Thus, we added the following few lines to the User include files to accomplish this task:
UserDeclare.h
-----------------------------------
TH2F *hilo;
-----------------------------------
UserInit.h
-----------------------------------
sprintf(str1,"lohi");
sprintf(str2,"lohi");
hilo = mkTH2F(str1,str2,2048,0,2047,512,0,511);
-----------------------------------
UserEv.h
-----------------------------------
for (j = 0; j < hdr.len; j++)
if (ev[j].ehi>0 || ev[j].elo>0)
if (ev[j].ehi<2048)
if (ev[j].elo<511)
hilo->Fill(ev[j].ehi,ev[j].elo);
-----------------------------------
After the sort, the 2D matrix "hilo" was available for inspection like
any other matrix from GSSort. The matrix showed that the lo- and hi-
res signals do not correlate perfectly and, thus, partially explained
the extra counts we found in the sumelo spectrum.