This new comment should make it easier to understand why there is no
need to encode the frame when outputting data in the raw format
supported by the device
 
                /* if we got the complete frame */
                if (got_picture) {
-                       /* convert it to YUV */
+                       /* 
+                        * Rescaling the picture also changes its pixel format
+                        * to the raw format supported by the projector if
+                        * this was set in video_output_init()
+                        */
                        sws_scale(sw_scale_ctx,
                                  (const uint8_t * const *)picture_raw->data,
                                  picture_raw->linesize,