35 KiB
koko-aio-slang documentation
RETROARCH OUTPUT DRIVERS
koko-aio does not work by on d3d12 and glitches on d3d11.
If you absolutely need it (Xbox?), you can edit the file
config\config-static.inc
and turn the line:
// #define D3D_WORKAROUND
into:
#define D3D_WORKAROUND
Vulkan ang Glcore have no problems, you can test both to see
which performs better.
d3d10 is completely unsupported.
USEFUL LOCATIONS/FILES:
config/config-static.inc:
Some shader parameters can't be changed within retroarch,
use this file instead.
Notable one is lcd antighosting feature.
config/config-user.txt:
shader parameters that can be changed within Retroarch.
can be set within this file too.
PRO: The shader will be faster
CON: The parameters can no longer be modified within Retroarch.
textures/background_under.png
This is the image that can is shown by default under the main content and under the bezel.
Read further for details.
textures/background_over.png
This is the image that can is shown by default over the main content and under the bezel.
Read further for details.
textures/monitor_body_curved.png, textures/monitor_body_straight.png
This is the image used to draw the bezel.
Read further for details.
Texture "sources", including the main gimp project file for the default curved and straight monitor frame are on the following repo: https://github.com/kokoko3k/koko-aio-slang-misc/tree/main
Color corrections:
Modify signal color at "input" stage.
Luminance, brightness, contrast and temperature:
self explanatory.
Input signal gain:
Gain applied in the chain just before the crt emulation stages.
Gamma in:
Gamma correction applied at the beginning of the chain.
Gamma out:
Gamma correction applied to the final processed picture.
Adaptive black level range:
On old CRTs the contrast was higher on high luminosity content,
and lower on low luminosity content.
This setting modulate the range of the effect; 0.0 disables it.
It is also possible to emulate a monochrome display with custom colors:
Monochrome screen colorization:
The amount of (de) colorization applied.
Hue bright, Hue dark:
Set the hue for bright and dark colors.
Hue bright-dark bias:
Controls the distribution of dark and bright hues.
FXAA:
Apply the well known antialiasing effect by Nvidia.
Use it if you don't want to blur the image and you still don't like
jagged or too much pixelated images.
** RF Noise:**
Emulates radio frequency noise with a given strength
** Dedither:
Try to smooth dithering patterns.
Enabling dedithering automatically disables NTSC color artifacts: generation.
Sensitivity: Avoid to dedither "legit" zones by lowering this.
Basic search strength: Blends basic dedithering and original image.
Extensive search strength: Blends extensive dedithering and original image .
may produce posterization effects.
CVBS: NTSC color artifacts:
Tries to emulate typical NTSC color artifacting without emulating
full NTSC coding/decoding pipeline.
While it improves the look of NTSC content, don't expect it to be
an accurate emulation (yet?)
As today, it is enough to emulate rainbowing effects on genesis.
If you enable Glow/Blur functions or Bandwidth limited chroma,
the image will be blurred in a way or another.
You can selectively keep the part of the image which does not contain
artifacts sharp by using the followin controls.
This allow to selectively blend artifacts.
Consider artifacts above this treshold:
Tune this to select more or less artifacts, depending on their strength.
Show the mask of selected artifacts (debug)
This will show only the part of the image that contains artifacts.
Use it to for a better visual feedback of the following parameters
1* Under treshold: Cancel blur (Glow)
How much the glow/blur function will skip blurring "unartifacted" areas.
2* Under treshold: Cancel Bandwidth limited chroma
How much the Bandwidth limited chroma function will skip blurring
"unartifacted" areas.
3* Under treshold: Cancel artifacts
How much the artifacts under the treshold will be completely removed.
CVBS: Bandwidth limited chroma:
Will cause an horizontal chroma bleed which cheaply mimics the effect of
poor composite video signals.
It can be used with RGB shifting and image blurring to give the picture
an ntsc look without dealing with specific encoding/decoding stuffs.
Persistence of phosphors:
This emulates the unexcited phosphors that continue to emit light.
Early decay: is the immediate light cut after the phosphor is no more/less excited.
Late persistence: modulates the time the residual light will stay on screen
Deconvergence:
Shift R,G,B components separately to mimic channel deconvergence.
By varying Red, Green and Blue offsets, the relative component will be
shifted column by column, row by row.
Glow/Blur:
Emulate the CRT glowing "feature", so that the brighter areas of
the image will light their surroundings,
with options to switch to classic blur.
Glow to blur bias:
Higher negative values -> more glow : brighter colors expands over darker ones.
Higher positive values -> means blur: all the colors are blurred.
0.0 means no blur, no glow.
Glow spread amount:
The higher, the more the bright colors will smoothly expand.
It emulates the natural antialiasing you see on CRTs on bright areas.
More pronunced as "Glow to blur bias" approaches 0.
Sharpness (horizontal, vertical):
Modulates the sharpness of the image.
- Max (actually 7.0) will not alter the image sharpness.
- More than 0: will use gauss blur
- Less than 0: will use box blur and will progressively
add visual sharpness to image when approaching lower values.
Tate mode:
Rotates mask and scanlines by 90°
Glitch if vertical resolution changes:
Emulates the crt circuits syncing to the new signal timing.
Will shake the screen for a while when the resolution changes.
The Hiher the value, the more the amplitude.
Hi-resolution scanlines handling:
There you can choose how to handle scanlines when a content is Hi-Resolution.
Special handling may be needed to mitigate glitches/moire at 1080p or lower resolutions.
Consider Hi-Resolution above # lines:
A value of 0.0 means that it will never consider anything hi-res.
A value of 1.0 means that it will always consider content hi-res.
With values > 1.0, it will consider a frame as Hi-resolution if the lines number is above the configured value.
Hi-Res scanlines type
-1: Use a number of scanlines that perfectly fits the screen, a good glitches/moire free tradeoff.
-2: As above, but tighter (~1.5x), another good glitches/moire free tradeoff.
0: Use interlaced scanlines, may need >1080p screen to avoid moire or weavy glitches
1: Avoid drawing scanlines gaps at all.
2: Use scanlines, but don't interlace them (bad for 1080p and lower resolutions)
Scanlines flicker (0=off,1=on,2=if Hi-res):
This setting emulates the flickering issues present on crt interlaced screens
where the brighter lines flickers when they are near dark ones.
You can choose to produce the flickering: never, always or only
when the input picture is considered High resolution.
Interlace Flicker power: The strength of the effect.
Low level phosphor grid:
This is a way to produce horizontal masks, scanlines and aperturegrille/slotmasks.
Parameters are tricky to setup, but possibilities are mny more quality is good at 1080p
and hopefully great at higher resolutions.
By reading the following explanaitions, you will realize that this section can also be used to emulate
handhelds screens, where pixels you see on screen have to be sized proportionally to the game one.
Overmask (1.0 = neutral):
Values > 1.0 give a gritty/gariny look to the whole mask.
It may cause moiree if combined with curvature, dedot, or sparkling look punch.
Values < 1.0 tend to nullify the whole mask effect.
Scanlines (*4)
Scanlines emulation, set the strength of the effect here.
Double-scan low input resolutions
Activate this if you want to double the number of scanlines when the content is low-res.
"low-res is defined via "Consider Hi-Resolution above # lines" parameter above.
This option is useful if you want to emulate 30khz VGA CRT monitors.
If you are on 1080p or lower, consider to activate
Fake integer scanlines or integer scaling.
Use fake integer scanlines
Use a number of scanlines that perfectly fits the lines on the screen, not accurate,
but avoids moire and weavy artifacts.
A negative value will cause the shader to choose when it is appropriate to activate them.
The decision will be based on the ratio of output dimensions and the core.
Phosphors height Min, Max:
Try to keep scanline height between those values, depending on content brightness.
Inflation Strength:
Scanlines appear as inflated depending on the pixel brightness.
Phosphors width min->max gamma:
Since emulating phosphors with high Min-Max range changes the apparent gamma of the final image,
it is advised, if needed, to use this option to compensate, instead of the main gamma correction.
It is also a quick way to make the image brighter or darker.
Inter-line extra steepness (good for integer scaling, nice when Phosphors height max > 1.0)
When you set maximum height to anything > 0.5, the phosphor light will bleed over
the adiacent (up/down) ones so that they will start to blend togheter.
Setting this option to anything >1.0 will avoid the bleeding.
You may want them to blend or not, depending on your preference to keep scanlines separated.
This function is useful when you want to emulate handhelds screens using integer scaling,
where cells are well separated.
Anti-moire sweet spot:
When dealing with curvature and deep scanlines gaps, moire patterns could appear on screen.
This setting staggers screen phosphors by the configured amount and that halps in mitigating
the disturbing effect.
I observed that a value of 0.17 does a good job for low-res content rendered at 1080p height.
Any value > 0.0 disables the, following functions: Slotmask(fake) and Deconvergence Y
Slotmask(fake) offset(*):
This will cause every cell to be vertically shifted by the configured amount to
emulate a slotmask phosphors layout.
It is true that for accurate reproduction of them, slotmasks are commonly emulated
at screen size, but this causes, on low resolution displays, weird artifacts,
primarily when using curvature and when you try to draw scanlines -and- slotmasks.
Here there is an added value given by the fact that the shift itself
can be relative to not only to the screen pixel height, but to game pixel height. (**)
By selecting Y resolution=0 (so core coordinates**) and enabling this slotmask offset,
you will have a staggered scanline.
This allows you to not drawing a scanline -and- a slotmask, but to draw a "slotmasked"
scanline.
While this does not exist at all in crt technology, it greatly mitigates the artifacts
just explained while producing a fairly convincing effect, very similar to a screen
with visible scanlines and visible slotmask.
Deconvergence Y: R,G,B phosphor"
This emulates Y deconvergence on phosphor level rather than on image level as seen in
the previous deconvergence section.
Emulating deconvergence here is good because phosphors will be able to brighten the
dark gap left by scanlines.
Dedot mask between scanlines
When using Horizontal masks, you mai notice a disturbing dot pattern left between high
scanlines, that's the residual of horizontal mask.
Use this parameter to clear it and use it only if needed or it would have counter-effects.
Also, mutating dots to straight lines would make moiree more visible when using curvature.
Horizontal mask (rgb subpixel mask strength)
X resolution: (core or screen) (**):
0: Phosphors width will be relative to the pixel width of the core (game).
1: Phosphors width will be relative to the pixel width of the screen.
Cell size multiplier x (neg=divider):
Multiply (or divide if the parameter is < 0) the mask (cell) size by a factor.
As stated(**), the size may be relative to screen or core, this allow you to
"zoom" the cell horizontally and can be useful if you have an HiDPI screen.
For example, you may choose to use screen sized masks and zoom them by a factor
of 2 or 3 to make room for phosphors and see them visually grow in width.
Likewise, you can use core/game(**) sized masks and divide them by a factor
if they appear too big.
Mask type preset:
You can have the shader generate a preconfigured mask for you:
1:gm 2:gmx 3:rgb 4:rgbx 5:rbg 6:rbgx 7:wx
1:GreenMagenta, 2:GreenMagentaGap, 3:RedGreenBlue, 4:RedGreenBlueGap, 5:RedBlueGreen, 6:RedBlueGreenGap
7:WhiteGap (means r,g and b phosphors are completely overlapped, nice with scanline deconvergence)
(beware that due to limitations of the actual implementation, masks ending in "x")
works reliable when emulating slotmasks only at screen coordinates with multiplier = 1.0)
bh selecting preset = 0, you can draft your own by using the following knobs:
Phosphors+gap count (mask size):
How much phosphors or blank spaces the final mask will have.
R,G,B, Phosphor position:
The position of every phosphor.
Example 1: Phosphors+gap count=4 and R=0 G=1 B=2
...will give you a mask with red,green,blue and a blank space.
Example 2: Phosphors+gap count=2 and R=1 G=0 B=1
...will give a mask with green + a mix of blue and red (magenta)
Example 3: Phosphors+gap count=3 and R=1 G=1 B=1
...will give a mask with a blank space, the neutral white color and another blank space
Example 4: Phosphors+gap count=1 and R=0 G=0 B=0
...like the previous one, but without any blank spaces.
Phosphors width Min, Max:
The width of each phosphors can grow or shrink, depending on the
luminosity of the underlying pixel luminance.
1.0 refers to the full mask width, so stay around 0.2 if you want
avoid them to blend.
Use Min and Max parameter to limit the minimum and maximum size
they can reach.
Phosphors width min->max gamma:
Since emulating phosphors with high Min-Max range changes the apparent gamma of the final image,
it is advised, if needed, to use this option to compensate, instead of the main gamma correction.
It is also a quick way to make the image brighter or darker.
Black level of the unexcided phosphor grid
Draw the vertical grid that hosts phosphors.
This is likely to produce moiree when using X resolution = core
Vertical cell Mask:
The shape of the mask generated ny this function is "boxed", while the one
generated by the previous function ("scanlines or vertical mask1") is more rounded.
Phosphor masks are usually vertically delimited by thin lines.
This parameter will set the visibility of them.
Resolution: (core or screen) (*1)
Should the vertical interval (height) between those likes be relative to screen or core pixel size?
Height divider (neg=multiplier) (*2):
The more, the less that intervall will be.
Interesting values for screen resolution: 1.0, 0.75, 0.5
Fadeout under light: How much they will be visible over bright pixels.
Even/odd offset (slotmask) (*3):
If you not used, the previous section to emulate scanlines(*), but still want to emulate
slotmasks layouts, you can set this to 1.0.
You can draw slotmasks at screen coordinates to emulate real crts or choose to paint
them at core coordinates to have a more defined slotmask
...if you like slotmasks so much :-)
Vertical shift (neg = auto):
This parameter allows you to move the whole vertical mask along the Y axis.
* When used with core resolution(*1) and integer divider/multiplier(*2), it is useful to mitigate
weird patterns from the screen when using slotmasks (*3) alongside scanline emulation (*4).
* The auto setting works with screen resolution(*1), and integer divider/multiplier(*2).
It will auto-select a shift that gives the sharpest result.
Steepness:
The more, the thinner they will be.
Setting this to very high values, may make them disappear unevenly.
Sparkling look punch:
Makes the "Vertical cell Mask" effect more pronunced and "pinchy/Sparky" by highering its contrast.
Beware, this may produce moiree.
Shadowmask:
Emulates crt's shadowmask, it is advised to turn off "Vertical cell Mask above" for accurate results
X staggering
how much even rows have to be shifted, probably you want 1.0 or 1.5 here
for 2 or 4 sized masks (gm,rgbx,rbgx,wx) use 1.5
for 3 sized masks (gmx,rgb,rbg) use 1.0
Phosphor height
If you are using a very High definition screen, you may want to set this higher than 1.0
I also observed nice results by setting this to 3 when using 4 sized masks like rgbx or rbgx.
Dot matrix emulation:
Emulates low refresh "boxed" screens used in old handheld consoles.
Grid strength: How much the grid is visible.
Gap brightness:
You may want to use a white grid eg: for gameboy mono emulation.
Grid sharpness:
Modulates the thickness of the grid around cells.
Paint on bright:
How much should the grid be visible on background?
More positive values -> more grid on bright
More negative values -> more grid on dark
Refresh inertia:
Emulates a low refresh screen, set the power of the effect.<br>
Inertia on:
0: display is always slow to refresh
1: display is slow to refresh bright pixels
2: display is slow to refresh dark pixels
Shadow strength:
Emulates the typical shadow seen on Gameboy mono handhelds
casted by on the underlying screen.
Shadow offset:
Moves the shadow left or right.
Halo:
Emulates the effect of the brighter pixels reflected by the CRT glass
that lights nearby zones (a sort of tight bloom).
The light is meant to be spreaded in a way that it nulls the effect of the
dark scanline parts, darklines and the RGB masks.
So you can use this to restore the brightness and color saturation
loss when using features like scanlines, darklines or RGB masks.
Pre-attenuate input signal gain to 1x:
Nullifies the input gain applied in the color correction section.
This way the halo effect will be consistent and will not depend on
it, avoiding hard to manage cascading effects.
Strength (negative = 10x precision)
The effect strength.
Negative values are interpreted as positive ones, divided by 10,
when fine tuning is needed.
Sharpness
The lower, the wider the halo.
Gamma in
Act like a soft treshold; the higher, the less the darker colors
will be "haloed"
Gamma out
Post-gamma correction applied to the halo.
Light up scanline gaps and dot grid gaps too:
Theoretically Halo have to be applied
"over" everything, because that is the way it works in nature.
But you can choose to cheat and instead apply scanlines over the halo
instead.
Do this if you like much more pronunced scanlines, even at the
price of some graphical artifacts visible on high contrasted areas.
The same apply for the grid emulated via dot matrix emulation feature.
Bloom:
Acts like Halo, but affects a much wider area and is more configurable.
By using this effect and playing with its parameters, you can achieve funny
or even artistic results.
Final mix:
Modulates between the original images and the bloomed one.
Radius:
Controls how much the bloom has to be wide.
Quality:
How much the shape of the bloomed picture will reflect the original one.
Input Gamma (threshold):
Use this as a threshold to control how much a pixel has to be bright
to produce a bloom effect.
Output Gamma (contour smoothness):
Lowering it will make the bloom contour more pronunced.
Handy to simulate well defined "Aura" effects.
Power multiplier:
Just apply a gain to the final bloom.
Modulate: Local exposure eye adaption strength
Simulate the process through which the pupil adapt itself to different
light conditions.
Modulate: Strength on bright areas (0 = aura)
Since the light produced by the bloom effect is added to the underlying
image, it can produce clipping effects on the already bright areas.
This is actually an hack that will avoid to bloom them.
Bypass/Solo:
See how the bloomed image looks alone.
Use 1.0 to see naked bloom without any modulation applied
Use 2.0 to see naked bloom with modulation applied
Curvature:
Emulates a curved CRT display.
WARP X, WARP Y:
control how much the display is curved along its axes.
Corners/Edges:
Corner radius, Edge sharpness:
Control the "smoothness" of the display edges.
Bezel:
Draws a monitor frame with simulated reflections from the game content.
The monitor frame is an image loaded by the shader and is shipped
in the "textures" shader subdirectory, named:
monitor_body_curved.png and monitor_body_straight.png
It has been made with the following rules that may come handy
only if you want to edit it; otherwise go on.
- The red channel represents the luminance information
- The green channel represents the highlights
- The alpha channel in the inner frame represents the part of the bezel
that will be filled by the game content
- The blue channel represents the part of the bezel that will be filled by the game reflection.
Straight
Use a straight bezel instead of a curved one.
Inner zoom:
Allows to shrink or expand the game content to fit the monitor frame.
"Fitting the monitor frame" is the ONLY scope of this option.
To have a smaller or larger than bezel screen, please use "Override content geometry"
options, or the reflections on the bezel will be misaligned.p
Frame zoom:
Allows to shrink or expand the monitor frame to fit the game content.
Bezel color (red,green,blue) and contrast:
Allows to choose the color of the monitor frame.
Reflections strength
The amount of reflections
Reflections sharpness
Modulates from totally blurred to totally sharp reflection appearance.
Reflections roughness
The amount of material roughness in reflection area
Diffusion strength
Mix an amount of high blurred reflection to simulate light diffusion
Light fadeout distance
Modulates the reflected and diffused light fadeout. higher numbers -> no fadeout.
Specularity strength
The amount of specular reflection
Darken corners
How much the bezel corners should be darkened
Backgound image:
Draws an image on screen picked from the "textures" shader subdirectory,
named by default background_over.png and background_under.png
Of course you can use other path/names, but then you have to edit the preset by modifying
bg_over and/or bg_under.
-> It is needed that you set retroarch aspect to "Full" <-
( Settings, Video, Scaling, Aspect Ratio = Full )
The image is painted "under" the game content and under the monitor frame by
default, and his alpha channel will let you see ambient lighs (see next).
Image over content (alpha channel driven)?:
...however you can paint the image over the game content and over the
monitor frame itself by selecting this option.
If you do so, the alpha channel of the background image will be used to
let you view the underlying content.
Shift(Zoom) Image over X(Y) axis:
move or zoom the whole background image.
Rotate/flip image
This could be needed when dealing with vertical games
or when the core flips the image for whatever reason.
0 = let the shader try to guess if the game content is rotated.
1, -1 = no change
>+1 = manual rotation for straight games
<-1 = manual rotation for rotated games
Wrap mode:
What to do outside the image:
0 Mirrored repeat because so is configured in main .slangp.
1 Clamp to border and it means black.
2 Clamp to edge and means that it repeats the edge color.
3 Plain repeat without mirroring.
Backdrop support:
Some old arcades used a mirror trick to overlay the
game content over an high definition printed image.
The image used by default, picked from the "textures" shader subdirectory,
is named: boothill.jpg
Shift(Zoom) Backdrop over X(Y) axis:
move or zoom the whole background image.
Ambient light leds:
Emulates the presence of led strips behind the monitor that lights the
surroundings according to the edges of the game content.
-> It is needed that you set retroarch aspect to "Full" <-
( Settings, Video, Scaling, Aspect Ratio = Full )
Slowness:
How much will the leds will take to reflect the game content.
It may sound silly to make them slow, but if they reacts too fast,
they may distract you.
Keep in mynd that there is a scene detection logic that will make them
react as fast as possible when a scene change is detected.
Led power/Falloff:
How input gain is applied to a single led, affects light wideness.
Led saturation:
Leds saturation post gain.
Internalness (Led position):
The distance between the virtual led strip and the content.
High values will move leds behind it, while lower values will move
leds around it.
Internalness (Sampling point):
Where the leds should take their color from, highering the value will make the color
follow the color of what is in the center of the screen, lowering the value will
color the leds as the edge of the screen.
Widen lights:
Dumb stretch of the visible texture, operates on the whole content, instead of the
single led.
Note: To avoid burn-in effects, keep Light Falloff + Led power not too high.
Bezel Colorization intensity:
Allow to add an amount of the ambient light over the bezel frame
Please use this only if you are using a bezel.
Bezel Colorization Size:
How much the ambient light should bleed over the monitor.
Pushing this too high would make it bleed over the tube.
This is not supported.
Back/Foreground image alpha blend:
When displaying background or foreground images, ambient lights are
shown on their transparent areas (alpha channel)
Switching this you can choose if you want to emulate leds:
under the image (0/mix) or over the image (1/add)
Foggy (power on dark colors, add mode only)
When using add mode or when colorizing the bezel via led lights, you may want
to lessen the foggy effect to turn it into a more realistic illumination.
Think to this as a way to turn the light target from the spectator (foggy)
to the background (not foggy).
Always colorize Back/Foreground image (add mode only)
When "Back/Foreground image alpha blend" is set to 1/add,
you can choose to always colorize the foreground image by a certain amount.
This can be used to produce big haloes all around the tube.
This internally works by adding the amount of your choice to the alpha channel
of the foreground image.
Vignette:
Will cause uneven brightness of the image, more at the center,
less at the edges.
Spot:
Simulates external light reflected by the monitor glass.
Aspect Ratio:
When using effects that need Retroarch aspect ratio option
to be set to "full", you have to provide the source aspect
ratio to the shader.
Use -6 for MAME cores that pre-rotates the game (TATE mode)
With Mame 2003 plus and fbneo cores, koko-aio detects if the
game is rotated or not without any user intervention.
Aspect Ratio Numerator:
Setting non positive value here will switch to a predefined
aspect ratio from the following list:
0 = 1.33 MAME
-1 = 1.55 NTSC
-2 = 1.25 PAL
-3 = 8/7 Snes
-4 = 10/7 Megadrive
-5 = Uncorrected
-6 = 0.75 MAME rotated 1.33 games in TATE mode
Aspect Ratio Denominator:
As long as Aspect Ratio Numerator is positive, this will
be used as the denominator of the fraction.
Global shift/zoom image:
Zoom and shift everything on screen, but background pictures.
Luminosity tied zoom:
On older CRT monitors, the picture gets bigger when the image was brighter.
Please TURN THIS OFF if you want to use integer scaling, since this obstructs it.
The higher, the more prominent the effect.
Autocrop: Clears solid bars around the frame.
Autocrop maximum amount:
The higher, the more solid borders wil be cropped around the image.
0.3 means 30%
Number of mandatory lines to crop:
The minimum lines to always crop; this is useful because sometimes
games have one or two "spurious" lines at the very edge of the screen that
won't allow autocrop to work at all.
This can be used to ignore them.
Samples per frame:
Higher values makes the shader search more in a single frame for solid areas.
This leads to more accurate result in less time, however it will also stress the gpu more.
Fortunately even low/lighter values like 10 will work good if you're ok
in waiting 2..3 seconds for the final crop value to be found.
Sample size:
Search multiple pixels at once, this provide a big performance boost, but less accuracy.
It means that some solid bar could remain around the image.
Scene change treshold
When autocrop finds a maximum crop value, it only tries to crop more when the scene changes.
By lowering this value, you tell the shader to try higher the crop more often.
Use 0.0 is probably useful only to trigger a new search.
Transition speed
This modulates the smoothness of the animation between various crop values.
Override content geometry:
Contrary to the global aspect ratio control, this changes only the game geometry.
Bezel stays the same.
Integer scale:
Game content zoom height is rounded to nearest integer.
Maximum integer scale:
Dont allow integer scaling more than this
* beware: the following options in this group overrides the integer scale.
Permit integer overscale by:
When doing integer scaling, allow the image to be slightly overscanned (goes off screen).
Aspect:
Change aspect ratio.
Vertical/Horizontal position:
Shifts the game position
Zoom:
Change the size
Sharp fix through offset:
When using integer scaling, it may be not possible to
draw sharp rounded lines.
This hack add a small (the higher, the smaller) offset
to the image that allow for sharp lines at the cost of a slightly
lower draw precision. Use 0.0 to disable the hack.
Tilt:
Put the bezel and the game content into perspective.
The implementation is basic, you can expect correct
results when tilting alongside a single axis or when
using both, but with small values.
Tilt along X axis:
Rotate the image in space
Fov: Modulates the field of view
Bezel multiplier:
Can be used to adjust the bezel rotation
in relation to the game tilt amount
Delta Render:
Koko-aio can render only the part of the screen that has been changed,
leading to a measurable power consumption reduction.
This feature can, however, produce artifacts in some cases, so the feature
is statically disabled by default by now.
To use it, you have to manually set to 1.0, in file config-user.txt:
#define DELTA_RENDER 0.0
to
#define DELTA_RENDER 1.0
Force refresh interval:
Forces a full screen refresh every #number of frames;
if there was artifacts on the screen, they will be cleared.
Power comsumption benefits will be lower.
Delta render area size
If you see artifacts, try to make this higher.
Artifacts come basically from bloom.
By highering this value, Delta render can take higher blur radiouses
into account.
Power comsumption benefits will be lower.