We generally always prefer the information from upstream for other
metadata (pixel-aspect-ration, etc.) and should also do so here.
Other parsers (h264parse) already do the same.
else
GST_DEBUG_OBJECT (mpvparse, "Invalid level - %u", level_c);
- gst_caps_set_simple (caps, "interlace-mode",
- G_TYPE_STRING,
- (mpvparse->sequenceext.progressive ? "progressive" : "mixed"), NULL);
+ if (!s || !gst_structure_has_field (s, "interlace-mode"))
+ gst_caps_set_simple (caps, "interlace-mode",
+ G_TYPE_STRING,
+ (mpvparse->sequenceext.progressive ? "progressive" : "mixed"), NULL);
}
gst_pad_set_caps (GST_BASE_PARSE_SRC_PAD (mpvparse), caps);