video4linux2: Avoid a floating point exception

This avoids a segfault in avconv_opt.c:opt_target when trying to
determine the norm.

avconv_opt.c

@@ -1757,7 +1757,8 @@ static int opt_target(void *optctx, const char *opt, const char *arg)
             for (j = 0; j < nb_input_files; j++) {
                 for (i = 0; i < input_files[j]->nb_streams; i++) {
                     AVCodecContext *c = input_files[j]->ctx->streams[i]->codec;
-                    if (c->codec_type != AVMEDIA_TYPE_VIDEO)
+                    if (c->codec_type != AVMEDIA_TYPE_VIDEO ||
+                        !c->time_base.num)
                         continue;
                     fr = c->time_base.den * 1000 / c->time_base.num;
                     if (fr == 25000) {