// Copyright (C) 2017 The Qt Company Ltd. // Copyright (C) 2017 Jolla Ltd, author: // SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only // Qt-Security score:significant reason:default import QtQuick import QtQuick.Window import Qt5Compat.GraphicalEffects.private /*! \qmltype GaussianBlur \inqmlmodule Qt5Compat.GraphicalEffects \since QtGraphicalEffects 1.0 \inherits QtQuick::Item \ingroup qtgraphicaleffects-blur \brief Applies a higher quality blur effect. \note The MultiEffect component provides a blur effect with some performance advantages. For new code, it should be considered as an alternative to using Qt Graphical Effects. GaussianBlur effect softens the image by blurring it with an algorithm that uses the Gaussian function to calculate the effect. The effect produces higher quality than \l{Qt5Compat.GraphicalEffects::FastBlur}{FastBlur}, but is slower to render. In particular, modifying properties of Gaussian blur will in many cases require rebuilding the shader code from scratch, so animating properties of the effect may perform badly as well as require a large amount of cache space. Therefore, \l{FastBlur} should usually be preferred for animated blurs. \table \header \li Source \li Effect applied \row \li \image Original_bug.png \li \image GaussianBlur_bug.png \endtable \note This effect is available when running with OpenGL. \section1 Example The following example shows how to apply the effect. \snippet GaussianBlur-example.qml example Performing blur live is a costly operation. Fullscreen gaussian blur with even a moderate number of samples will only run at 60 fps on highend graphics hardware. */ Item { id: root /*! This property defines the source item that is going to be blurred. \note It is not supported to let the effect include itself, for instance by setting source to the effect's parent. */ property variant source /*! This property defines the distance of the neighboring pixels which affect the blurring of an individual pixel. A larger radius increases the blur effect. The ideal blur is achieved by selecting \c samples and \c radius such that \c {samples = 1 + radius * 2}, such as: \table \header \li Radius \li Samples \row \li 0 \e{(no blur)} \li 1 \row \li 1 \li 3 \row \li 2 \li 5 \row \li 3 \li 7 \endtable The value ranges from 0.0 (no blur) to inf. By default, the property is set to \c floor(samples / 2.0). \table \header \li Output examples with different radius values \li \li \row \li \image GaussianBlur_radius1.png \li \image GaussianBlur_radius2.png \li \image GaussianBlur_radius3.png \row \li \b { radius: 0 } \li \b { radius: 4 } \li \b { radius: 8 } \row \li \l samples: 16 \li \l samples: 16 \li \l samples: 16 \row \li \l deviation: 3 \li \l deviation: 3 \li \l deviation: 3 \endtable */ property real radius: Math.floor(samples / 2); /*! This property defines how many samples are taken per pixel when blur calculation is done. Larger value produces better quality, but is slower to render. Ideally, this value should be twice as large as the highest required radius value plus 1, for example, if the radius is animated between 0.0 and 4.0, samples should be set to 9. By default, the property is set to \c 9. \note This property is not intended to be animated. Changing this property may cause the underlying OpenGL shaders to be recompiled. */ property int samples: 9 /*! This property is a parameter to the gaussian function that is used when calculating neighboring pixel weights for the blurring. A larger deviation causes image to appear more blurry, but it also reduces the quality of the blur. A very large deviation value causes the effect to look a bit similar to what, for exmple, a box blur algorithm produces. A too small deviation values makes the effect insignificant for the pixels near the radius. \image GaussianBlur_deviation_graph.png \caption The image above shows the Gaussian function with two different deviation values, yellow (1) and cyan (2.7). The y-axis shows the weights, the x-axis shows the pixel distance. The value ranges from 0.0 (no deviation) to inf (maximum deviation). By default, deviation is bound to radius. When radius increases, deviation is automatically increased linearly. With the radius value of 8, the deviation default value becomes approximately 2.7034. This value produces a compromise between the blur quality and overall blurriness. \table \header \li Output examples with different deviation values \li \li \row \li \image GaussianBlur_deviation1.png \li \image GaussianBlur_deviation2.png \li \image GaussianBlur_deviation3.png \row \li \b { deviation: 1 } \li \b { deviation: 2 } \li \b { deviation: 4 } \row \li \l radius: 8 \li \l radius: 8 \li \l radius: 8 \row \li \l samples: 16 \li \l samples: 16 \li \l samples: 16 \endtable */ property real deviation: (radius + 1.0) / 3.3333 /*! This property defines the blur behavior near the edges of the item, where the pixel blurring is affected by the pixels outside the source edges. If the property is set to \c true, the pixels outside the source are interpreted to be transparent, which is similar to OpenGL clamp-to-border extension. The blur is expanded slightly outside the effect item area. If the property is set to \c false, the pixels outside the source are interpreted to contain the same color as the pixels at the edge of the item, which is similar to OpenGL clamp-to-edge behavior. The blur does not expand outside the effect item area. By default, the property is set to \c false. \table \header \li Output examples with different transparentBorder values \li \li \row \li \image GaussianBlur_transparentBorder1.png \li \image GaussianBlur_transparentBorder2.png \row \li \b { transparentBorder: false } \li \b { transparentBorder: true } \row \li \l radius: 8 \li \l radius: 8 \row \li \l samples: 16 \li \l samples: 16 \row \li \l deviation: 2.7 \li \l deviation: 2.7 \endtable */ property bool transparentBorder: false /*! This property allows the effect output pixels to be cached in order to improve the rendering performance. Every time the source or effect properties are changed, the pixels in the cache must be updated. Memory consumption is increased, because an extra buffer of memory is required for storing the effect output. It is recommended to disable the cache when the source or the effect properties are animated. By default, the property is set to \c false. */ property bool cached: false // private members... /*! \internal */ property int _paddedTexWidth: transparentBorder ? width + 2 * radius: width; /*! \internal */ property int _paddedTexHeight: transparentBorder ? height + 2 * radius: height; /*! \internal */ property int _kernelRadius: Math.max(0, samples / 2); /*! \internal */ property int _kernelSize: _kernelRadius * 2 + 1; /*! \internal */ property real _dpr: Screen.devicePixelRatio; /*! \internal */ property bool _alphaOnly: false; /*! \internal */ property var _maskSource: undefined /*! \internal */ property alias _output: sourceProxy.output; /*! \internal */ property alias _outputRect: sourceProxy.sourceRect; /*! \internal */ property alias _color: verticalBlur.color; /*! \internal */ property real _thickness: 0; /*! \internal */ property bool _componentIsComplete: false onSamplesChanged: _rebuildShaders(); on_KernelSizeChanged: _rebuildShaders(); onDeviationChanged: _rebuildShaders(); on_DprChanged: _rebuildShaders(); on_MaskSourceChanged: _rebuildShaders(); Component.onCompleted: { _componentIsComplete = true _rebuildShaders(); } /*! \internal */ function _rebuildShaders() { if (!_componentIsComplete) return var params = { radius: _kernelRadius, // Limit deviation to something very small avoid getting NaN in the shader. deviation: Math.max(0.00001, deviation), alphaOnly: root._alphaOnly, masked: _maskSource != undefined, fallback: root.radius != _kernelRadius } var shaders = ShaderBuilder.gaussianBlur(params); horizontalBlur.fragmentShader = shaders.fragmentShader; horizontalBlur.vertexShader = shaders.vertexShader; } SourceProxy { id: sourceProxy interpolation: SourceProxy.LinearInterpolation input: root.source sourceRect: root.transparentBorder ? Qt.rect(-root.radius, 0, root._paddedTexWidth, parent.height) : Qt.rect(0, 0, 0, 0) } ShaderEffect { id: horizontalBlur width: root.transparentBorder ? root._paddedTexWidth : root.width height: root.height; // Used by all shaders property Item source: sourceProxy.output; property real spread: root.radius / root._kernelRadius; property var dirstep: Qt.vector2d(1 / (root._paddedTexWidth * root._dpr), 0); // Used by fallback shader (sampleCount exceeds number of varyings) property real deviation: root.deviation // Only in use for DropShadow and Glow property color color: "white" property real thickness: Math.max(0, Math.min(0.98, 1 - root._thickness * 0.98)); // Only in use for MaskedBlur property var mask: root._maskSource; layer.enabled: true layer.smooth: true layer.sourceRect: root.transparentBorder ? Qt.rect(0, -root.radius, width, root._paddedTexHeight) : Qt.rect(0, 0, 0, 0) visible: false blending: false } ShaderEffect { id: verticalBlur x: transparentBorder ? -root.radius : 0 y: x; width: root.transparentBorder ? root._paddedTexWidth: root.width height: root.transparentBorder ? root._paddedTexHeight : root.height; fragmentShader: horizontalBlur.fragmentShader vertexShader: horizontalBlur.vertexShader property Item source: horizontalBlur property real spread: horizontalBlur.spread property var dirstep: Qt.vector2d(0, 1 / (root._paddedTexHeight * root._dpr)); property real deviation: horizontalBlur.deviation property color color: "black" property real thickness: horizontalBlur.thickness; property var mask: horizontalBlur.mask; visible: true } ShaderEffectSource { id: cacheItem anchors.fill: verticalBlur visible: root.cached smooth: true sourceItem: verticalBlur hideSource: visible } }