►
Description
GHS Stretching 101 – Part B – Applications
David Payne
Following up on the understanding of brightness, contrast and control of the Generalized Hyperbolic Stretch provided in Part A of this talk, Dave will be providing examples of the application of GHS to a variety of image types to customize them to what you want to see and to your own taste. He will illustrate how GHS can accomplish an initial stretch from linear to non-linear, used to make finer adjustment to an already non-linear image, and how GHS can be incorporated throughout your workflow to work in concert with other image enhancing processes. Applications will range from terrestrial images to solar system to a variety of deep sky targets. Examples will be drawn from monochrome to broadband colour to narrowband data and include starry, starless, and star images. Processes demonstrated will include colour calibration, HDMRT dynamic range regain, and control of star halos. Finally, Dave will be illustrating how the components of an image: stars and starless, RGB and narrowband, luminance and colour information can all be brought together with the help of GHS to create your final image.
A
B
Hi
everyone
I'm
back
to
talk,
give
the
second
part,
at
least
of
the
generalized
hyperbolic,
stretching
script
and
module
talk
again.
This
is
a
strip,
a
script
in
a
process
I'm
going
to
be
focused
on
the
process
tonight
for
stretching
your
astral
images.
This
was
put
together
by
Mike
Cranfield
and
myself
and
we're
just
too
astronomers
amateur
astronomers,
with
some
synergistic
backgrounds
and
without
further
Ado
I'm,
going
to
kick
off
the
top,
but
before
before,
I
go
on.
B
So
I'm
going
to
open
up
ghs,
so
the
first
talk
was
about
understanding
the
stretch
and
in
particular,
we
want
to
talk
just
go
over
these
parameters
because
they're
important
to
know
as
we're
conducting
stretch.
So
you
can
follow
what
I'm
doing
remember.
The
first
of
our
five
key
parameters
was
the
amount
of
stretch
we
called
Big
D
for
Daddy,
and
you
can
tell
how
much
it's
stretching
an
image
and
I'm
still
now
focusing
on
this
grayscale
image.
B
Here
by
how
far
our
stretched
curve
varies
from
a
45
degree
line
on
our
histogram
plot
and
the
distance
that
it
gets
is
away
from
that
line
is
how
much
it's
brightening
if
it's
above
the
line
or
dimming.
If
it's
blowing
below
the
line
and,
as
we
add
stretch,
as
we
add
stretch,
as
we
add
D
to
our
stretch,
the
stretch
gets
further
and
further
away
and
that
brightening
causes
the
histogram
to
move
to
the
right.
B
That
we're
redistributing
around
the
image
so
where
the,
where
the
stretch
is
steepest,
we're
actually
adding
contrast,
because
it's
steeper
than
I
our
identity
line
and
where
it's
shallower
and
less
steep
we're
actually
taking
contrast
away.
That's
because
there's
only
so
much
contrast
in
an
image
to
go
around.
So
if
we
add
it
to
some
place,
we
have
to
take
it
from
someplace
else.
B
Where
we
are
adding
stretches
contrast,
it
tends
to
push
the
histogram
down
and
where
it
is
being
taken
away.
The
histogram
goes
up.
The
second
parameter
is
actually
a
third
parameter
listed
here,
but
the
second
one
is,
with
ghs
we're
able
to
control
exactly
where
we're
adding
that
most
contrast
at
zero.
We
can
see
we're
adding
the
most
contrast
down
here
at
zero,
but
we
can
put
that
maximum
contrast
at
wherever
we
want.
B
B
B
The
next
parameter
I'm
going
to
talk
about
the
last
of
the
five.
Is
this
B
baby
for
little
B
for
babies
that
I'm
going
to
be
talking
about?
It
controls
the
shape
of
the
stretch?
It
starts
off
at
zero,
which
is
the
broadest
stretch
that
can
take
place,
but
as
I
move
away
from
zero.
First,
on
the
positive
side,
you
can
see
it
intensifies
that
contrast
ad
right
at
SP
and
decreases
that
farther
away
from
SP
at
the
end,
so
I
can
still
move
SP
wherever
I
want
and
I
can
change.
The
shape.
B
I
can
also
go
away
from
zero
on
the
negative
side,
and
this
is
for
very
fine
adjustments,
but
again
the
further
I
get
and
now
in
a
negative
sense
away
from
zero
B,
actually
fights
D,
but
it.
But
it
becomes
a
very
much
more
gentle
stretch,
all
in
all
so
I'm
going
to
be
using
these
parameters
to
both
perform
initial
stretches
on
images,
as
well
as
as
well
as
some
refinement
stretches
as
we
go
along.
B
If
you
recall
from
our
first
video
oh
I,
closed
the
ghs
accidentally
recall
from
our
first
video,
we
did
a
stretch
on
the
heart
and
soul.
Nebula
and
I
do
want
to
reset
the
stretch
here,
because
I
want
to
talk
a
little
bit
about
the
histogram
and
what
it
should
look
like,
and
this
character
is
going
to
come
out
through
the
rest
of
the
presentation.
Today,
I
tend
to
look
at
the
log
scale
and
you
see
we've
got
an
image.
This
is
showing
the
histogram
of
this
image
on
the
left.
B
It's
a
it's
a
it's
a,
not
a
bad
stretch.
It's
not
a
bad
image
and
I
just
wanted
to
bring
it
up.
So
I
could
show
you
what
on
looking
at
a
log
scale,
the
histogram
tends
to
look
like
for
successful
stretches
and
successful
images.
Basically,
the
components
are
first
arise
from
the
left
to
a
histogram
Peak.
B
B
It
can
have
some
slight
curves
to
it,
but
it
won't
have
bumps
and
valleys
as
you
go
along
and
then
end
up
on
the
right
hand,
side
on
the
right
side,
either
switching
to
a
steeper
slope
or
actually
coming
to
a
cliff,
and
what
you
want
to
make
sure
you
avoid
in
this
case
is
too
much
curve
on
the
right
hand,
side
curving
up
setting
that
star.
Brightness
is
bunching
up
on
the
right
hand,
side
of
my
histogram,
making
the
Stars
overly
bright
without
any
contrast
and
and
bloated.
Potentially.
B
B
So
one
last
thing
before
I
start,
we
start
stretching
some
images.
I
want
to
just
go
over
those
grand
rules
and
all
I
want
to
say
about
the
ground
ground
rules.
Here
is
that
because
of
the
formulation
and
the
implementation,
your
data
will
retain
intact.
When
you
go
through
ghs,
it's
one
of
the
few
processes
that
can
actually
say
that
so
you've
always
got
a
way
to
undo.
B
You
can
either
either
go
back
to
a
previous
image,
but
you
can
always
undo
there's
an
invert
button
that
you
can
click
and
undo
and
I'm
actually
going
to
show
you
a
few
tricks
with
that
invert
button,
because
it
really
stands
ghs
on
its
head
and
allows
for
a
whole
bunch
of
extra
things,
but
because
of
these
formulations,
you
can
do
stretches
all
day
long.
Your
data
won't
get
tired
of
being
stretched,
so
you
cannot
do
a
whole
bunch
of
little
refinements
and
it
won't
get
tired.
B
At
the
same
time,
it
is
a
bit
of
a
limitation
because
of
its
formulation.
The
rank
order
of
pixel
brightness
will
remain
the
same
in
your
image,
no
matter
how
you
basically,
no
matter
how
you
stretch
it.
Other
processes
will
be
changing
that
rank
order
and
sometimes
that's
necessary
in
order
to
get
the
image
that
you
want
and
essentially
changing
the
rank
over.
The
pixels
is
what
all
the
other
processes
are
for.
B
So
the
first
image
I'm
going
to
stretch
here
is,
was
really
a
drive-by
Imaging,
not
much
thought
gone
into
it
of
the
Moon
monochrome
image
and
I
open
it
up,
and
you
can
see
even
in
linear
form,
you
can
see
the
outline
of
the
Moon
barely
and
here's
the
here's,
the
histogram
that
goes
with
it
and
I'm
going
to
go
through
the
same
process
for
this
image
as
we
went
through
for
that
heart
and
soul.
Nebula
that
I
showed
you
in
the
last
talk.
B
It's
a
little
bit
different
image,
but
I
can
see
where
my
histogram
Peak
is.
It's
still
on
the
far
left
hand
side,
and
the
first
thing
I
want
to
do-
is
do
a
black
point
adjustment
because
I'm
looking
at
the
log
scale,
there's
my
non-log
scale
of
the
histogram
I
can
see.
There
are
no
data
points
here
and
I
can
achieve
a
Black
Point
by
switching
over
to
the
linear
type
stretch
clicking
on
where
I
want
the
black
point
to
be
said.
B
B
Now
this
is
my
standard
procedure
for
initializing,
then
I
start
to
add
D,
and
you
can
see
my
histogram
move
to
the
right
and
what
I
can
do
is
just
keep
moving
it
until
I'm,
getting
something
on
my
histogram.
So
here's
a
preview
what
I'm
doing
so
I'm
going
to
take
that
d
back
and
as
I
add
D.
You
can
see
my
moon
come
out
of
the
Shadows.
B
Now
and
I
can
see
I'm
doing
something
that
I
need
to
watch
for
as
I'm
stretching
images.
If
you
recall
I'm,
is
that
I'm
bifurcating,
the
histogram
I'm
actually
splitting
it
in
two
and
actually
what
you'll
find
is
this
part
of
the
histogram
is
actually
my
moon
and
this
part
of
the
histogram
is
actually
everything
else.
Pretty
much
I
can
move
SP
back
and
forth.
B
Right
but
then
you
say
well:
I
haven't
I,
haven't
stretched
my
image
quite
enough
and
to
to
my
liking
and
and
that's
true
and
just
sometimes
with
certain
images
when
you're
too
far
in
front
of
the
histogram
Peak
and
that's
the
place
you
want
to
be.
You
will
bifurcate
the
image
and
there's
nothing
wrong
with
executing
the
stretch
at
that
point.
B
B
You
can
add
some
B
and
now
you
can
see
I'm
back
to
stretching
the
image
without
bifurcating
I
am
starting
to
bifurcate
and
you'll
notice
that
even
if
I'm,
not
bifurcating,
you
see
this
image
is
going
down
and
it's
going
back
up.
Well,
that's
not
a
great
thing
to
be
doing
to
my
image.
You
remember
if
you,
if
you
recall
when
I
was
describing
that
really
good
histogram
it
goes
over.
The
peak
goes
to
a
line,
comes
down
the
peak
and
then
moves
to
a
sort
of
straight
line.
That's
decreasing!
B
B
B
So
if
I
move
SP
back,
get
rid
of
that
I've
got
it
all
the
way
at
zero.
That's
where
it
belongs,
then
get
rid
of
this
and
I'm
not
doing
too
bad.
Here
I've
got
a
little
bit
of
a
hump
here
and
a
little
bit
of
a
secondary
hump
there.
So
I'm
gonna
take
that
stretch
and
it's
going
to
re-stretch
for
me
once
again.
B
C
B
B
So
at
this
stage,
I've
got
a
fairly
decent
stretch
could
be
a
little
brighter,
but
I
don't
want
to
get
too
bright
or
I'm
going
to
lose
some
of
the
some
of
the
details
on
the
edge
of
the
Moon
here.
But
you
can
see
from
this
image
that
I've
got
a
bit
of
a
hump
in
the
middle
and
whenever
I
see
a
hump
in
my
histogram
that
tends
to
indicate
I've
got
a
little
bit
of
bunching
up
of
pixel
brightness
I.
Don't
have
enough
contrast
in
part
of
my
image.
You
can
see.
B
A
lot
of
this
Moon
area
is
basically
the
same
brightness
and
that's
what's
causing
this
hump.
There.
I
can
check
that
out
by
just
clicking
in
this
space
somewhere,
and
you
can
see
that
it's
kind
of
Landing
oops,
that's
kind
of
Landing
in
this
hump
in
the
middle,
so
I
can
fix
that
hump
in
the
middle,
by
kind
of
putting
it
in
the
middle
here.
Making
that
my
SP
point
I
just
want
to
go
gently
on
my
b
here
and
you
can
see.
B
I
can
stretch
that
out
at
that
point
and
make
it
a
little
bit
flatter
and
you
can
see.
I
am
putting
a
little
bit
of
contrast
into
my
image.
So
when
it's
pretty
good
and
flat
I
can
say:
okay,
I'll
accept
that
and
you'll
stretch,
my
image
and
reset.
You
can
see
I've
taken
that
bump
out
of
the
middle
and
added
a
little
bit
of
contrast
in
the
move,
see
on
the
right
hand,
side
there's
a
little
bit
of
a
bump
there.
Let's
get
rid
of
that.
B
I
can
send
that
to
SP
and
this
time,
because
the
bump
isn't
very
broad
I'm,
going
to
start
using
a
negative
B
somewhere
in
there
say
and
then
start
adding
some
contrast,
and
you
can
see.
I
am
kind
of
flattening
it
out,
but
I'm
also
pushing
it
to
the
left,
hand,
side
and
I
think
I
need
to
adjust
SP
a
little
bit
to
the
right
here
and
add
a
little
bit
more
contrast
until
I
spread
that
Peak
out
I
don't
want
to
create
a
a
bulge
but
I'm
getting
rid
of
that
secondary
Peak.
B
B
B
You
can
see
I
almost
bypassed
the
the
Peak
at
all.
You
can
only
see
it
when
I
zoom
in
you
can't
really
see
the
peak
on
this
left
hand
side.
Well
this!
That's
all
this
black
area
in
the
Blackness,
so
I
can
still
make
out
my
Terminator
I
haven't
darkened
it
too
much,
but
but
that's
about
the
limit
of
what
I
can
do
with
ghs
on
this
image.
B
I
can't
add
contrast
in
here,
because
I'm
going
to
be
darkening
parts
of
it
that
will
darken
another
part
and
I
can't
brighten
some
other
parts,
because
it'll
brighten
other
pixels,
so
what
I
have
to
do
is
I
have
to
go
to
some
sort
of
other
methodology
here
and
this
one
I'm
just
going
to
show
you
local
histogram
Equalization,
so
by
making
this
fairly
flat,
I'm,
actually
setting
up
local
histogram
Equalization.
To
add
contrast.
B
Now
the
parameters
are
in
here.
How
far
away
does
it
have
to
to
look
really?
That's
like
that
kernel
size,
but
it's
going
to
trick
by
adding
some
contrast
locally
and,
for
example,
if
we
look
at
our
preview
here
of
of
the
Moon,
you
can
see
that
there's
some
parts
that
are
showing
quite
a
bit
darker.
B
Then,
when
just
stretched
with
ghs
and
parts
that
are
brighter
and
what
it
is,
is
it's
violating
that
that
rank
order,
pixels
rule?
So
it's
doing
a
mind
trick
on
you
now
you
can
obviously
add
too
much
contrast
and
you're
going
to
say.
Oh,
that's,
that's
doing
something
funny
it's
playing
a
trick,
so
your
brain
won't
accept
it.
B
B
B
B
B
I,
do
put
the
add
the
hydrogen
Alpha
back
into
the
color
image
for
my
RB
RGB
image,
as
you
see
here
and
I
also
had
a
luminance
image
which
I
also
added
some
hydrogen
Alpha
and
some
O3
data
into
into
this
into
my
luminous
image,
I'm,
showing
you
a
noise
reduced
version,
that's
Auto
stretched
here
here
is
the
the
non-noised
reduced
version.
I
did,
and
sometimes
I
do
this
sometimes
to
do
with
stf.
B
D
B
If
I
put
SDF
back
on
here,
you
can
see.
I
have
zoomed
in
on
the
the
histogram
here
and
I
do
want
to
set
the
Black
Point
and
I
can
tell
well
there's
a
couple
of
a
couple
of
few
pixels
back
here
near
zero,
but
you
can
see
I'm
pretty
good
to
to
set
my
black
point
here,
but
I'm
not
going
to
use
the
linear
image.
B
What
that's
doing
is
making
the
stretch
to
the
right
of
SP,
completely
linear,
and
then,
when
I
add
stretch
to
this,
it's
actually
going
to
shrink
the
right
hand,
side
and
I
can
move
the
SP
I
can
move
the
Black
Point
to
wherever
I
want.
It's
actually
not
creating
a
black
point.
It's
just
compressing
all
these
pixels,
so
the
information
isn't
totally
lost.
Although
if
you
approach
machine
language
machine
resolution,
you
can
lose
some
data,
but
let's
just
execute
that
on
our
image.
B
B
B
B
So
I'm
going
to
back
off
on
my
amount
of
stretch,
maybe
I
should
back
off
a
little
bit
on
my
b
Factor
find
out
what
kind
of
gives
me
this
more
or
less
straight
line
to
the
right
of
my
stretch,
and
you
can
see
I'm
pretty
good
here,
I'm,
pretty
close
to
where
I
want
to
be
I
can
do
another
little
assessment
of
of
SP.
Here
you
see
if
I
got
SP
too
low.
I
start
brightening.
Everything
in
the
background
becomes
Gray.
B
D
B
B
B
B
B
B
D
B
And
now
it's
going
to
use
wavelets
to
basically
form
another
trick.
It's
going
to
boost
the
Boost,
the
amplitude
of
some
of
the
wavelets
decrease
some
of
the
larger
wavelet
amplitudes.
So
I
now
can
bring
some
of
the
details
of
the
the
Galaxy
into
my
image.
Now
what
this
will
do
is
actually
playing
a
trick.
It
will
darken
some
pixels
like
by
rights.
The
this
little
dust
Lane
here
should
is,
is
being
made
pretty
much
as
dark
as
out
here
and
that
and
that's
not
really
the
case,
but
it's
it's
doing
this
trick.
B
So
it's
adding
contrast.
You
start
to
see
the
structure
of
the
Galaxy
by
doing
that,
but
it
does
have
some
unintended
consequences
if
I
looked
over
at
my,
if
you,
let
me
get
rid
of
the
stretch,
that's
on
here
here
was.
My
here
was
my
starting
image.
Actually,
it's
made
the
in
this
case.
It's
made
the
image
a
little
bit
better.
In
some
cases,
it's
made
it
actually
straighter.
There
is
a
little
bit
of
a
hump
here,
there's
a
little
bit
of
a
hump
here.
B
B
B
This
one
also
has
a
black
point:
that's
that's
fairly
high,
so
I
do
want
to
click
where
I
want
the
black
point
to
go.
I'm,
not
setting
the
black
point
I'm,
sending
that
to
SP
my
SP
parameter.
I
am
sending
HP
down
to
SP,
so
I'm
linearizing
the
whole
right
hand,
side
I'm,
adding
some
stretch
and
I
can
move
which
I
am,
on
the
left
hand,
side.
E
B
And
then
apply
that
again,
it's
affected
by
stf
now
I
want
to
do
my
real
stretch.
Send
my
SP
I,
put
it
near
my
histogram
Peak
now
I
can
zoom
out
use
a
large
B
start
increasing
my
stretch
as
I
move.
The
stretch
to
the
right
I
always
forget
to
look
at
what
the
images
looks
like
take
my
stf
back
and
you
say:
oh,
what
wait
a
second
here,
I'm
sort
of
missing
my
Shark.
It's
not
it's
not
coming
out.
B
B
You
can
live
with
it
if
it's
trying
to
do
some
sort
of
effect
like
it's
really
making
the
shark
stand
out
from
the
background,
but
it's
kind
of
unrealistic
because
there's
other
nebulosity
in
the
background
I,
don't
like
it
generally
when
you're
bifurcating
the
image
like
this,
you
want
to
back
off,
probably
on
both
B
and
D
right.
So
I
want
to
go
up
with
a
more
modest
B
I'm,
really
not.
B
D
B
So
let's
reset
the
stretch
now
I
want
to
look
at
my
image
and
say:
okay.
Well,
where
is
this
shark
on
the
histogram?
So
I
can
click
on
a
preview
and
I
can
click
on
the
shark
and
you
can
see
this
shark
exists
actually
in
my
histogram
Peak,
it
doesn't
exist
out
here,
like
the
Galaxy
did
in
the
previous
example.
B
It's
actually
Within
this
histogram
Peak
and
if
I
want
to
brighten
this
brighten,
the
shark
I
know
I
want
to
be
to
the
left
of
where
the
shark
brightness
is,
but
I
want
to
be
to
the
right
of
the
background.
So
if
I
click
on
some
of
the
background,
I
can
see
that's
near
the
histogram
Peak.
So
that
gives
me
an
indication
where
I
want
to
add
contrast
in
between
those
two
points.
So
I'm
going
to
put
sp
there.
B
Now,
just
some
modest
B
I'm
going
to
start
with,
because
I'm
starting
from
a
non-linear
image
and
add
a
little
stretch
to
here
and
you
can
see
now,
the
shark
is
coming
out
from
the
background
and
I'm
not
being
too
bright
on
the
stars.
But
if
I
increase
B
too
much
I
start
to
increase
that
start
to
create
that
Valley
again,
which
I
really
don't
want.
B
B
Now,
what
I
did
find
is
a
funny
characteristic,
where
it's
trying
to
fool
my
eyes,
but
what
it
actually
ends
up
doing
with
my
image,
because
it's
trying
to
increase
the
contrast
where
there's
some
contrast
and
internal
is
making
the
shark
a
bit
blotchy,
but
it's
also
creating
a
highlight
around.
It's
almost
like
someone
went
with
a
white
marker
and
and
marked
around
the
edge
of
the
the
dark
shark,
and
that's
not
going
to
fool
anybody.
B
A
B
Could
I
set
you
up
for
this?
Okay,
good,
okay,
so
normally
I
do
exactly
what
you're
saying
I
separate
the
stars
out
from
the
subject
matter
this
star
and
a
couple
of
the
stars
in
this
image
are
a
problem
because
at
least
starnet
I
haven't
I,
don't
have
a
star.
Exterminator
wreck
doesn't
recognize
that
Halo
as
a
star.
It
thinks
there's
a
star
on
top
of
this
circular
nebulosity.
B
That's
there
Okay!
So
yes,
you're
right
and
I'm
going
to
show
you
that,
as
as
we
go
through,
let's
see
how
far
we
get
in
this
recording
I'm
going
to
show
you
that,
but
but
that's
just
not
an
option
for
this
image,
so
I
trying
to
show
what
you
can
do.
Hdmrt
will
minimize
it,
but
I
do
want
to
make.
Another
point
is
that
star
is
in
that
nebulosity
and
it
is
creating
a
Halo?
It
is
in
my
data
ghs.
Isn't
making
up
that
Halo
it
is.
There
remember
I,
said
ghs
won't
do
anything.
B
It
just
doesn't
know
that
it's
a
star
or
not
a
star
or
a
Halo
or
not
a
Halo.
It's
not
that
smart,
so
that
star
is
got.
Does
have
that
bright
handle
around
it
now
it
may
look
better
on
a
poster
if
you
get
rid
of
that
Halo
and
it
may
be
more
artistically
pleasing
if
you
get
rid
of
that
Halo,
it
depends
on
what
your
purpose
of
your
image
is.
B
B
B
B
B
B
So
I
can
stretch
a
little
bit
too
much
move
SP
to
see
where
am
I
bringing
the
background
out
I,
don't
really
want
that
background.
I
want
it
to
fade,
but
I
don't
want
to
fade
the
dimmest
star,
so
I
remove
HP
up
until
I
get
a
nice
image.
The
other
thing
I
want
to
be
careful
of
is
now
I'm,
creating
a
hump,
so
I
don't
want
to
do
that
so
I'm,
going
to
back
off
that
stretch
right.
B
Now
the
last
thing
I
want
to
check
is
the
core
of
this
cluster,
because
I
want
to
make
sure
my
stars
are
resolved
right
to
the
core
of
that
structure
and
that
really
has
to
do
with
the
B
parameter
I'm
using.
If
B
is
that's
a
bit
too
much
B,
you
can
see
that
I'm
I'm
losing
some
of
the
contrast
in
the
cluster
itself.
That's
pretty
good,
so
I
can
say:
let's
execute
that.
B
So
the
problem
with
this
image
is
that
I
don't
have
a
lot
of
color
in
this,
and
that
tends
to
happen
as
your
as
you're
increasing
brightness.
The
saturation
is
actually
saying
the
same.
Saturation
is
really
the
difference
between
the
color
intensities
at
a
given
pixel.
It's
really
the
difference
and
the
brightness
is
really
the
average.
So
as
I'm
moving
the
average
up,
the
ratio
of
the
difference
to
the
average
decreases
and
it
tends
to
bleach
out
images
somewhat
and
there's
two
solutions
to
this.
D
D
B
B
B
D
B
The
other
way,
I
can
add
color
to
my
stretched
images,
or
maybe
that's
too
much
color,
oh
because
it's
repeating
the
stretch
well
I
do
want
to
show
you
that
little
Galaxy
because
it
came
out
kind
of
nice
in
here
you
can
see
some
colors
in
that
that's
not
its
color
Star
by
the
way
it
just
it's
just
a
foreground
star.
That
makes
it
look
like
it,
but
pretty
good,
pretty
good
result.
If
I
go
to
my.
B
I
can
take
out.
I
can
pick
this
Hercules
image
that
I
have
there
it's
this
image
and
it's
listing
the
stretches
that
I've
done.
I
know
this
was
the
second
stretch,
the
third,
the
first
one
was
my
my
Black
Point.
The
second
one
was
my
stretch
and
the
third
one
was
my
saturation
stretch
I'm
after
this
stretch,
I
can
pull
that
out
of
History
if
I
double
click
on
that
that's
the
original
Stretch
I
did
on
this
image.
B
Does
and
here's
what
it
looks
like
so
what
it
does
is
not
an
RGB
stretch.
What
it
does
is
a
brightness
stretch,
but
it
doesn't
apply
that
brightness
stretch
to
the
image.
It
actually
applies
that
ratio
of
the
brightness
of
the
stretched
image
to
the
brightness
of
the
starting
image.
It
applies
that
ratio
to
each
of
the
individual
colors,
and
that
means
that
each
of
the
the
saturation
increases
as
the
brightness
increases,
and
that
also
retains
the
colors
in
the
image.
Now
some
clipping
occurs
here.
You've
got
various
ways
to
look
after
that.
B
Mike
came
up
with
this
RGB
blend
method.
It's
the
best
method,
just
leave
it
there
you
can
try
clipping.
Also
was
our
initial
stab
at
this,
but
I
would
just
leave
it
there.
It
does
a
great
job,
you're,
actually
blending
in
the
RGB
image,
so
that
no
clipping
is
actually
taking
place.
So
it
was
minor,
minimal
effect
on
your
image.
By
using
this,
hey,
Dave
I
got
a
question,
reply
that
and
that's
two
ways
of
bringing
out
the
color.
A
Dave
can
I
ask
a
question
on
that
image.
Sure
I
I
see
this
when
I've
used
ghs
and
I
see
it
on
your
image.
Now,
too,
you
have
a
number
of
stars
that
sort
of
look
like
billiard
balls
right
where
they're
colorful
and
have
like
a
white
dot
in
the
middle
yeah,
and
what
is
why
is
that
Is,
It
Anyway
I
haven't
how
you
deal
with
that.
So.
B
B
Just
for
now
what
it
is
is
the
saturation
cannot
be
accommodated
like
if
you
think
of
saturation
as
the
difference
between
the
colors.
Let's
say
we
got
a
two
color
thing:
we
got
red
and
blue
and
let's
say
the
saturation
there's
0.1
brightness
saturation
difference
between
red
and
blue
and
that's
giving
you
the
color
of
the
Stars.
B
B
B
So
that's
the
only
way
you
can
get
the
0.8.
So
as
you
brighten
things
past
the
the
level
that
they
can
accommodate
that
saturation,
it
whitens
out
the
center
of
the
star
closer
where
the
star
is
less
bright.
It
can
accommodate
that
saturation
and
that's
why
you
see
the
color
sort
of
appear
in
the
circles
around
the
Stars.
B
F
I
ask
why
a
one
little
question
on
that
I'm
gonna
have
to
go
in
a
second
but
I
I've
scratched
my
head
on
that
a
lot
I
think.
That's
that
that
issue
also
has
something
to
do
with
acquisition
that
if
you
have
acquired
too
much
brightness
in
the
RGB
channels,
you
tend
to
completely
saturate
the
center
of
the
star.
Is
that
also
a
factor
here
or
is
that
just
is
this
mostly
a
processing
problem.
C
F
E
B
When
I,
if
I'm
taking
a
cluster
like
that,
I
just
make
sure
I
don't
overexpose
my
image,
particularly
in
the
sense
of
that
cluster,
like
you
want
to
see
the
stars
right
at
the
center
of
that
cluster.
So
you
you
tend
to
dial
back
your
exposure
to
make
sure
you
don't
do
that.
But
when
you
stretch
back
up
you're
trying
to
use
your
entire
Spectrum,
so
the
solution
to
that
and
I'll
I'll
jump
to
it,
but
essentially
I'll
jump
to
that.
And
then
we
can
finish
off
the
evening
and
maybe
we
can.
B
So
if
I
go
to
ghs
on
here
and
the
the
normal
way,
you
sort
of
stretch,
a
an
image
here
is,
you
would
get
rid
of
the
stf
and
you
would
say,
oh
I
want
to
be
protective
of
the
stars
and
I
would
stretch
the
image
normally
and
you
get
your
stars,
and
maybe
you
say
you
want
to
keep
some
contrast
in
the
Stars.
You
don't
want
to
over
stretch
them.
If
you
don't
want
these
big
Halos,
but
you
want
some
contrast
in
them.
B
B
The
point
is:
what
I
want
to
do
here
is
actually
put
SP
all
the
way
to
the
top
so
I'm
going
to
try
and
make
stars
with
the
least
amount
of
contrast.
I
can
so
I
can
do
that
by
stretching
in
invert
mode
with
SP
all
the
way
right
hand,
side
I
can
stretch
the
Stars.
You
can
see
I'm
bifurcating
them
there
I'm
just
stretching
the
brightest
stars.
What
I
have
to
do
is
reduce
my
B
Factor.
B
You
can
see
how
that's
changing
my
my
stretch.
If
I
reduce
my
B
Factor
I
can
create
basically
contrastless
star
Mass,
so
I've
taken
a
heck
of
a
lot
of
the
contrast
out
of
these
Stars.
You
can
still
see
the
spikes
and
you
can
still
see
some
dim
pillows
there.
So
maybe
I
back
off
that
so
much.
But
essentially
you
can
see
these.
These
stars
are
basically
round
and
flat
if
I
increase
the
stretch,
I
increase
their
size.
B
B
The
dim
Stars
they're
kind
of
round
very
little
contrast
in
them
they're
only
0.75
right,
so
this
is
going
on
a
black
background.
It's
they're,
not
the
brightest
things
going,
but
the
nice
thing
about
keeping
this
below
a
hundred
percent
below
completely
completely
bright
is
now
they
can
accommodate
color
saturation.
B
B
B
B
D
F
Yes,
it
does
it's
very
it's
very
interesting.
There's
I
understand
that
I
understand
exactly
what
you're
doing.
That's
very
sophisticated
I
have
to
put
that
in
as
a
method
47.3
for
the
for
dealing
with
this
problem.
It's
really
a
a
mind.
Bender
and
I
very
I
find
it's
very
interesting.
Now
your
lecture
I
have
to
go,
unfortunately,
I'm
going
to
get
divorced,
but
I.
B
A
A
Right
so
that
that
that
did
answer
my
question.
So
your
question,
you
know
my
question
was:
what
about
those
donuts
I
keep
getting?
That's
a
good
word
for
him
and
your
answer
was
sort
of
you
can
flat.
You
know
if
you've
separated
out
the
Stars,
you
can
flatten
them
in
in
that
way
and
get
and
make
room
for
the
saturation
and
then
the
root
cause
of
the
donuts
is
sort
of
the
inability
to
saturate
further
you
know:
I
I,
didn't
mind
the
white
in
the
center.
What
bothered
me
was
sort
of
the
discontinuity.
A
You
know
what
I
mean
the
fact
that
it
was
you
know
colored
and
then
white,
and
then
there
was
no
like
sort
of
ramp
to
white
I
mean
it's
pretty
discontinuous
the
way
it
changes
colors,
but
I.
Guess
that
happens
where
you
run
out
of
of
room
to
for
your
saturation.
All
of
a
sudden,
you
can't
do
it
anymore,
yeah.
B
G
C
B
B
B
Gotta
be
protective
of
the
Stars.
You
can
see
that
I'm
starting
to
go
up
on
the
right
hand
side.
So,
even
though
I've
got
B,
all
the
far
to
right,
I
still
need
to
add
a
little
bit
of
HP
to
be
packed
up
to
the
the
Stars
I
can
zoom
into
the
globular
cluster.
In
my
preview
now
you
can
see.
I
actually
did
a
deconvolution
on
this
and
I
wasn't
able
to
take
all
of
the
stars
out.
B
It
didn't
recognize
the
center
of
this
globular
cluster
as
just
a
bunch
of
stars,
so
it
left
them
in
my
nebulosity
image
and
I
ended
up
having
a
little
bit
of
a
little
bit
of
ringing
going
on,
but
ignore
that
it
was
a
shame,
but
I
didn't
want
to
sacrifice
the
deconvolution
I'm.
Getting
pretty
good
resolution
of
my
Stars
I
can
put
the
entire
image
and
I
can
move
SP.
B
A
B
If
I
look
at
the
preview
and
get
rid
of
the
stf,
that's
basically
duplicating
now,
we
did
talk
about
if
I
increased,
SP
I'm
going
to
go
on
to
the
front
of
that
histogram
Peak.
That's
where
my
ifn
kind
of
lives
in
this
Peak
going
on
here.
So
if
I
move
SP
in
front
of
it,
it'll
push
that
Peak
back
and
you
can
see
I'm
getting
rid
of
my
ifm.
But
I
am
starting
to
create
this
Valley
I.
B
Don't
want
to
create
valleys
in
my
histogram,
so
I
can
back
off
B
A
little,
but
I
don't
want
to
back
out
B
too
much
because
I
start
impacting
those
Stars
so
I'm
going
to
have
to
back
off
on
the
overall
stretch
now
I
could
and
when
I
did
this.
This
is
exactly
what
I
did,
but
I
had
to
return
for
a
struct
second
stretch
to
make
the
whole
thing
brighter.
B
B
B
D
C
G
B
D
B
D
B
B
B
E
B
B
B
Taking
some
contrast
away
so
I
can
make
that
curve
closer
to
to
straight
going
down
and
and
that
sort
of
repairing
what
I
did
with
hdmrt.
So
I'll
do
that
repair
to
the
image
see
it's
brought
out
some
of
the
dark
nebulosity
again
there
are
kind
of
quite
a
few
bumps
still
along
the
road
here,
but
then,
if
I
do
another
round
of
of
hdmrt.
B
B
It's
showing
me
some
of
the
details
in
here,
but
it's
also
creating
some
effects
like
like
this
here,
so
I
wanna
get
rid
of
that
concavity
that
that's
in
that
image,
so
I'm
going
to
do
an
invert
stretch
again,
send
that
to
SP
and
a
little
bit
raising
that
up
now.
I've
got
another
bulge
coming
here
that
I
need
to
get
rid
of.
B
D
B
D
B
B
D
D
D
B
B
Artificial
intelligence
deconvolution:
this
is
obeying
the
the
true
math
behind
deconvolution,
although
it
is
an
iterative
process,
so
there's
only
approximation
to
to
the
convolution,
but
this
is
just
a
fantastic
process
to
do
on.
If
you
can
make
your
image
starless,
you
don't
have
to
worry
about
any
of
those
ringing
artifacts
and
it's
just
so
easy
to
do.
You
can
use
the
starry
image
to
get
a
psf
and
I've
loaded
it
here.
B
D
B
B
B
E
B
Couldn't
go
for
any
fake
stuff,
it
had
to
be
real,
but
you
can
see
the
the
resolution
it's
bringing
out
now.
I
can
do
an
undo
and
that's
only
five
iterations
on
deconvolution.
Hopefully
you
can
see
that,
but
it
it
pays
out
huge
dividends
going
forward.
So
I
always
do
the
starless
part
at
the
at
the
linear
stage.
B
B
B
Wherever
I
want,
it
makes
the
initial
stretch
just
so
much
easier,
not
having
to
worry
about
those
Stars.
So
a
few
reasons
why
I
do
it
the
linear
stage
one
is
so
I
can
get
that
deconvolution
in
and
second
of
all,
I
can
do.
My
initial
stretch
and
I
can
really
focus
in
on
the
core
of
my
nebulosity.
Instead
of
worrying
about
all
these
extraneous
Stars,
so.
D
D
B
B
Then
I
did
local
histogram,
Equalization
I
probably
went
to
heavy-handed
on
that
one
another
ghs
stretch
and
then
finally
a
sharpening
and
so
I
have
ghs
and
you
can
see
as
I
went
through
those
how
this
histogram
is
is
being
changed
and
you
see
at
the
end
I
end
up
with
some
contrast.
I
need
to
add
here.
You
can
see
that
in
the
nebulosity
here,
I
want
to
add
some
contrast,
so
I
can
be
be
zero
alone
at
the
time.
Maybe.
D
B
When
I'm
doing
these
fixes,
you
see
I
take
the
old
histogram
and
you
can
see
the
old
histogram
has
this
bit
of
a
bulge
before
diving
down
at
the
end
and
what
I've
done
is
had
a
little
bit
of
stretch
so
that
the
histogram
comes
down
more
evenly
kind
of
crosses
over
and
extends
out
further
and
comes
down
so
really
using
the
old
histogram
and
the
new
histogram
as
an
edge
to
to
what
is
going
on.
I
need
to
make
sure
that
that
looks
good
and
brightening
up
too
much.
B
The
center,
because
I
want
to
leave
that
focus
on
that
Galaxy
Core
and
then,
when
I,
when
I've
done,
that
I
can
execute
it
and
then
I
may
be
going
on
to
the
next
process.
I
want
to
do
so.
It's
this
interlacing
of
running
the
process,
taking
what
it
does
right
and
then
fixing
any
unintended
consequences
as
a
result
from
it
and
then
moving
on
as
I
as
I
process.
B
B
And
just
a
couple
of
little
tricks
that
you
might
be
used
to
a
lot
of
people,
do
the
curves
right
at
the
end
and
they
add
a
little
gentle
S
curves,
to
take
some
of
the
contrast
out
of
the
low
end.
Some
of
the
contrast
out
of
the
high
end.
You
can
do
that
with
ghs2.
You
can
just
say
you're
doing
an
invert,
an
image
and
leave
SP
at
zero.
B
B
The
bottom
half
of
an
S
curve
sort
of
stretch
we'd
have
a
lower
part
of
the
S
and
and
late
in
a
second
I'm
going
to
put
an
upper
part
of
the
S
there.
But
what
I
recommend
is
you
look
at
the
image
and
see
what
you're
doing
to
it
as
you
decrease
that
contrast
right
at
the
bottom
end,
I
see
I'm
darkening
the
background
I'm
doing
it
too
much
here
so
I
want
to
sorry.
I
want
to
bring
my
SPN
till
it
it's
kind
of
where
I
don't
want
to
darken.
B
B
Working
on
now,
I've
got
two
new
images
open
and
then
all
I
have
to
do
is
move
SP
to
the
very
top
and
now
use
LP
to
control
what
I'm
doing
at
the
very
top
of
the
image
now
here,
I'm,
basically
looking
at
the
stars
and
the
galactic
colors
to
how
much
contrast
I
want
to
take
out
of
it
in
this
case,
I
don't
want
to
take
much
contrast
out,
but
but
there
you
have
it.
So
that's
how
you
can
do
it.
B
So
a
couple
of
really
quick
ones
to
show
you
here
are
some
just
raw
drizzle
files
that
you
may
have
after
taking
narrow
bands,
I'm
going
to
talk
a
little
bit
about
narrow
Bend
in
this
case
now,
one
of
the
first
things
you
probably
want
to
do
is
dbe
and
I'll.
Show
you
a
bit
of
a
shortcut
using
ghs
to
do
dbe,
I
always
well.
I
generally.
Do
a
channel
combination
here
and
I
just
pick.
D
B
Now
this
hasn't
been
color
calibrated
or
it
hasn't
had
anything.
It
hasn't
had
dbe
done
to
it.
So
I'm
going
to
go
in
here
and
I
can
see
where
the
colors
are
in
relation
to
another
and
I
see
that
red
happens
to
be
weaker
than
green.
Well,
that's
not
surprising
in
an
SHO
image.
What
I'll
do
is
I'll
boost
red.
D
B
D
B
B
B
B
Before
and
say,
okay
well,
I
can
check
out
the
the
model
there.
What
is
it?
What
does
it
show
me?
It
does
show
me
a
fairly
regular
sort
of
background,
which
makes
me
happy
I'm,
not
getting
any
of
my
image
in
the
background
and
then
when
I've
done
that
I
just
simply
save
it
as
a
process.
I
cancel
that
session
I
get
rid
of
this
horrible
stretch,
but
now
I
go
back
to
my
individual
Master
files,
open
it
up
and
it'll
Place.
D
B
Do
my
dbe
on
my
images,
get
rid
of
that
open
up
my
next
image.
I.
Can
my
points
are
all
saved
in
this
this
process
here?
So
I
can
just
go
through
that
and
quickly
do
that.
So
that's
how
ghfs
can
even
help
with
dbe
a
little
bit
more
sophisticated
color
is
when
I've
done
my
dbe
and
I
do
my
channel
combination
and
yeah
I
might
get
an
image
such
as
this?
This
is
another
Sho
image,
so
I
need
to
do
some
color
color
calibration
on
this
now.
B
B
I
don't
have
to
touch
anything,
but
the
amount
of
stretch
I
have
to
make
sure
I'm
doing
the
red,
Channel
and
I
want
to
stretch
this
very
slowly
with
the
D
I
want
to
stretch
it
past
where
I
want
to
go
so,
let's
say
to
the
right
of
the
blue
Channel
and
then
I
want
to
bring
HP
back
to
line
it
up
with
green.
Now,
if
I
do
this
with
the
stf,
you
want
to
make
sure
it's
locked
on
the
color.
You
can
see.
B
I
have
indeed
boosted
red
and
now
the
background's
showing
up
somewhat
as
purple
so
I
can
fine
tune
if
I
select
HP
to
fine
tune
here.
I
don't
want
the
highest
resolution,
but
I
can
fine
tune
red
to
overly
on
my
green
and
generally
I'm.
Wanting
the
left
hand
side
of
my
histogram
Peak
rise
to
show
there
now
I
can
execute
that
and
now
I
want
to
work
on
Blue
Blues
the
opposite
way.
B
D
B
So
how
do
I
get
enough
blue
and
get
enough
red
into
the
image?
Well
here,
if
I
just
let's
say
we
want
to
work
on
red
first,
if
I
put
SP
right
into
the
middle
of
my
histogram,
that's
where
I'm
going
to
add
the
most
contrast.
I
want
to
use
as
much
b
as
I,
possibly
can
because
I
want
to
inject
some
contrast
into
that
blue
into
that
red.
B
D
C
D
B
B
D
B
D
B
D
B
C
B
B
B
Color
adjusted
stretch,
stretch
red
and
blue
individually
even
further,
so
I'm
getting
even
more
red
after
I
did
my
color
and
even
more
green.
Oh
sorry,
more
blue
in
amongst
this
green.
At
this
stage,
I've
got
basically
a
couple
of
different
options:
I
can
do
I
can
do
an
lrgb
combination
by
setting
green
at
eighty
percent.
I.
Think
I
did
for
this
case,
which
is
one
way
I
can
bring
out
these
Reds
and
this
foxman's
cone.
B
B
G
B
B
B
So
this
is
all
fine
and
well-
and
you
can
do
this
really
well,
you
can
do
deconvolution,
because
it's
starless,
you
can
do
all
sorts
of
things
because
it's
starless,
but
at
the
end
of
the
day,
I
wanted
to
add,
in
my
add
in
my
RGB
stars,
because
I
wanted
the
Stars
to
be
RGB.
Now
I
can't
do
that
directly
using
that
star
mask
stretch,
adding
it
to
the
luminance
and
then
because
the
fundamental
color
of
my
RGB,
all
of
my
Luminosity,
is
a
different
color.
B
E
D
D
B
C
D
B
B
Now
the
only
thing
I
need
to
worry
about
is
that
my
stars
are
bigger
than
my
mask
or
I
could
potentially
get
some
ringing
going
on.
I
want
to
make
sure
I
absolutely
do
everything
to
capture
the
color,
so
I
don't
want
to
stretch
my
stars
too
much
I
do
want
to
do
a
color
image,
because
I
really
want
that
star.
Color.
D
B
B
B
B
Me
take
those
stars
out
for
a
second
and
let's
do
another
invert
image,
and
what
I'm
going
to
do
here?
No
I
won't
I
won't
do
the
invert
image
I'll
just
do
a
negative
B
move,
my
SP,
so
I'm
going
to
do
a
gentle
dimming
of
this
image,
so
I
want
to
make
sure
my
mask
is
not
enabled
because
I
just
want
to
dim
this
image
somewhat
now
I
can
enable
my
mask.
B
So
I
brighten
up
the
image
again,
but
I'm
also
brightening
up
the
star.
So
if
I
want
light,
really
bright,
stars
well
I'm
going
to
lose
some
color
because
I
brightened
up
the
stars,
but
now
I
get
the
overall
thing.
So
that's
how
you
can
use
ghs
to
help
put
whatever
RGB
Stars
into
whatever
image
you
want
without
losing
the
fundamental
star
color
to
it.