take a quarter as long uh total okay now that's done i'll just close the resample box i can close these uh extra masks that it's generated just to get rid of any clutter from the screen and it's actually done a beautiful job of denoising the image uh while leaving it still in a linear state which is perfect uh but as it happens right now we we're actually going to take this image from the linear phase uh into non-linear and we're going to use the easy processing script again for this so if we just go to easy processing suite and easy soft stretch again there'll be another warning so you could crash but we need to select the right image again so there's only one on my screen you can reset these settings uh back to the standard values i'd actually turn these down for processing another image where i wanted a slightly more general stretch but for the sake of this and going with everything basically at default values as much as possible we'll just run it fully reset and we'll run that now unfortunately that one doesn't take too long to perform and this is now a non-linear image okay so the next thing we're going to do now is we'll actually run starnet plus plus you can find this process all processors down in the s's here you go start it we do want to tick create a star mask because we're actually going to use that for our star image um and this is as simple as just drag and drop again leave it on the same standard stride setting no need to mess with that and again i'll just come back to you in a moment once this is done processing all right that one didn't take too long at all thankfully um probably due to the resample but we can now close starter and you what you'll be left with is a stylus image and one that contains all your stars hopefully it should have done as good a job as this one has which is almost perfect uh and what we can now move on to doing is working on this stylus image by breaking it up into its component parts to begin with so to do that we're going to make sure we have this image selected go to process channel management and channel extraction we're going to name our channels real quickly very very simple naming convention uh g and b because i want very short identifiers uh reduces some visual clutter make sure you're in the rgb space and just drag and drop that on the image okay we can close that down i'm actually going to minimize blue because it's quite a noisy channel in this image whereas green uh hopefully it's apparent for you on this recording has considerably less noise while containing all the same data and some more really so again minimizing blue we're not going to be using that perhaps moving off to one side so we're going to be mainly focusing on this red and green images now you might be wondering what about using a third channel for when we recombine things and that's what we're about to do right now so i'm going to take this red image put it somewhere where i can keep an eye on it move my green image to one side and just using curves while selecting the green image and a live preview i want to stretch the green until it has a similar brightness to this red and to perform this i'm literally just going to drag this center point upwards so you can hopefully see it's starting to get close to matching the brightness now you don't need to go quite all the way and then i'm going to drop the background brightness to again match the red channel and to do that we're going to try and drag this uh curve into a gentle s hopefully that is showing for you guys at home but that's now got a fairly similar background level uh to this red image and that's perfect for what we want to do so i'm going to go ahead and apply that and now reset the tool and close it we can also get rid of this preview so now to make third image which we're actually going to use in this case as a green channel green will become blue and red will still remain as red and actually also our luminance image in this case we're actually going to use a little bit of pixel math which is a very simple expression uh which you so you'll find that here process pixel math and open that up now because we've got very simple and short identifiers just r and g it's going to be as simple as this so r times 0.6 plus g times 0.4 so it's going to give us uh an image containing 60 percent of the red channel 40 of the green channel and it'll be a nice blend between the two to use in the the actual green channel on the recombine uh now we need to set a new destination uh of create a new image you can leave an auto identifier but we don't want it to replace either of these it's very important uh go ahead and apply that and as you can see the pixelmath is now finished and we're left with a new uh new image so we're actually going to change the identifier on green now and just for the sake of keeping things kind of easy to understand in the next part of recombining the image i'm actually going to rename this to blue so now green is going to become blue i'll just have to close this i shouldn't really have left that one open all right identifier b and identifier now that we've freed up g because we're going to use this in the green channel which call it g all right so we're left with three separate images now um and we can actually begin to recombine these using the lrgb combination tool uh i've actually got a shortcut to it right here so i'm just going to use that so in the place of l we're going to use red because i want that to actually be our luminance image it's very low noise and it looks beautiful the red channel is going to be red the green channel we're actually going to use the freshly created pixel math green and the blue channel we're going to use the natural uh what was the green channel sorry if this is getting confusing but if you've done it a few times it'll start to make sense uh we are going to apply some chrominance noise reduction and just go ahead oops and apply that globally and this should generate us now uh an almost hubble-esque image right from the the get-go rather than just those uh so many images that you often find you get with just a basic process kind of like the one that we were working on here in this stylus form um so you're probably going to notice there's not nearly enough blue in this and the way we're going to deal with that is just by selecting it and opening the curves tool now you probably know that you can select individual rgb channels and that's exactly what we're going to be manipulating to bring those blues up so i'm going to select the blue channel again drag the midpoint of this slider and just drag it up sorry i need a uh a preview open and so we're making the background blue but if we just drag this again into an s you'll notice the background values go back towards looking more uh neutral and black whereas the car where it should be turning blue is actually doing so and we'll leave in this region of the image hopefully that's apparent for your arm largely untouched without the use of any mask and things like that which they take extra time but by all means you should feel free to perform that if you're confident with masking um that actually would be a better way to do this but this is quite a rapid way to do it so i'm satisfied that that's brought the blues up quite nicely for a single pass i'm going to go ahead and apply that and reset the tool so analyzing it again i'd say this is too yellow and i like to see a more reddish tinge so i'm going to go to the red channel now and gently bring that up a little again while bringing it into an ever so slight s and apply that satisfied there reset the tool again and now the background is looking quite green so we're going to bring likely down the greens to deal with that a little bit you see from naturally left alone there to just an embassy slight reduction and all the colors change quite dramatically so you need to be careful with this that you don't overdo it but i'm happy at that and i think a final touch up on those blues again just drag it upwards because i want them to increase again so if you if you drag things up it'll increase the value down and it'll decrease it um so get the blues somewhere where i want them around about there i feel looks quite nice and just drag it into that very gentle s again okay for the sake of perfectionism i think in terms of my eye uh i'm gonna actually bring the reds up again just a nervous slight touch again you can preview this what effect it's had back and forth uh no i'll leave it as it was okay so uh you can hopefully see now that we've got quite a nice uh colorful yet slightly muted um stylus hobble-esque image so again we could use curves to bring this saturation up just using the saturation tool on the end again this works just like all the others uh dragged upwards it'll give you a brighter in this case more saturated value and that's perfect so yeah around about there looks great close the tool i'm actually going to pass this through a color calibration before continuing to make sure things are aligned nicely and yeah i'd say that is an improvement so i'm actually going to go with that even though it's took it back to slightly more yellowish um the color calibration and the subsequent scnr i did afterwards i've taken this into quite a nice uh a nice looking state i would say by all means this is just wes i suppose when you're processing in hubble uh colors or hubble s colours in this case since it's just a bi-color image it's really down to you the processor as to what you want to settle on and what colors you find appealing so i mean you might like this one better uh i happen to like this one better and that's all fine it's just artists choice if you like because at this point i suppose in processing it does just become more art so now we're left with a nice colored image there the last thing we're going to want to do is you could you could save this out as is by all means lots of people like stylus images but i i'd like to put the stars back in but i don't want them to overwhelm the image uh and i also know that some of them look errors are slightly green so i'm just going to pass scr over this real quick sorry if i'm rambling a little bit of this it's just it's kind of a late night tutorial and i really should be in bed but i just thought i'd make this uh as i want to keep content coming out for you guys as you've all been so supportive um and i thought this was worth sharing so yeah i think the scnr has had a change very subtle um but yeah before i put these stars back in there's one more tool that we can use as a really lovely uh effect especially on a narrowband image where you don't really want to swamp the image with big bright stars so we're going to go to process and use morphology now so morphological transformation because this is already going to act as just a star image we don't need to create a mask for this um at one value which is full strength for one pass i think it'll probably be a little bit strong so we're going to turn that down to 0.8 and just drag and drop and hopefully you can see it looks like someone's turned the lights out on most of those stars they're still visible but they're greatly diminished and certainly not going to dominate the image anymore so back to identifiers now before we move on to the last phase of this image um so this one's currently image 21 it's just where we're currently at this point in processing i'm just going to change the identifier to stylus to keep things nice and simple and this one that's created a star mask i'm going to change that to just stars because we're going to be using pixel math again uh and i can never be bothered typing out all this massive long identifier when i could just change the uh the identifier to something more suitable for the purpose um so without messing around too much more we're just going to go on our process uh and pixel math we can clear this previous expression and this is a very simple expression it's just going to be stylus plus stars uh an auto identifier is fine as long as it's going to create a new image oops i always get this the wrong way around i'm gonna reply that uh and hopefully you should see now it's created a lovely image with the stars put back in not dominating it i'll just make this a little bigger yeah the stars definitely aren't dominating this image and we've got these beautiful hubble colors uh coming through from just what was initially a bi-color image uh this is the same data as i used when i uh made my recent video photographing the rosette nebula uh taken on a half moon night or it might have even been more illuminated than that but it's really limited data but we've got a fantastic result just from using these few basic tools that really don't take much installing but the power that they give you over the image is not really to be underestimated in any way i'd say that that's certainly a finer result that i got first time around with processing um so yeah without any more messing around i guess because as i mentioned earlier it's all down to you what you actually want to sell on um i think i'll call that finished for this tutorial uh and hopefully you're able to follow along with that at home and get some use from it maybe pick up a few tips and if it was useful and you liked it please consider leaving a like uh commenting subscribing it all really helps this little channel of mine to grow which is ideally what i'd like to do because i'm really enjoying making these videos for everyone so uh yeah thank you very much for watching and see you next time cheers

Channel: lukomatico

Views: 7,132

Keywords: OSC, colour, one shot colour, narrowband, hubble, hubble pallette, hubble palette, dual narrowband, optolong, l-extreme, l extreme, l-enhance, filter, 2600mc pro, zwo, asi, skywatcher, PixInsight

Id: OVb1_Nqcs5I

Length: 23min 12sec (1392 seconds)

Published: Fri Mar 12 2021