diff --git a/player/main.c b/player/main.c
index 7f334d1128..9ecdbc3ed5 100644
--- a/player/main.c
+++ b/player/main.c
@@ -452,9 +452,12 @@ int mp_initialize(struct MPContext *mpctx)
 
     if (opts->force_vo) {
         opts->fixed_vo = 1;
-        mpctx->video_out = init_best_video_out(mpctx->global, mpctx->input,
-                                               mpctx->osd,
-                                               mpctx->encode_lavc_ctx);
+        struct vo_extra ex = {
+            .input_ctx = mpctx->input,
+            .osd = mpctx->osd,
+            .encode_lavc_ctx = mpctx->encode_lavc_ctx,
+        };
+        mpctx->video_out = init_best_video_out(mpctx->global, &ex);
         if (!mpctx->video_out) {
             MP_FATAL(mpctx, "Error opening/initializing "
                     "the selected video_out (-vo) device.\n");
diff --git a/player/video.c b/player/video.c
index ccd0bbf6c6..3c46bb0ce8 100644
--- a/player/video.c
+++ b/player/video.c
@@ -265,9 +265,12 @@ int reinit_video_chain(struct MPContext *mpctx)
 
     //================== Init VIDEO (codec & libvo) ==========================
     if (!opts->fixed_vo || !mpctx->video_out) {
-        mpctx->video_out = init_best_video_out(mpctx->global, mpctx->input,
-                                               mpctx->osd,
-                                               mpctx->encode_lavc_ctx);
+        struct vo_extra ex = {
+            .input_ctx = mpctx->input,
+            .osd = mpctx->osd,
+            .encode_lavc_ctx = mpctx->encode_lavc_ctx,
+        };
+        mpctx->video_out = init_best_video_out(mpctx->global, &ex);
         if (!mpctx->video_out) {
             MP_FATAL(mpctx, "Error opening/initializing "
                     "the selected video_out (-vo) device.\n");
diff --git a/video/out/vo.c b/video/out/vo.c
index c2ec6c5813..82da6b5369 100644
--- a/video/out/vo.c
+++ b/video/out/vo.c
@@ -203,9 +203,7 @@ static void dealloc_vo(struct vo *vo)
 }
 
 static struct vo *vo_create(bool probing, struct mpv_global *global,
-                            struct input_ctx *input_ctx, struct osd_state *osd,
-                            struct encode_lavc_context *encode_lavc_ctx,
-                            char *name, char **args)
+                            struct vo_extra *ex, char *name, char **args)
 {
     struct mp_log *log = mp_log_new(NULL, global->log, "vo");
     struct m_obj_desc desc;
@@ -220,11 +218,12 @@ static struct vo *vo_create(bool probing, struct mpv_global *global,
         .driver = desc.p,
         .opts = &global->opts->vo,
         .global = global,
-        .encode_lavc_ctx = encode_lavc_ctx,
-        .input_ctx = input_ctx,
-        .osd = osd,
+        .encode_lavc_ctx = ex->encode_lavc_ctx,
+        .input_ctx = ex->input_ctx,
+        .osd = ex->osd,
         .event_fd = -1,
         .monitor_par = 1,
+        .extra = *ex,
         .probing = probing,
         .in = talloc(vo, struct vo_internal),
     };
@@ -260,10 +259,7 @@ error:
     return NULL;
 }
 
-struct vo *init_best_video_out(struct mpv_global *global,
-                               struct input_ctx *input_ctx,
-                               struct osd_state *osd,
-                               struct encode_lavc_context *encode_lavc_ctx)
+struct vo *init_best_video_out(struct mpv_global *global, struct vo_extra *ex)
 {
     struct m_obj_settings *vo_list = global->opts->vo.video_driver_list;
     // first try the preferred drivers, with their optional subdevice param:
@@ -273,8 +269,8 @@ struct vo *init_best_video_out(struct mpv_global *global,
             if (strlen(vo_list[n].name) == 0)
                 goto autoprobe;
             bool p = !!vo_list[n + 1].name;
-            struct vo *vo = vo_create(p, global, input_ctx, osd, encode_lavc_ctx,
-                                      vo_list[n].name, vo_list[n].attribs);
+            struct vo *vo = vo_create(p, global, ex, vo_list[n].name,
+                                      vo_list[n].attribs);
             if (vo)
                 return vo;
         }
@@ -283,8 +279,8 @@ struct vo *init_best_video_out(struct mpv_global *global,
 autoprobe:
     // now try the rest...
     for (int i = 0; video_out_drivers[i]; i++) {
-        struct vo *vo = vo_create(true, global, input_ctx, osd, encode_lavc_ctx,
-                                  (char *)video_out_drivers[i]->name, NULL);
+        char *name = (char *)video_out_drivers[i]->name;
+        struct vo *vo = vo_create(true, global, ex, name, NULL);
         if (vo)
             return vo;
     }
diff --git a/video/out/vo.h b/video/out/vo.h
index fa35f65d13..60982c9d65 100644
--- a/video/out/vo.h
+++ b/video/out/vo.h
@@ -165,6 +165,12 @@ struct osd_state;
 struct mp_image;
 struct mp_image_params;
 
+struct vo_extra {
+    struct input_ctx *input_ctx;
+    struct osd_state *osd;
+    struct encode_lavc_context *encode_lavc_ctx;
+};
+
 struct vo_driver {
     // Encoding functionality, which can be invoked via --o only.
     bool encode;
@@ -273,6 +279,7 @@ struct vo {
     struct encode_lavc_context *encode_lavc_ctx;
     struct vo_internal *in;
     struct mp_vo_opts *opts;
+    struct vo_extra extra;
 
     // --- The following fields are generally only changed during initialization.
 
@@ -297,10 +304,7 @@ struct vo {
 };
 
 struct mpv_global;
-struct vo *init_best_video_out(struct mpv_global *global,
-                               struct input_ctx *input_ctx,
-                               struct osd_state *osd,
-                               struct encode_lavc_context *encode_lavc_ctx);
+struct vo *init_best_video_out(struct mpv_global *global, struct vo_extra *ex);
 int vo_reconfig(struct vo *vo, struct mp_image_params *p, int flags);
 
 int vo_control(struct vo *vo, uint32_t request, void *data);