Original subtitles

hello my name is Frank seckenheim and

you're watching the Astro photo cast in

this video I would like to introduce the

latest version of grep Greer 3.0 which

now includes AI based noise reduction so

let's

go the most important innovation of

Greer 3.0 is the ability to do noise

reduction of astronomical images based

on artificial intelligence this AI D

noise tool was already available for a

couple of month through GitHub and some

of you may have been used it already but

we've managed now to make it available

in the cper GUI so let's head over to

the computer and have a look at the new

version of Greer with Greer AI D this is

a new user interface of expert 3.0 maybe

you already have 3.0.1 up here because I

recorded the video a few days ago you

can see that everything has become a

little clearer the sliders to stretch

the images and adjust the color

saturation have now moved down here also

new is this channel link button you may

know this function from pix inside I go

through it step by step and unfold the

first function here you can load the

image and I'll do that

once as soon as the image appears you

can adjust the stretch factor and the

color saturation and you can see that

there are still edges that I should crop

away first by the way this is extremely

important for gradient

removal so I closed the first function

and open the crop function I adjust the

image field slightly and Trigger the

crop

function that's done and I can open the

next window and continue with the

gradient removal you can also select the

AI functionality here but if you want to

choose a different method you can of

course do that too and then the other

options will pop up but I'll just use

the AI here and here's the new feature D

noising there's not much to adjust with

these AI things that's just in the

nature of these algorithms you can only

set the strength of the denosing which

is set to open .5 by

default if you are using GRE bird for

the first time you may see a message

asking you to download the appropriate

AI please follow the instructions and

confirm with yes you may also receive a

message asking you to select an AI

version please open the advanced

settings on the right hand side and

select the latest version in the

corresponding fields which I did already

here now I'll click on the noise button

and depending on your computer this may

take a

while the D noising has been completed

and we can now click through the

individual images above here starting

from the original image to the gradient

free image and finally we also look at

the Denis image it may be that the

brightness of the image changes slightly

because the background has been denoised

and the image display in graer is

calculated ated on the basis of the

background level this can make it

difficult to assess the quality of the D

noising at least that's how I feel I can

judge this better in stretched images

but that says nothing about the quality

of the D noising you can also use D

noising on stretched images but the AI

was trained on linear images and should

therefore deliver the best results here

not much else has changed in the last

step you can now save the result and

that's

it and download GRE directly from our

homepage at

www.grar.com there are download buttons

for the different operating systems make

sure that if you have a Mac computer

there are those with Intel processors

and those with the so-called silicon

chips silicon chips uh means M1 M2 Etc

GRE is also available for Linux you can

also download Greer directly from GitHub

where you can also find the release

notes for example there is information

about the minimum requirements for Mac

OS and there is a note for Ubunto users

under Windows when you start the

installation a warning message will

appear something like the computer is

protected by windows then please click

on the further information field and

then open anyway then you can continue

with the installation and what has been

new for a while is that there is now a

proper installer on the Mac you will

also get a message something like this

Greer can not be opened because Mac OS

cannot scan for mare or something like

this then please click on the Apple at

the top left of your Mac and go to the

control panel scroll down a little under

data protection and security and allow

Greer to be

opened you can still use Greer AI

functions under pigs inside there is a

script and there's now also a process

for Greer however I would like to

emphasize that these are other

developers unfortunately the Greer team

cannot provide support for this if you

have any problems please contact the

script developers

directly you can find links in the video

description you can find the greit

script in the toolbox it also contains

other great Scripts which I showed you

in an earlier video Let's install the

process together to do this go to

resources manage

repositories here you enter the

following link although it is important

that you put the backlash at the end do

not open this link in an internet

browser it is just a repository link you

confirm twice with okay and now carry

out the actual app date therefore you

have to shut down pix inside and reopen

it and then the uh

grep process will be

installed in the processes menu you will

actually find a Greer

process I once loaded an image and

Trigger the screen transfer function I

will now first start the gradient

removal and I will also have the

background model

output after a short time we get the

result and the background is also shown

to us which we can make it appear a

little bit smoother in the next step I

carry out the den noising here with the

maximum

value that took a while but we compare

the result with the original image and

see that there was a lot of the noising

again under the video in the video

description you will find all the

necessary

links a few more words about training

the D noise AI in order to ensure D

noising that looks as natural as

possible we trained the AI according to

the following idea we took images that

were exposed for a very long time and

then only took a subset a small portion

of these exposures and stacked them as

well well so we have two stacks one with

the full exposure time and one will say

about A fifth or so of the exposure time

if you look at these images you see

exactly the same image twice but of

course there's more noise in the shter

St we trained the AI with exactly these

pairs of images and so we managed to

create an AI that shows d noising as

naturally as

possible and one more thing all the

develop velers of greir and also of

these scripts are working in their free

time on the weekends to provide you some

wonderful software which cost you

nothing so if you encounter any problems

with G grur or the Pix inside scits

please uh bear with us and be patient

and we try to help you as best as we can

but sometimes yeah it's a little bit

difficult to um to find a way to solve

the the problems and it might take some

time to solve some problems so please be

patient with us I hope you enjoyed the

video so have fun with the latest

version of Greer and Greer do noise and

see you next time bye-bye