diff --git a/filters/f_hwtransfer.c b/filters/f_hwtransfer.c
index 9488c5536c..b281a15838 100644
--- a/filters/f_hwtransfer.c
+++ b/filters/f_hwtransfer.c
@@ -198,7 +198,8 @@ static void process(struct mp_filter *f)
     }
 
     if (!mp_update_av_hw_frames_pool(&p->hw_pool, p->av_device_ctx, p->hw_imgfmt,
-                                     p->last_hw_output_fmt, src->w, src->h))
+                                     p->last_hw_output_fmt, src->w, src->h,
+                                     src->imgfmt == IMGFMT_CUDA))
     {
         MP_ERR(f, "failed to create frame pool\n");
         goto error;
@@ -353,7 +354,7 @@ static bool probe_formats(struct mp_hwupload *u, int hw_imgfmt)
         // Creates an AVHWFramesContexts with the given parameters.
         AVBufferRef *frames = NULL;
         if (!mp_update_av_hw_frames_pool(&frames, ctx->av_device_ref,
-                                         hw_imgfmt, imgfmt, 128, 128))
+                                         hw_imgfmt, imgfmt, 128, 128, false))
         {
             MP_WARN(u->f, "failed to allocate pool\n");
             continue;
diff --git a/video/filter/vf_vavpp.c b/video/filter/vf_vavpp.c
index 5b30c46163..52be148111 100644
--- a/video/filter/vf_vavpp.c
+++ b/video/filter/vf_vavpp.c
@@ -164,7 +164,8 @@ static struct mp_image *alloc_out(struct mp_filter *vf)
     int src_h = hw_frames->height;
 
     if (!mp_update_av_hw_frames_pool(&p->hw_pool, p->av_device_ref,
-                                     IMGFMT_VAAPI, IMGFMT_NV12, src_w, src_h))
+                                     IMGFMT_VAAPI, IMGFMT_NV12, src_w, src_h,
+                                     false))
     {
         MP_ERR(vf, "Failed to create hw pool.\n");
         return NULL;
diff --git a/video/mp_image_pool.c b/video/mp_image_pool.c
index 2f0259146a..8f982b414c 100644
--- a/video/mp_image_pool.c
+++ b/video/mp_image_pool.c
@@ -15,6 +15,8 @@
  * License along with mpv.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include "config.h"
+
 #include <stddef.h>
 #include <stdbool.h>
 #include <pthread.h>
@@ -22,7 +24,11 @@
 
 #include <libavutil/buffer.h>
 #include <libavutil/hwcontext.h>
+#if HAVE_VULKAN_INTEROP
+#include <libavutil/hwcontext_vulkan.h>
+#endif
 #include <libavutil/mem.h>
+#include <libavutil/pixdesc.h>
 
 #include "mpv_talloc.h"
 
@@ -354,7 +360,8 @@ done:
 
 bool mp_update_av_hw_frames_pool(struct AVBufferRef **hw_frames_ctx,
                                  struct AVBufferRef *hw_device_ctx,
-                                 int imgfmt, int sw_imgfmt, int w, int h)
+                                 int imgfmt, int sw_imgfmt, int w, int h,
+                                 bool disable_multiplane)
 {
     enum AVPixelFormat format = imgfmt2pixfmt(imgfmt);
     enum AVPixelFormat sw_format = imgfmt2pixfmt(sw_imgfmt);
@@ -385,6 +392,18 @@ bool mp_update_av_hw_frames_pool(struct AVBufferRef **hw_frames_ctx,
         hw_frames->sw_format = sw_format;
         hw_frames->width = w;
         hw_frames->height = h;
+
+#if HAVE_VULKAN_INTEROP
+        if (format == AV_PIX_FMT_VULKAN && disable_multiplane) {
+            const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(sw_format);
+            if ((desc->flags & AV_PIX_FMT_FLAG_PLANAR) &&
+                !(desc->flags & AV_PIX_FMT_FLAG_RGB)) {
+                AVVulkanFramesContext *vk_frames = hw_frames->hwctx;
+                vk_frames->flags = AV_VK_FRAME_FLAG_DISABLE_MULTIPLANE;
+            }
+        }
+#endif
+
         if (av_hwframe_ctx_init(*hw_frames_ctx) < 0) {
             av_buffer_unref(hw_frames_ctx);
             return false;
diff --git a/video/mp_image_pool.h b/video/mp_image_pool.h
index ab9065e6e9..8cb2a5f995 100644
--- a/video/mp_image_pool.h
+++ b/video/mp_image_pool.h
@@ -36,7 +36,8 @@ bool mp_image_hw_upload(struct mp_image *hw_img, struct mp_image *src);
 struct AVBufferRef;
 bool mp_update_av_hw_frames_pool(struct AVBufferRef **hw_frames_ctx,
                                  struct AVBufferRef *hw_device_ctx,
-                                 int imgfmt, int sw_imgfmt, int w, int h);
+                                 int imgfmt, int sw_imgfmt, int w, int h,
+                                 bool disable_multiplane);
 
 struct mp_image *mp_av_pool_image_hw_upload(struct AVBufferRef *hw_frames_ctx,
                                             struct mp_image *src);