diff --git a/libavfilter/qsvvpp.c b/libavfilter/qsvvpp.c
index a588a37610..0adcdc41c6 100644
--- a/libavfilter/qsvvpp.c
+++ b/libavfilter/qsvvpp.c
@@ -1015,3 +1015,16 @@ int ff_qsvvpp_create_mfx_session(void *ctx,
 }
 
 #endif
+
+AVFrame *ff_qsvvpp_get_video_buffer(AVFilterLink *inlink, int w, int h)
+{
+    /* When process YUV420 frames, FFmpeg uses same alignment on Y/U/V
+     * planes. VPL and MSDK use Y plane's pitch / 2 as U/V planes's
+     * pitch, which makes U/V planes 16-bytes aligned. We need to set a
+     * separate alignment to meet runtime's behaviour.
+    */
+    return ff_default_get_video_buffer2(inlink,
+                                        FFALIGN(inlink->w, 32),
+                                        FFALIGN(inlink->h, 32),
+                                        16);
+}
diff --git a/libavfilter/qsvvpp.h b/libavfilter/qsvvpp.h
index a8cfcc565a..6f7c9bfc15 100644
--- a/libavfilter/qsvvpp.h
+++ b/libavfilter/qsvvpp.h
@@ -127,4 +127,5 @@ int ff_qsvvpp_print_warning(void *log_ctx, mfxStatus err,
 int ff_qsvvpp_create_mfx_session(void *ctx, void *loader, mfxIMPL implementation,
                                  mfxVersion *pver, mfxSession *psession);
 
+AVFrame *ff_qsvvpp_get_video_buffer(AVFilterLink *inlink, int w, int h);
 #endif /* AVFILTER_QSVVPP_H */
diff --git a/libavfilter/vf_deinterlace_qsv.c b/libavfilter/vf_deinterlace_qsv.c
index 98ed7283ad..6c94923f02 100644
--- a/libavfilter/vf_deinterlace_qsv.c
+++ b/libavfilter/vf_deinterlace_qsv.c
@@ -581,6 +581,7 @@ static const AVFilterPad qsvdeint_inputs[] = {
         .name         = "default",
         .type         = AVMEDIA_TYPE_VIDEO,
         .filter_frame = qsvdeint_filter_frame,
+        .get_buffer.video = ff_qsvvpp_get_video_buffer,
     },
 };
 
diff --git a/libavfilter/vf_overlay_qsv.c b/libavfilter/vf_overlay_qsv.c
index d947a1faa1..1a2c1b1e96 100644
--- a/libavfilter/vf_overlay_qsv.c
+++ b/libavfilter/vf_overlay_qsv.c
@@ -399,11 +399,13 @@ static const AVFilterPad overlay_qsv_inputs[] = {
         .name          = "main",
         .type          = AVMEDIA_TYPE_VIDEO,
         .config_props  = config_main_input,
+        .get_buffer.video = ff_qsvvpp_get_video_buffer,
     },
     {
         .name          = "overlay",
         .type          = AVMEDIA_TYPE_VIDEO,
         .config_props  = config_overlay_input,
+        .get_buffer.video = ff_qsvvpp_get_video_buffer,
     },
 };
 
diff --git a/libavfilter/vf_scale_qsv.c b/libavfilter/vf_scale_qsv.c
index 758e730f78..36d5f3a6ec 100644
--- a/libavfilter/vf_scale_qsv.c
+++ b/libavfilter/vf_scale_qsv.c
@@ -641,6 +641,7 @@ static const AVFilterPad qsvscale_inputs[] = {
         .name         = "default",
         .type         = AVMEDIA_TYPE_VIDEO,
         .filter_frame = qsvscale_filter_frame,
+        .get_buffer.video = ff_qsvvpp_get_video_buffer,
     },
 };
 
diff --git a/libavfilter/vf_vpp_qsv.c b/libavfilter/vf_vpp_qsv.c
index a14b3de7b5..64ab8bc5a4 100644
--- a/libavfilter/vf_vpp_qsv.c
+++ b/libavfilter/vf_vpp_qsv.c
@@ -634,6 +634,7 @@ static const AVFilterPad vpp_inputs[] = {
         .name          = "default",
         .type          = AVMEDIA_TYPE_VIDEO,
         .config_props  = config_input,
+        .get_buffer.video = ff_qsvvpp_get_video_buffer,
     },
 };