Hello. When I generate a graph, I want all numbers and english characters in the fontfact of "times new roman", and all Chinese characters in "SimSun". I set the "times new roman" as the default fontface of graph. In Stata 14, when I export a graph, Chinese Characters which is not applicable for "times new roman" will be by default in the fontface of "SimSun" (as attached bar graph in figure 1). But in Stata 16, when I export the graph, Chinese Characters will neither in the fontface of "times new roman" nor "SimSun", but in a fontface which is default for characters which is not applicable for the user defined fonface (as figure 2). Though I can define fontface for each Chinese character in the graph program, it's truely complicated (as example code). And I also has to define the fontface of the labels of group variable through relabeling, which is supposed to be in mixed fontface of both "times new roman" and "SimSun"(as figure 3). But it obviously cannot be achieved through relabeling. I am wondering is there any easier way to achieve my purpose in Stata 16, like modifying default fontface of Chinese chracters?

Code:
graph bar (asis) 种粮比例, over(规模组,sort(排序) relabel(3 "{fontface 宋体:200-500亩}" ///
                                                          2 "{fontface 宋体:500-1000亩}" 1 "{fontface 宋体:1000亩以上}" ///
                                                          ) label(labsize(5.5) angle(20))) ///
      ytitle("{fontface 宋体:粮}" "{fontface 宋体:食}" "{fontface 宋体:种}" "植" "比" "例" "" "%" "", size(5.5) orientation(horizontal)) ///
      ylabel(,labsize(5.5)) ///
      bargap(20) ///
      blabel(total, size(4.5) format(%9.1f)) ///
      bar(1, fcolor(gs12) lcolor(black) lwidth(0.2) lpattern(solid)) ///
      xsize(12.42) ysize(5.11) ///
      scheme(s1color)
Array
Array
Array