diff --git a/girs/GstVulkan-1.0.gir b/girs/GstVulkan-1.0.gir
index 6acc5b7fd3..f49839ac39 100644
--- a/girs/GstVulkan-1.0.gir
+++ b/girs/GstVulkan-1.0.gir
@@ -6324,6 +6324,9 @@ properties.</doc>
           <doc xml:space="preserve" filename="../subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c">query for vulkan formats without multiple images</doc>
           <type name="gboolean" c:type="gboolean"/>
         </parameter>
+        <parameter name="requested_usage" transfer-ownership="none">
+          <type name="Vulkan.ImageUsageFlags" c:type="VkImageUsageFlags"/>
+        </parameter>
         <parameter name="fmts" direction="out" caller-allocates="1" transfer-ownership="none">
           <doc xml:space="preserve" filename="../subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c">Vulkan formats per image/plane</doc>
           <array zero-terminated="0" c:type="VkFormat*" fixed-size="4">
diff --git a/subprojects/gst-plugins-bad/ext/vulkan/vkupload.c b/subprojects/gst-plugins-bad/ext/vulkan/vkupload.c
index d25ba5bd26..da593056ba 100644
--- a/subprojects/gst-plugins-bad/ext/vulkan/vkupload.c
+++ b/subprojects/gst-plugins-bad/ext/vulkan/vkupload.c
@@ -229,6 +229,8 @@ _raw_to_buffer_set_caps (gpointer impl, GstCaps * in_caps, GstCaps * out_caps)
   guint out_width, out_height;
   guint i;
   VkFormat vk_fmts[4] = { VK_FORMAT_UNDEFINED, };
+  VkImageUsageFlags usage =
+      VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT;
   int n_imgs;
 
   if (!gst_video_info_from_caps (&raw->in_info, in_caps))
@@ -242,7 +244,7 @@ _raw_to_buffer_set_caps (gpointer impl, GstCaps * in_caps, GstCaps * out_caps)
 
   if (!gst_vulkan_format_from_video_info_2 (raw->upload->
           device->physical_device, &raw->out_info, VK_IMAGE_TILING_OPTIMAL,
-          FALSE, vk_fmts, &n_imgs, NULL))
+          FALSE, usage, vk_fmts, &n_imgs, NULL))
     return FALSE;
 
   for (i = 0; i < n_imgs; i++) {
@@ -250,8 +252,7 @@ _raw_to_buffer_set_caps (gpointer impl, GstCaps * in_caps, GstCaps * out_caps)
 
     img_mem = (GstVulkanImageMemory *)
         gst_vulkan_image_memory_alloc (raw->upload->device, vk_fmts[i],
-        out_width, out_height, VK_IMAGE_TILING_OPTIMAL,
-        VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT,
+        out_width, out_height, VK_IMAGE_TILING_OPTIMAL, usage,
         VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT);
 
     raw->alloc_sizes[i] = img_mem->requirements.size;
diff --git a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c
index 1c2a605113..f4dbb11d29 100644
--- a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c
+++ b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.c
@@ -580,7 +580,8 @@ _get_usage (guint64 feature)
 gboolean
 gst_vulkan_format_from_video_info_2 (GstVulkanPhysicalDevice * physical_device,
     GstVideoInfo * info, VkImageTiling tiling, gboolean no_multiplane,
-    VkFormat fmts[GST_VIDEO_MAX_PLANES], int *n_imgs, VkImageUsageFlags * usage)
+    VkImageUsageFlags requested_usage, VkFormat fmts[GST_VIDEO_MAX_PLANES],
+    int *n_imgs, VkImageUsageFlags * usage_ret)
 {
   int i;
 #if (defined(VK_VERSION_1_3) || defined(VK_VERSION_1_2) && VK_HEADER_VERSION >= 195)
@@ -609,6 +610,7 @@ gst_vulkan_format_from_video_info_2 (GstVulkanPhysicalDevice * physical_device,
   for (i = 0; i < G_N_ELEMENTS (vk_formats_map); i++) {
     gboolean basics_primary = FALSE, basics_secondary = FALSE;
     guint64 feats_primary = 0, feats_secondary = 0;
+    VkImageUsageFlags usage = 0;
 
     if (vk_formats_map[i].format != GST_VIDEO_INFO_FORMAT (info))
       continue;
@@ -648,48 +650,65 @@ gst_vulkan_format_from_video_info_2 (GstVulkanPhysicalDevice * physical_device,
     if (GST_VIDEO_INFO_IS_RGB (info)) {
       if (basics_primary && (GST_VIDEO_INFO_COLORIMETRY (info).transfer ==
               GST_VIDEO_TRANSFER_SRGB)) {
-        if (fmts)
-          fmts[0] = vk_formats_map[i].vkfrmt;
-        if (n_imgs)
-          *n_imgs = 1;
-        if (usage)
-          *usage = _get_usage (feats_primary);
-      } else if (basics_secondary
+        usage = _get_usage (feats_primary);
+        if ((requested_usage & usage) == requested_usage) {
+          if (fmts)
+            fmts[0] = vk_formats_map[i].vkfrmt;
+          if (n_imgs)
+            *n_imgs = 1;
+          if (usage_ret)
+            *usage_ret = usage;
+          return TRUE;
+        }
+      }
+
+      if (basics_secondary
           && GST_VIDEO_INFO_COLORIMETRY (info).transfer !=
           GST_VIDEO_TRANSFER_SRGB) {
-        if (fmts)
-          fmts[0] = vk_formats_map[i].vkfrmts[0];
-        if (n_imgs)
-          *n_imgs = 1;
-        if (usage)
-          *usage = _get_usage (feats_secondary);
-      } else {
-        return FALSE;
+        usage = _get_usage (feats_secondary);
+        if ((requested_usage & usage) == requested_usage) {
+          if (fmts)
+            fmts[0] = vk_formats_map[i].vkfrmts[0];
+          if (n_imgs)
+            *n_imgs = 1;
+          if (usage_ret)
+            *usage_ret = usage;
+          return TRUE;
+        }
       }
+      return FALSE;
     } else {
       if (basics_primary && !no_multiplane
           && GST_VIDEO_INFO_N_PLANES (info) > 1) {
-        if (fmts)
-          fmts[0] = vk_formats_map[i].vkfrmt;
-        if (n_imgs)
-          *n_imgs = 1;
-        if (usage)
-          *usage = _get_usage (feats_primary);
-      } else if (basics_secondary) {
-        if (fmts) {
-          memcpy (fmts, vk_formats_map[i].vkfrmts,
-              GST_VIDEO_MAX_PLANES * sizeof (VkFormat));
+        usage = _get_usage (feats_primary);
+        if ((requested_usage & usage) == requested_usage) {
+          if (fmts)
+            fmts[0] = vk_formats_map[i].vkfrmt;
+          if (n_imgs)
+            *n_imgs = 1;
+          if (usage_ret)
+            *usage_ret = usage;
+          return TRUE;
         }
-        if (n_imgs)
-          *n_imgs = GST_VIDEO_INFO_N_PLANES (info);
-        if (usage)
-          *usage = _get_usage (feats_secondary);
-      } else {
-        return FALSE;
       }
-    }
 
-    return TRUE;
+      if (basics_secondary) {
+        usage = _get_usage (feats_secondary);
+        if ((requested_usage & usage) == requested_usage) {
+          if (fmts) {
+            memcpy (fmts, vk_formats_map[i].vkfrmts,
+                GST_VIDEO_MAX_PLANES * sizeof (VkFormat));
+          }
+          if (n_imgs)
+            *n_imgs = GST_VIDEO_INFO_N_PLANES (info);
+          if (usage_ret)
+            *usage_ret = usage;
+
+          return TRUE;
+        }
+      }
+      return FALSE;
+    }
   }
 
   return FALSE;
diff --git a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.h b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.h
index ad0a1e1509..56d108b74f 100644
--- a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.h
+++ b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkformat.h
@@ -145,6 +145,7 @@ gboolean                        gst_vulkan_format_from_video_info_2            (
                                                                                 GstVideoInfo * info,
                                                                                 VkImageTiling tiling,
                                                                                 gboolean no_multiplane,
+                                                                                VkImageUsageFlags requested_usage,
                                                                                 VkFormat fmts[GST_VIDEO_MAX_PLANES],
                                                                                 int * n_imgs,
                                                                                 VkImageUsageFlags * usage);
diff --git a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkimagebufferpool.c b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkimagebufferpool.c
index a1f059579b..15bb553afd 100644
--- a/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkimagebufferpool.c
+++ b/subprojects/gst-plugins-bad/gst-libs/gst/vulkan/gstvkimagebufferpool.c
@@ -183,8 +183,8 @@ gst_vulkan_image_buffer_pool_set_config (GstBufferPool * pool,
 
   tiling = priv->raw_caps ? VK_IMAGE_TILING_LINEAR : VK_IMAGE_TILING_OPTIMAL;
   found = gst_vulkan_format_from_video_info_2 (vk_pool->device->physical_device,
-      &priv->v_info, tiling, no_multiplane, priv->vk_fmts, &priv->n_imgs,
-      &supported_usage);
+      &priv->v_info, tiling, no_multiplane, priv->usage, priv->vk_fmts,
+      &priv->n_imgs, &supported_usage);
   if (!found)
     goto no_vk_format;
 
diff --git a/subprojects/gst-plugins-bad/tests/check/libs/vkformat.c b/subprojects/gst-plugins-bad/tests/check/libs/vkformat.c
index 0a04685ba3..dd36fdb015 100644
--- a/subprojects/gst-plugins-bad/tests/check/libs/vkformat.c
+++ b/subprojects/gst-plugins-bad/tests/check/libs/vkformat.c
@@ -57,13 +57,15 @@ GST_START_TEST (test_format_from_video_info_2)
           480));
 
   fail_unless (gst_vulkan_format_from_video_info_2 (phy_dev, &vinfo,
-          VK_IMAGE_TILING_OPTIMAL, TRUE, vk_fmts, &n_imgs, &supported_usage));
+          VK_IMAGE_TILING_OPTIMAL, TRUE, 0, vk_fmts, &n_imgs,
+          &supported_usage));
 
   fail_unless (n_imgs == 2 && vk_fmts[0] == VK_FORMAT_R8_UNORM
       && vk_fmts[1] == VK_FORMAT_R8G8_UNORM);
 
   fail_unless (gst_vulkan_format_from_video_info_2 (phy_dev, &vinfo,
-          VK_IMAGE_TILING_LINEAR, FALSE, vk_fmts, &n_imgs, &supported_usage));
+          VK_IMAGE_TILING_LINEAR, FALSE, 0, vk_fmts, &n_imgs,
+          &supported_usage));
 
   fail_unless (n_imgs == 1 && vk_fmts[0] == VK_FORMAT_G8_B8R8_2PLANE_420_UNORM);
 
@@ -73,7 +75,7 @@ GST_START_TEST (test_format_from_video_info_2)
       GST_VIDEO_TRANSFER_SRGB);
 
   fail_unless (gst_vulkan_format_from_video_info_2 (phy_dev, &vinfo,
-          VK_IMAGE_TILING_LINEAR, TRUE, vk_fmts, &n_imgs, &supported_usage));
+          VK_IMAGE_TILING_LINEAR, TRUE, 0, vk_fmts, &n_imgs, &supported_usage));
 
   fail_unless (n_imgs == 1 && vk_fmts[0] == VK_FORMAT_R8G8B8A8_SRGB);
 
@@ -83,7 +85,7 @@ GST_START_TEST (test_format_from_video_info_2)
       (&GST_VIDEO_INFO_COLORIMETRY (&vinfo), "smpte240m"));
 
   fail_unless (gst_vulkan_format_from_video_info_2 (phy_dev, &vinfo,
-          VK_IMAGE_TILING_LINEAR, TRUE, vk_fmts, &n_imgs, &supported_usage));
+          VK_IMAGE_TILING_LINEAR, TRUE, 0, vk_fmts, &n_imgs, &supported_usage));
 
   fail_unless (n_imgs == 1 && vk_fmts[0] == VK_FORMAT_R8G8B8A8_UNORM);
 }