diff --git a/src/renderer/wg_engine/tvgWgBindGroups.cpp b/src/renderer/wg_engine/tvgWgBindGroups.cpp
index e25e4181..20cab5d4 100644
--- a/src/renderer/wg_engine/tvgWgBindGroups.cpp
+++ b/src/renderer/wg_engine/tvgWgBindGroups.cpp
@@ -103,9 +103,15 @@ WGPUBindGroup WgBindGroupLayouts::createBindGroupStrorage3RO(WGPUTextureView tex
 
 
 WGPUBindGroup WgBindGroupLayouts::createBindGroupBuffer1Un(WGPUBuffer buff)
+{
+    return createBindGroupBuffer1Un(buff, 0, wgpuBufferGetSize(buff));
+}
+
+
+WGPUBindGroup WgBindGroupLayouts::createBindGroupBuffer1Un(WGPUBuffer buff, uint64_t offset, uint64_t size)
 {
     const WGPUBindGroupEntry bindGroupEntrys[] = {
-        { .binding = 0, .buffer = buff, .size = wgpuBufferGetSize(buff) }
+        { .binding = 0, .buffer = buff, .offset = offset, .size = size }
     };
     const WGPUBindGroupDescriptor bindGroupDesc { .layout = layoutBuffer1Un, .entryCount = 1, .entries = bindGroupEntrys };
     return wgpuDeviceCreateBindGroup(device, &bindGroupDesc);
diff --git a/src/renderer/wg_engine/tvgWgBindGroups.h b/src/renderer/wg_engine/tvgWgBindGroups.h
index 9bab0e77..ac959700 100644
--- a/src/renderer/wg_engine/tvgWgBindGroups.h
+++ b/src/renderer/wg_engine/tvgWgBindGroups.h
@@ -49,6 +49,7 @@ public:
     WGPUBindGroup createBindGroupStrorage2RO(WGPUTextureView texView0, WGPUTextureView texView1);
     WGPUBindGroup createBindGroupStrorage3RO(WGPUTextureView texView0, WGPUTextureView texView1, WGPUTextureView texView2);
     WGPUBindGroup createBindGroupBuffer1Un(WGPUBuffer buff);
+    WGPUBindGroup createBindGroupBuffer1Un(WGPUBuffer buff, uint64_t offset, uint64_t size);
     WGPUBindGroup createBindGroupBuffer2Un(WGPUBuffer buff0, WGPUBuffer buff1);
     WGPUBindGroup createBindGroupBuffer3Un(WGPUBuffer buff0, WGPUBuffer buff1, WGPUBuffer buff2);
     void releaseBindGroup(WGPUBindGroup& bindGroup);
diff --git a/src/renderer/wg_engine/tvgWgCompositor.cpp b/src/renderer/wg_engine/tvgWgCompositor.cpp
index 071fd3a0..44629e06 100644
--- a/src/renderer/wg_engine/tvgWgCompositor.cpp
+++ b/src/renderer/wg_engine/tvgWgCompositor.cpp
@@ -28,7 +28,7 @@ void WgCompositor::initialize(WgContext& context, uint32_t width, uint32_t heigh
 {
     // pipelines (external handle, do not release)
     pipelines.initialize(context);
-    stageBuffer.initialize(context);
+    stageBufferGeometry.initialize(context);
     // initialize opacity pool
     initPools(context);
     // allocate global view matrix handles
@@ -64,7 +64,8 @@ void WgCompositor::release(WgContext& context)
     context.layouts.releaseBindGroup(bindGroupViewMat);
     context.releaseBuffer(bufferViewMat);
     // release stage buffer
-    stageBuffer.release(context);
+    stageBufferPaint.release(context);
+    stageBufferGeometry.release(context);
     // release pipelines
     pipelines.release(context);
 }
@@ -190,27 +191,31 @@ void WgCompositor::endRenderPass()
 
 void WgCompositor::reset(WgContext& context)
 {
-    stageBuffer.clear();
+    stageBufferGeometry.clear();
+    stageBufferPaint.clear();
 }
 
 
 void WgCompositor::flush(WgContext& context)
 {
-    stageBuffer.append(&meshDataBlit);
-    stageBuffer.flush(context);
+    stageBufferGeometry.append(&meshDataBlit);
+    stageBufferGeometry.flush(context);
+    stageBufferPaint.flush(context);
 }
 
 
 void WgCompositor::requestShape(WgRenderDataShape* renderData)
 {
-    stageBuffer.append(renderData);
+    stageBufferGeometry.append(renderData);
+    renderData->bindGroupPaintInd = stageBufferPaint.append(renderData->paintSettings);
     // TODO: expand for fill settings
 }
 
 
 void WgCompositor::requestImage(WgRenderDataPicture* renderData)
 {
-    stageBuffer.append(renderData);
+    stageBufferGeometry.append(renderData);
+    renderData->bindGroupPaintInd = stageBufferPaint.append(renderData->paintSettings);
     // TODO: expand for fill settings
 }
 
@@ -332,8 +337,8 @@ void WgCompositor::drawMesh(WgContext& context, WgMeshData* meshData)
     uint64_t icount = meshData->ibuffer.count;
     uint64_t vsize = meshData->vbuffer.count * sizeof(Point);
     uint64_t isize = icount * sizeof(uint32_t);
-    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBuffer.vbuffer_gpu, meshData->voffset, vsize);
-    wgpuRenderPassEncoderSetIndexBuffer(renderPassEncoder, stageBuffer.ibuffer_gpu, WGPUIndexFormat_Uint32, meshData->ioffset, isize);
+    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBufferGeometry.vbuffer_gpu, meshData->voffset, vsize);
+    wgpuRenderPassEncoderSetIndexBuffer(renderPassEncoder, stageBufferGeometry.ibuffer_gpu, WGPUIndexFormat_Uint32, meshData->ioffset, isize);
     wgpuRenderPassEncoderDrawIndexed(renderPassEncoder, icount, 1, 0, 0, 0);
 };
 
@@ -345,7 +350,7 @@ void WgCompositor::drawMeshFan(WgContext& context, WgMeshData* meshData)
     uint64_t icount = (meshData->vbuffer.count - 2) * 3;
     uint64_t vsize = meshData->vbuffer.count * sizeof(Point);
     uint64_t isize = icount * sizeof(uint32_t);
-    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBuffer.vbuffer_gpu, meshData->voffset, vsize);
+    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBufferGeometry.vbuffer_gpu, meshData->voffset, vsize);
     wgpuRenderPassEncoderSetIndexBuffer(renderPassEncoder, context.bufferIndexFan, WGPUIndexFormat_Uint32, 0, isize);
     wgpuRenderPassEncoderDrawIndexed(renderPassEncoder, icount, 1, 0, 0, 0);
 };
@@ -358,9 +363,9 @@ void WgCompositor::drawMeshImage(WgContext& context, WgMeshData* meshData)
     uint64_t icount = meshData->ibuffer.count;
     uint64_t vsize = meshData->vbuffer.count * sizeof(Point);
     uint64_t isize = icount * sizeof(uint32_t);
-    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBuffer.vbuffer_gpu, meshData->voffset, vsize);
-    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 1, stageBuffer.vbuffer_gpu, meshData->toffset, vsize);
-    wgpuRenderPassEncoderSetIndexBuffer(renderPassEncoder, stageBuffer.ibuffer_gpu, WGPUIndexFormat_Uint32, meshData->ioffset, isize);
+    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 0, stageBufferGeometry.vbuffer_gpu, meshData->voffset, vsize);
+    wgpuRenderPassEncoderSetVertexBuffer(renderPassEncoder, 1, stageBufferGeometry.vbuffer_gpu, meshData->toffset, vsize);
+    wgpuRenderPassEncoderSetIndexBuffer(renderPassEncoder, stageBufferGeometry.ibuffer_gpu, WGPUIndexFormat_Uint32, meshData->ioffset, isize);
     wgpuRenderPassEncoderDrawIndexed(renderPassEncoder, icount, 1, 0, 0, 0);
 };
 
@@ -376,7 +381,7 @@ void WgCompositor::drawShape(WgContext& context, WgRenderDataShape* renderData)
     WGPURenderPipeline stencilPipeline = (renderData->fillRule == FillRule::NonZero) ? pipelines.nonzero : pipelines.evenodd;
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, stencilPipeline);
     // draw to stencil (first pass)
     ARRAY_FOREACH(p, renderData->meshGroupShapes.meshes)
@@ -384,7 +389,7 @@ void WgCompositor::drawShape(WgContext& context, WgRenderDataShape* renderData)
     // setup fill rules
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     WgRenderSettings& settings = renderData->renderSettingsShape;
     if (settings.fillType == WgRenderSettingsType::Solid) {
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, settings.bindGroupSolid, 0, nullptr);
@@ -418,7 +423,7 @@ void WgCompositor::blendShape(WgContext& context, WgRenderDataShape* renderData,
     WGPURenderPipeline stencilPipeline = (renderData->fillRule == FillRule::NonZero) ? pipelines.nonzero : pipelines.evenodd;
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, stencilPipeline);
     // draw to stencil (first pass)
     ARRAY_FOREACH(p, renderData->meshGroupShapes.meshes)
@@ -426,7 +431,7 @@ void WgCompositor::blendShape(WgContext& context, WgRenderDataShape* renderData,
     // setup fill rules
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 3, targetTemp0.bindGroupTexure, 0, nullptr);
     uint32_t blendMethodInd = (uint32_t)blendMethod;
     WgRenderSettings& settings = renderData->renderSettingsShape;
@@ -457,7 +462,7 @@ void WgCompositor::clipShape(WgContext& context, WgRenderDataShape* renderData)
     WGPURenderPipeline stencilPipeline = (renderData->fillRule == FillRule::NonZero) ? pipelines.nonzero : pipelines.evenodd;
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, stencilPipeline);
     // draw to stencil (first pass)
     ARRAY_FOREACH(p, renderData->meshGroupShapes.meshes)
@@ -470,7 +475,7 @@ void WgCompositor::clipShape(WgContext& context, WgRenderDataShape* renderData)
     // setup fill rules
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     WgRenderSettings& settings = renderData->renderSettingsShape;
     if (settings.fillType == WgRenderSettingsType::Solid) {
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, settings.bindGroupSolid, 0, nullptr);
@@ -500,14 +505,14 @@ void WgCompositor::drawStrokes(WgContext& context, WgRenderDataShape* renderData
         // setup stencil rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
         // draw to stencil (first pass)
         drawMesh(context, renderData->meshGroupStrokes.meshes[i]);
         // setup fill rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         WgRenderSettings& settings = renderData->renderSettingsStroke;
         if (settings.fillType == WgRenderSettingsType::Solid) {
             wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, settings.bindGroupSolid, 0, nullptr);
@@ -543,14 +548,14 @@ void WgCompositor::blendStrokes(WgContext& context, WgRenderDataShape* renderDat
         // setup stencil rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
         // draw to stencil (first pass)
         drawMesh(context, renderData->meshGroupStrokes.meshes[i]);
         // setup fill rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 3, targetTemp0.bindGroupTexure, 0, nullptr);
         uint32_t blendMethodInd = (uint32_t)blendMethod;
         WgRenderSettings& settings = renderData->renderSettingsStroke;
@@ -584,7 +589,7 @@ void WgCompositor::clipStrokes(WgContext& context, WgRenderDataShape* renderData
         // setup stencil rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
         // draw to stencil (first pass)
         drawMesh(context, renderData->meshGroupStrokes.meshes[i]);
@@ -596,7 +601,7 @@ void WgCompositor::clipStrokes(WgContext& context, WgRenderDataShape* renderData
         // setup fill rules
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         WgRenderSettings& settings = renderData->renderSettingsStroke;
         if (settings.fillType == WgRenderSettingsType::Solid) {
             wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, settings.bindGroupSolid, 0, nullptr);
@@ -623,13 +628,13 @@ void WgCompositor::drawImage(WgContext& context, WgRenderDataPicture* renderData
     // draw stencil
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
     drawMeshImage(context, &renderData->meshData);
     // draw image
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, renderData->bindGroupPicture, 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.image);
     drawMeshImage(context, &renderData->meshData);
@@ -650,14 +655,14 @@ void WgCompositor::blendImage(WgContext& context, WgRenderDataPicture* renderDat
     // setup stencil rules
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
     drawMeshImage(context, &renderData->meshData);
     // blend image
     uint32_t blendMethodInd = (uint32_t)blendMethod;
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, renderData->bindGroupPicture, 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 3, targetTemp0.bindGroupTexure, 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.image_blend[blendMethodInd]);
@@ -674,7 +679,7 @@ void WgCompositor::clipImage(WgContext& context, WgRenderDataPicture* renderData
     // setup stencil rules
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.direct);
     drawMeshImage(context, &renderData->meshData);
     // merge depth and stencil buffer
@@ -685,7 +690,7 @@ void WgCompositor::clipImage(WgContext& context, WgRenderDataPicture* renderData
     // draw image
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, renderData->bindGroupPicture, 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.image);
     drawMeshImage(context, &renderData->meshData);
@@ -734,7 +739,7 @@ void WgCompositor::blendScene(WgContext& context, WgRenderTarget* scene, WgCompo
 void WgCompositor::markupClipPath(WgContext& context, WgRenderDataShape* renderData)
 {
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
     // markup stencil
     if (renderData->meshGroupStrokes.meshes.count > 0) {
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 255);
@@ -765,7 +770,7 @@ void WgCompositor::renderClipPath(WgContext& context, WgRenderDataPaint* paint)
     // copy stencil to depth
     wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData0->bindGroupPaint, 0, nullptr);
+    wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData0->bindGroupPaintInd], 0, nullptr);
     wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[128], 0, nullptr);
     wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.copy_stencil_to_depth);
     drawMeshFan(context, &renderData0->meshDataBBox);
@@ -777,31 +782,31 @@ void WgCompositor::renderClipPath(WgContext& context, WgRenderDataPaint* paint)
         markupClipPath(context, renderData);
         // copy stencil to depth (clear stencil)
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[190], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.copy_stencil_to_depth_interm);
         drawMeshFan(context, &renderData->meshDataBBox);
         // copy depth to stencil
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 1);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[190], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.copy_depth_to_stencil);
         drawMeshFan(context, &renderData->meshDataBBox);
         // clear depth current (keep stencil)
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[255], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.clear_depth);
         drawMeshFan(context, &renderData->meshDataBBox);
         // clear depth original (keep stencil)
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData0->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData0->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[255], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.clear_depth);
         drawMeshFan(context, &renderData0->meshDataBBox);
         // copy stencil to depth (clear stencil)
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[128], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.copy_stencil_to_depth);
         drawMeshFan(context, &renderData->meshDataBBox);
@@ -822,7 +827,7 @@ void WgCompositor::clearClipPath(WgContext& context, WgRenderDataPaint* paint)
         // set transformations
         wgpuRenderPassEncoderSetStencilReference(renderPassEncoder, 0);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 0, bindGroupViewMat, 0, nullptr);
-        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, renderData->bindGroupPaint, 0, nullptr);
+        wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 1, stageBufferPaint[renderData->bindGroupPaintInd], 0, nullptr);
         wgpuRenderPassEncoderSetBindGroup(renderPassEncoder, 2, bindGroupOpacities[255], 0, nullptr);
         wgpuRenderPassEncoderSetPipeline(renderPassEncoder, pipelines.clear_depth);
         drawMeshFan(context, &renderData->meshDataBBox);
diff --git a/src/renderer/wg_engine/tvgWgCompositor.h b/src/renderer/wg_engine/tvgWgCompositor.h
index 648248d9..b04069c4 100644
--- a/src/renderer/wg_engine/tvgWgCompositor.h
+++ b/src/renderer/wg_engine/tvgWgCompositor.h
@@ -39,7 +39,8 @@ private:
     // pipelines
     WgPipelines pipelines{};
     // stage buffers
-    WgRenderDataStageBuffer stageBuffer{};
+    WgStageBufferGeometry stageBufferGeometry{};
+    WgStageBufferUniform<WgShaderTypePaintSettings> stageBufferPaint;
     // global stencil/depth buffer handles
     WGPUTexture texDepthStencil{};
     WGPUTextureView texViewDepthStencil{};
diff --git a/src/renderer/wg_engine/tvgWgPipelines.cpp b/src/renderer/wg_engine/tvgWgPipelines.cpp
index 4466cbbc..c8c8918b 100644
--- a/src/renderer/wg_engine/tvgWgPipelines.cpp
+++ b/src/renderer/wg_engine/tvgWgPipelines.cpp
@@ -167,17 +167,17 @@ void WgPipelines::initialize(WgContext& context)
 
     const WgBindGroupLayouts& layouts = context.layouts;
     // bind group layouts helpers
-    const WGPUBindGroupLayout bindGroupLayoutsStencil[] { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un };
-    const WGPUBindGroupLayout bindGroupLayoutsDepth[]   { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutBuffer1Un };
+    const WGPUBindGroupLayout bindGroupLayoutsStencil[] { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un };
+    const WGPUBindGroupLayout bindGroupLayoutsDepth[]   { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutBuffer1Un };
     // bind group layouts normal blend
-    const WGPUBindGroupLayout bindGroupLayoutsSolid[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutBuffer1Un };
-    const WGPUBindGroupLayout bindGroupLayoutsGradient[] { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutTexSampledBuff2Un };
-    const WGPUBindGroupLayout bindGroupLayoutsImage[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutTexSampled };
+    const WGPUBindGroupLayout bindGroupLayoutsSolid[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutBuffer1Un };
+    const WGPUBindGroupLayout bindGroupLayoutsGradient[] { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutTexSampledBuff2Un };
+    const WGPUBindGroupLayout bindGroupLayoutsImage[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutTexSampled };
     const WGPUBindGroupLayout bindGroupLayoutsScene[]    { layouts.layoutTexSampled, layouts.layoutBuffer1Un };
     // bind group layouts custom blend
-    const WGPUBindGroupLayout bindGroupLayoutsSolidBlend[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutBuffer1Un, layouts.layoutTexSampled };
-    const WGPUBindGroupLayout bindGroupLayoutsGradientBlend[] { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutTexSampledBuff2Un, layouts.layoutTexSampled };
-    const WGPUBindGroupLayout bindGroupLayoutsImageBlend[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer2Un, layouts.layoutTexSampled, layouts.layoutTexSampled };
+    const WGPUBindGroupLayout bindGroupLayoutsSolidBlend[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutTexSampled };
+    const WGPUBindGroupLayout bindGroupLayoutsGradientBlend[] { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutTexSampledBuff2Un, layouts.layoutTexSampled };
+    const WGPUBindGroupLayout bindGroupLayoutsImageBlend[]    { layouts.layoutBuffer1Un, layouts.layoutBuffer1Un, layouts.layoutTexSampled, layouts.layoutTexSampled };
     const WGPUBindGroupLayout bindGroupLayoutsSceneBlend[]    { layouts.layoutTexSampled, layouts.layoutTexSampled, layouts.layoutBuffer1Un };
     // bind group layouts scene compose
     const WGPUBindGroupLayout bindGroupLayoutsSceneCompose[] { layouts.layoutTexSampled, layouts.layoutTexSampled };
diff --git a/src/renderer/wg_engine/tvgWgRenderData.cpp b/src/renderer/wg_engine/tvgWgRenderData.cpp
index f6eea81e..eb7bd52d 100644
--- a/src/renderer/wg_engine/tvgWgRenderData.cpp
+++ b/src/renderer/wg_engine/tvgWgRenderData.cpp
@@ -284,23 +284,14 @@ void WgRenderSettings::release(WgContext& context)
 
 void WgRenderDataPaint::release(WgContext& context)
 {
-    context.layouts.releaseBindGroup(bindGroupPaint);
-    context.releaseBuffer(bufferModelMat);
-    context.releaseBuffer(bufferBlendSettings);
     clips.clear();
 };
 
 
 void WgRenderDataPaint::update(WgContext& context, const tvg::Matrix& transform, tvg::ColorSpace cs, uint8_t opacity)
 {
-    WgShaderTypeMat4x4f modelMat(transform);
-    WgShaderTypeVec4f blendSettings(cs, opacity);
-    bool bufferModelMatChanged = context.allocateBufferUniform(bufferModelMat, &modelMat, sizeof(modelMat));
-    bool bufferBlendSettingsChanged = context.allocateBufferUniform(bufferBlendSettings, &blendSettings, sizeof(blendSettings));
-    if (bufferModelMatChanged || bufferBlendSettingsChanged) {
-        context.layouts.releaseBindGroup(bindGroupPaint);
-        bindGroupPaint = context.layouts.createBindGroupBuffer2Un(bufferModelMat, bufferBlendSettings);
-    }
+    paintSettings.modelMat.update(transform);
+    paintSettings.blendSettings.update(cs, opacity);
 }
 
 
@@ -676,10 +667,10 @@ void WgRenderDataEffectParamsPool::release(WgContext& context)
 }
 
 //***********************************************************************
-// WgRenderDataStageBuffer
+// WgStageBufferGeometry
 //***********************************************************************
 
-void WgRenderDataStageBuffer::append(WgMeshData* meshData)
+void WgStageBufferGeometry::append(WgMeshData* meshData)
 {
     assert(meshData);
     uint32_t vsize = meshData->vbuffer.count * sizeof(meshData->vbuffer[0]);
@@ -712,13 +703,13 @@ void WgRenderDataStageBuffer::append(WgMeshData* meshData)
 }
 
 
-void WgRenderDataStageBuffer::append(WgMeshDataGroup* meshDataGroup)
+void WgStageBufferGeometry::append(WgMeshDataGroup* meshDataGroup)
 {
     ARRAY_FOREACH(p, meshDataGroup->meshes) append(*p);
 }
 
 
-void WgRenderDataStageBuffer::append(WgRenderDataShape* renderDataShape)
+void WgStageBufferGeometry::append(WgRenderDataShape* renderDataShape)
 {
     append(&renderDataShape->meshGroupShapes);
     append(&renderDataShape->meshGroupShapesBBox);
@@ -730,7 +721,7 @@ void WgRenderDataStageBuffer::append(WgRenderDataShape* renderDataShape)
 }
 
 
-void WgRenderDataStageBuffer::append(WgRenderDataPicture* renderDataPicture)
+void WgStageBufferGeometry::append(WgRenderDataPicture* renderDataPicture)
 {
     append(&renderDataPicture->meshData);
     ARRAY_FOREACH(p, renderDataPicture->clips)
@@ -738,28 +729,28 @@ void WgRenderDataStageBuffer::append(WgRenderDataPicture* renderDataPicture)
 }
 
 
-void WgRenderDataStageBuffer::release(WgContext& context)
+void WgStageBufferGeometry::release(WgContext& context)
 {
     context.releaseBuffer(vbuffer_gpu);
     context.releaseBuffer(ibuffer_gpu);
 }
 
 
-void WgRenderDataStageBuffer::clear()
+void WgStageBufferGeometry::clear()
 {
     vbuffer.clear();
     ibuffer.clear();
 }
 
 
-void WgRenderDataStageBuffer::flush(WgContext& context) 
+void WgStageBufferGeometry::flush(WgContext& context) 
 {
     context.allocateBufferVertex(vbuffer_gpu, (float *)vbuffer.data, vbuffer.count);
     context.allocateBufferIndex(ibuffer_gpu, (uint32_t *)ibuffer.data, ibuffer.count);
 }
 
 
-void WgRenderDataStageBuffer::bind(WGPURenderPassEncoder renderPass, size_t voffset, size_t toffset)
+void WgStageBufferGeometry::bind(WGPURenderPassEncoder renderPass, size_t voffset, size_t toffset)
 {
     wgpuRenderPassEncoderSetVertexBuffer(renderPass, 0, vbuffer_gpu, voffset, vbuffer.count - voffset);
     wgpuRenderPassEncoderSetVertexBuffer(renderPass, 1, vbuffer_gpu, toffset, vbuffer.count - toffset);
diff --git a/src/renderer/wg_engine/tvgWgRenderData.h b/src/renderer/wg_engine/tvgWgRenderData.h
index ba3805a2..7d5b14b7 100644
--- a/src/renderer/wg_engine/tvgWgRenderData.h
+++ b/src/renderer/wg_engine/tvgWgRenderData.h
@@ -94,9 +94,8 @@ struct WgRenderSettings
 
 struct WgRenderDataPaint
 {
-    WGPUBuffer bufferModelMat{};
-    WGPUBuffer bufferBlendSettings{};
-    WGPUBindGroup bindGroupPaint{};
+    WgShaderTypePaintSettings paintSettings;
+    uint32_t bindGroupPaintInd{};
     RenderRegion viewport{};
     BBox aabb{{},{}};
     float opacity{};
@@ -221,7 +220,7 @@ public:
     void release(WgContext& context);
 };
 
-class WgRenderDataStageBuffer {
+class WgStageBufferGeometry {
 private:
     Array<uint8_t> vbuffer;
     Array<uint8_t> ibuffer;
@@ -240,4 +239,50 @@ public:
     void bind(WGPURenderPassEncoder renderPass, size_t voffset, size_t toffset);
 };
 
+// typed uniform stage buffer with related bind groups handling
+template<typename T>
+class WgStageBufferUniform {
+private:
+    Array<T> ubuffer;
+    WGPUBuffer ubuffer_gpu{};
+    Array<WGPUBindGroup> bbuffer;
+public:
+    // append uniform data to cpu staged buffer and return related bind group index
+    uint32_t append(const T& value) {
+        ubuffer.push(value);
+        return ubuffer.count - 1;
+    }
+
+    void flush(WgContext& context) {
+        // flush data to gpu buffer from cpu memory including reserved data to prevent future gpu buffer reallocations
+        bool bufferChanged = context.allocateBufferUniform(ubuffer_gpu, (void*)ubuffer.data, ubuffer.reserved*sizeof(T));
+        // if gpu buffer handle was changed we must to remove all created binding groups
+        if (bufferChanged) releaseBindGroups(context);
+        // allocate bind groups for all new data items
+        for (uint32_t i = bbuffer.count; i < ubuffer.count; i++)
+            bbuffer.push(context.layouts.createBindGroupBuffer1Un(ubuffer_gpu, i*sizeof(T), sizeof(T)));
+        assert(bbuffer.count >= ubuffer.count);
+    }
+
+    // please, use index that was returned from append method
+    WGPUBindGroup operator[](const uint32_t index) const {
+        return bbuffer[index];
+    }
+
+    void clear() {
+        ubuffer.clear();
+    }
+
+    void release(WgContext& context) {
+        context.releaseBuffer(ubuffer_gpu);
+        releaseBindGroups(context);
+    }
+
+    void releaseBindGroups(WgContext& context) {
+        ARRAY_FOREACH(p, bbuffer)
+            context.layouts.releaseBindGroup(*p);
+        bbuffer.clear();
+    }
+};
+
 #endif // _TVG_WG_RENDER_DATA_H_
diff --git a/src/renderer/wg_engine/tvgWgShaderSrc.cpp b/src/renderer/wg_engine/tvgWgShaderSrc.cpp
index 53fa134a..052bbe9a 100644
--- a/src/renderer/wg_engine/tvgWgShaderSrc.cpp
+++ b/src/renderer/wg_engine/tvgWgShaderSrc.cpp
@@ -32,14 +32,15 @@
 const char* cShaderSrc_Stencil = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat  : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     return out;
 }
 
@@ -56,15 +57,16 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Depth = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat  : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var<uniform> uDepth : f32;
 
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.position.z = uDepth;
     return out;
 }
@@ -82,23 +84,23 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Solid = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat      : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var<uniform> uSolidColor : vec4f;
 
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     return out;
 }
 
 @fragment
 fn fs_main(in: VertexOutput) -> @location(0) vec4f {
     let Sc = uSolidColor;
-    let So = uBlendSettings.a;
+    let So = uPaintSettings.blendSettings.a;
     return vec4f(Sc.rgb * Sc.a * So, Sc.a * So);
 }
 )";
@@ -110,12 +112,12 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Linear = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position : vec4f, @location(0) vGradCoord : vec4f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 struct GradSettings { settings: vec4f, focal: vec4f };
 
 // uniforms
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var uSamplerGrad : sampler;
 @group(2) @binding(1) var uTextureGrad : texture_2d<f32>;
 @group(2) @binding(2) var<uniform> uSettingGrad : GradSettings;
@@ -124,7 +126,7 @@ struct GradSettings { settings: vec4f, focal: vec4f };
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.vGradCoord = uTransformGrad * vec4f(in.position.xy, 0.0, 1.0);
     return out;
 }
@@ -137,7 +139,7 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
     let ba = ed - st;
     let t = dot(pos - st, ba) / dot(ba, ba);
     let Sc = textureSample(uTextureGrad, uSamplerGrad, vec2f(t, 0.5));
-    let So = uBlendSettings.a;
+    let So = uPaintSettings.blendSettings.a;
     return vec4f(Sc.rgb * Sc.a * So, Sc.a * So);
 }
 )";
@@ -149,11 +151,11 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Radial = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position : vec4f, @location(0) vGradCoord : vec4f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 struct GradSettings { settings: vec4f, focal: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var uSamplerGrad : sampler;
 @group(2) @binding(1) var uTextureGrad : texture_2d<f32>;
 @group(2) @binding(2) var<uniform> uSettingGrad : GradSettings;
@@ -162,7 +164,7 @@ struct GradSettings { settings: vec4f, focal: vec4f };
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.vGradCoord = uTransformGrad * vec4f(in.position.xy, 0.0, 1.0);
     return out;
 }
@@ -179,7 +181,7 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
     let c = 1.0*dot(d0, d0) - 1.0*r0*r0;
     let t = (-b + sqrt(b*b - 4*a*c))/(2*a);
     let Sc = textureSample(uTextureGrad, uSamplerGrad, vec2f(1.0 - t, 0.5));
-    let So = uBlendSettings.a;
+    let So = uPaintSettings.blendSettings.a;
     return vec4f(Sc.rgb * Sc.a * So, Sc.a * So);
 }
 )";
@@ -191,17 +193,17 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Image = R"(
 struct VertexInput { @location(0) position: vec2f, @location(1) texCoord: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f, @location(0) vTexCoord: vec2f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat      : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var uSampler     : sampler;
 @group(2) @binding(1) var uTextureView : texture_2d<f32>;
 
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    out.position = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    out.position = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.vTexCoord = in.texCoord;
     return out;
 }
@@ -209,7 +211,7 @@ fn vs_main(in: VertexInput) -> VertexOutput {
 @fragment
 fn fs_main(in: VertexOutput) -> @location(0) vec4f {
     var Sc: vec4f = textureSample(uTextureView, uSampler, in.vTexCoord.xy);
-    let So: f32 = uBlendSettings.a;
+    let So: f32 = uPaintSettings.blendSettings.a;
     return vec4f(Sc.rgb * Sc.a * So, Sc.a * So);
 };
 )";
@@ -248,10 +250,10 @@ fn fs_main(in: VertexOutput) -> @location(0) vec4f {
 const char* cShaderSrc_Solid_Blend = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f, @location(1) vScrCoord: vec2f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var<uniform> uSolidColor : vec4f;
 @group(3) @binding(0) var uSamplerDst : sampler;
 @group(3) @binding(1) var uTextureDst : texture_2d<f32>;
@@ -259,7 +261,7 @@ struct VertexOutput { @builtin(position) position: vec4f, @location(1) vScrCoord
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    let pos = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    let pos = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.position = pos;
     out.vScrCoord = vec2f(0.5 + pos.x * 0.5, 0.5 - pos.y * 0.5);
     return out;
@@ -274,7 +276,7 @@ fn getFragData(in: VertexOutput) -> FragData {
     var data: FragData;
     data.Sc = colorSrc.rgb;
     data.Sa = colorSrc.a;
-    data.So = uBlendSettings.a;
+    data.So = uPaintSettings.blendSettings.a;
     data.Dc = colorDst.rgb;
     data.Da = colorDst.a;
     data.Sc = data.Sa * data.So * data.Sc;
@@ -288,11 +290,11 @@ fn postProcess(d: FragData, R: vec4f) -> vec4f { return R; };
 const char* cShaderSrc_Linear_Blend = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f, @location(0) vGradCoord : vec4f, @location(1) vScrCoord: vec2f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 struct GradSettings { settings: vec4f, focal: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var uSamplerGrad : sampler;
 @group(2) @binding(1) var uTextureGrad : texture_2d<f32>;
 @group(2) @binding(2) var<uniform> uSettingGrad : vec4f;
@@ -303,7 +305,7 @@ struct GradSettings { settings: vec4f, focal: vec4f };
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    let pos = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    let pos = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.position = pos;
     out.vGradCoord = uTransformGrad * vec4f(in.position.xy, 0.0, 1.0);
     out.vScrCoord = vec2f(0.5 + pos.x * 0.5, 0.5 - pos.y * 0.5);
@@ -324,7 +326,7 @@ fn getFragData(in: VertexOutput) -> FragData {
     var data: FragData;
     data.Sc = colorSrc.rgb;
     data.Sa = colorSrc.a;
-    data.So = uBlendSettings.a;
+    data.So = uPaintSettings.blendSettings.a;
     data.Dc = colorDst.rgb;
     data.Da = colorDst.a;
     data.Sc = mix(data.Dc, data.Sc, data.Sa * data.So);
@@ -338,11 +340,11 @@ fn postProcess(d: FragData, R: vec4f) -> vec4f { return R; };
 const char* cShaderSrc_Radial_Blend = R"(
 struct VertexInput { @location(0) position: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f, @location(0) vGradCoord : vec4f, @location(1) vScrCoord: vec2f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 struct GradSettings { settings: vec4f, focal: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
 @group(2) @binding(0) var uSamplerGrad : sampler;
 @group(2) @binding(1) var uTextureGrad : texture_2d<f32>;
 @group(2) @binding(2) var<uniform> uSettingGrad : GradSettings;
@@ -353,7 +355,7 @@ struct GradSettings { settings: vec4f, focal: vec4f };
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    let pos = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    let pos = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.position = pos;
     out.vGradCoord = uTransformGrad * vec4f(in.position.xy, 0.0, 1.0);
     out.vScrCoord = vec2f(0.5 + pos.x * 0.5, 0.5 - pos.y * 0.5);
@@ -376,7 +378,7 @@ fn getFragData(in: VertexOutput) -> FragData {
     var data: FragData;
     data.Sc = colorSrc.rgb;
     data.Sa = colorSrc.a;
-    data.So = uBlendSettings.a;
+    data.So = uPaintSettings.blendSettings.a;
     data.Dc = colorDst.rgb;
     data.Da = colorDst.a;
     data.Sc = mix(data.Dc, data.Sc, data.Sa * data.So);
@@ -390,11 +392,11 @@ fn postProcess(d: FragData, R: vec4f) -> vec4f { return R; };
 const char* cShaderSrc_Image_Blend = R"(
 struct VertexInput { @location(0) position: vec2f, @location(1) texCoord: vec2f };
 struct VertexOutput { @builtin(position) position: vec4f, @location(0) vTexCoord : vec2f, @location(1) vScrCoord: vec2f };
+struct PaintSettings { modelMat : mat4x4f, blendSettings: vec4f };
 
 @group(0) @binding(0) var<uniform> uViewMat : mat4x4f;
-@group(1) @binding(0) var<uniform> uModelMat      : mat4x4f;
-@group(1) @binding(1) var<uniform> uBlendSettings : vec4f;
-@group(2) @binding(0) var uSamplerSrc     : sampler;
+@group(1) @binding(0) var<uniform> uPaintSettings : PaintSettings;
+@group(2) @binding(0) var uSamplerSrc : sampler;
 @group(2) @binding(1) var uTextureSrc : texture_2d<f32>;
 @group(3) @binding(0) var uSamplerDst : sampler;
 @group(3) @binding(1) var uTextureDst : texture_2d<f32>;
@@ -402,7 +404,7 @@ struct VertexOutput { @builtin(position) position: vec4f, @location(0) vTexCoord
 @vertex
 fn vs_main(in: VertexInput) -> VertexOutput {
     var out: VertexOutput;
-    let pos = uViewMat * uModelMat * vec4f(in.position.xy, 0.0, 1.0);
+    let pos = uViewMat * uPaintSettings.modelMat * vec4f(in.position.xy, 0.0, 1.0);
     out.position = pos;
     out.vTexCoord = in.texCoord;
     out.vScrCoord = vec2f(0.5 + pos.x * 0.5, 0.5 - pos.y * 0.5);
@@ -418,7 +420,7 @@ fn getFragData(in: VertexOutput) -> FragData {
     var data: FragData;
     data.Sc = colorSrc.rgb;
     data.Sa = colorSrc.a;
-    data.So = uBlendSettings.a;
+    data.So = uPaintSettings.blendSettings.a;
     data.Dc = colorDst.rgb;
     data.Da = colorDst.a;
     data.Sc = data.Sc * data.So;
diff --git a/src/renderer/wg_engine/tvgWgShaderTypes.h b/src/renderer/wg_engine/tvgWgShaderTypes.h
index 66853217..400426b0 100644
--- a/src/renderer/wg_engine/tvgWgShaderTypes.h
+++ b/src/renderer/wg_engine/tvgWgShaderTypes.h
@@ -55,6 +55,16 @@ struct WgShaderTypeVec4f
     void update(const RenderRegion& r);
 };
 
+// base paint settings - transform and blend settings (alligned to 256 bytes)
+struct WgShaderTypePaintSettings
+{
+    WgShaderTypeMat4x4f modelMat;
+    WgShaderTypeVec4f blendSettings;
+    uint8_t _padding[256 - sizeof(modelMat) - sizeof(blendSettings)];
+};
+// must be alligned with in-shader data type
+static_assert(sizeof(WgShaderTypePaintSettings) == 256);
+
 // sampler, texture, vec4f
 #define WG_TEXTURE_GRADIENT_SIZE 512
 struct WgShaderTypeGradient