diff --git a/vkplayer.cpp b/vkplayer.cpp
index 610852f..afcdc81 100644
--- a/vkplayer.cpp
+++ b/vkplayer.cpp
@@ -53,6 +53,7 @@ struct DecoderImage
 	VkDeviceMemory image_memory;
 	// todo: deduplicate shared parameters
 	VkDeviceSize memory_offset_plane0, memory_offset_plane1;
+	VkDeviceSize stride0, stride1;
 	VkImageView image_view;
 	unsigned char *pMappedData;
 };
@@ -130,6 +131,13 @@ void CreateSoftwareImage(VkInstance inst, VulkanDevice &dev, DecoderImage &image
 	VkCommandBuffer cbuf = dev.CreateCommandBuffer();
 	VulkanTexture::SetImageLayout(cbuf, image.image, VK_IMAGE_LAYOUT_PREINITIALIZED, VK_IMAGE_LAYOUT_GENERAL, { VK_IMAGE_ASPECT_COLOR_BIT, 0, 1, 0, 1 });
 	dev.FlushCommandBuffer(cbuf,dev.defautQueue);
+	VkImageSubresource plane = {VK_IMAGE_ASPECT_PLANE_0_BIT};
+	VkSubresourceLayout layout = {};
+	vkGetImageSubresourceLayout(dev.device, image.image, &plane, &layout);
+	image.stride0 = layout.rowPitch;
+	plane.aspectMask = VK_IMAGE_ASPECT_PLANE_1_BIT;
+	vkGetImageSubresourceLayout(dev.device, image.image, &plane, &layout);
+	image.stride1 = layout.rowPitch;
 }
 
 #define MAX_DECODER_FRAMES 4
@@ -192,13 +200,14 @@ static void *DecoderThread(void*)
 				idx++;
 				DecoderImage &img = gFF.images[idx & 3];
 				// todo: wait fence?
-				memcpy(img.pMappedData + img.memory_offset_plane0 , frame->data[0],  frame->linesize[0] * frame->height);
+				for(int i = 0; i < frame->height; i++)
+					memcpy(img.pMappedData + img.memory_offset_plane0 + img.stride0 * i, frame->data[0] + frame->linesize[0]*i,  frame->linesize[0]);
 				// ffmpeg cannot NV12????
 				for(int i = 0; i < frame->height / 2; i++)
 					for(int j = 0; j < frame->linesize[1]; j++)
 					{
-						*(img.pMappedData + img.memory_offset_plane1 + frame->linesize[1]*2*i +j * 2) = *(frame->data[1] + frame->linesize[1]* i + j);
-						*(img.pMappedData + img.memory_offset_plane1 + frame->linesize[1]*2*i +j * 2+1) = *(frame->data[2] + frame->linesize[2]* i + j);
+						*(img.pMappedData + img.memory_offset_plane1 + img.stride1*i +j * 2) = *(frame->data[1] + frame->linesize[1]* i + j);
+						*(img.pMappedData + img.memory_offset_plane1 + img.stride1*i +j * 2+1) = *(frame->data[2] + frame->linesize[2]* i + j);
 					}
 				//memcpy(img.pMappedData + img.memory_offset_plane1 , frame->data[1],  frame->linesize[1] * frame->height/2);
 				//assert(!frame->data[2]);