diff --git a/decodevarname.m b/decodevarname.m
new file mode 100644
index 0000000..4a4e3fa
--- /dev/null
+++ b/decodevarname.m
@@ -0,0 +1,66 @@
+function newname = decodevarname(name,varargin)
+%
+%    newname = decodevarname(name)
+%
+%    Decode a hex-encoded variable name (from encodevarname) and restore
+%    its original form
+%
+%    This function is sensitive to the default charset
+%    settings in MATLAB, please call feature('DefaultCharacterSet','utf8')
+%    to set the encoding to UTF-8 before calling this function.
+%
+%    author: Qianqian Fang (q.fang <at> neu.edu)
+%
+%    input:
+%        name: a string output from encodevarname, which converts the leading non-ascii
+%              letter into "x0xHH_" and non-ascii letters into "_0xHH_"
+%              format, where hex key HH stores the ascii (or Unicode) value
+%              of the character.
+%              
+%    output:
+%        newname: the restored original string
+%
+%    example:
+%        decodevarname('x0x5F_a)   % returns _a
+%        decodevarname('a_')   % returns a_ as it is a valid variable name
+%        decodevarname('x0xE58F98__0xE9878F_')  % returns '变量' 
+%
+%    this file is part of EasyH5 Toolbox: https://github.com/fangq/easyh5
+%
+%    License: GPLv3 or 3-clause BSD license, see https://github.com/fangq/easyh5 for details
+%
+
+isunpack=jsonopt('UnpackHex',1,varargin{:});
+newname=name;
+if(isempty(regexp(name,'0x([0-9a-fA-F]+)_','once')))
+    return
+end
+if(isunpack)
+    if(exist('native2unicode','builtin'))
+        h2u=@hex2unicode;
+        newname=regexprep(name,'(^x|_){1}0x([0-9a-fA-F]+)_','${h2u($2)}');
+    else
+        pos=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','start');
+        pend=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','end');
+        if(isempty(pos))
+            return;
+        end
+        str0=name;
+        pos0=[0 pend(:)' length(name)];
+        newname='';
+        for i=1:length(pos)
+            newname=[newname str0(pos0(i)+1:pos(i)-1) char(hex2dec(str0(pos(i)+3:pend(i)-1)))];
+        end
+        if(pos(end)~=length(name))
+            newname=[newname str0(pos0(end-1)+1:pos0(end))];
+        end
+    end
+end
+
+%--------------------------------------------------------------------------
+function str=hex2unicode(hexstr)
+val=hex2dec(hexstr);
+id=histc(val,[0 2^8 2^16 2^32 2^64]);
+type={'uint8','uint16','uint32','uint64'};
+bytes=typecast(cast(val,type{id~=0}),'uint8');
+str=native2unicode(fliplr(bytes(:,1:find(bytes,1,'last'))));
diff --git a/encodevarname.m b/encodevarname.m
new file mode 100644
index 0000000..b336f2d
--- /dev/null
+++ b/encodevarname.m
@@ -0,0 +1,67 @@
+function str = encodevarname(str,varargin)
+%
+%    newname = encodevarname(name)
+%
+%    Encode an invalid variable name using a hex-format for bi-directional
+%    conversions. 
+
+%    This function is sensitive to the default charset
+%    settings in MATLAB, please call feature('DefaultCharacterSet','utf8')
+%    to set the encoding to UTF-8 before calling this function.
+%
+%    author: Qianqian Fang (q.fang <at> neu.edu)
+%
+%    input:
+%        name: a string, can be either a valid or invalid variable name
+%
+%    output:
+%        newname: a valid variable name by converting the leading non-ascii
+%              letter into "x0xHH_" and non-ascii letters into "_0xHH_"
+%              format, where HH is the ascii (or Unicode) value of the
+%              character.
+%
+%              if the encoded variable name CAN NOT be longer than 63, i.e. 
+%              the maximum variable name specified by namelengthmax, and
+%              one uses the output of this function as a struct or variable
+%              name, the name will be trucated at 63. Please consider using
+%              the name as a containers.Map key, which does not have such
+%              limit.
+%
+%    example:
+%        encodevarname('_a')   % returns x0x5F_a
+%        encodevarname('a_')   % returns a_ as it is a valid variable name
+%        encodevarname('变量')  % returns 'x0xE58F98__0xE9878F_' 
+%
+%    this file is part of EasyH5 Toolbox: https://github.com/fangq/easyh5
+%
+%    License: GPLv3 or 3-clause BSD license, see https://github.com/fangq/easyh5 for details
+%
+
+    if(~isempty(regexp(str,'^[^A-Za-z]','once')))
+        if(exist('unicode2native','builtin'))
+            str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once');
+        else
+            str=sprintf('x0x%X_%s',char(str(1))+0,str(2:end));
+        end
+    end
+    if(isvarname(str))
+        return;
+    end
+    if(exist('unicode2native','builtin'))
+        str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_');
+    else
+        cpos=regexp(str,'[^0-9A-Za-z_]');
+        if(isempty(cpos))
+            return;
+        end
+        str0=str;
+        pos0=[0 cpos(:)' length(str)];
+        str='';
+        for i=1:length(cpos)
+            str=[str str0(pos0(i)+1:cpos(i)-1) sprintf('_0x%X_',str0(cpos(i))+0)];
+        end
+        if(cpos(end)~=length(str))
+            str=[str str0(pos0(end-1)+1:pos0(end))];
+        end
+    end
+end
\ No newline at end of file
diff --git a/loadjson.m b/loadjson.m
index 87b6b2a..58d95e2 100644
--- a/loadjson.m
+++ b/loadjson.m
@@ -413,7 +413,7 @@
             end
             pos=parse_char(inputstr, pos, ':');
             [val, pos,index_esc] = parse_value(inputstr, pos, esc, index_esc, varargin{:});
-            object.(valid_field(str,varargin{:}))=val;
+            object.(encodevarname(str,varargin{:}))=val;
             [cc,pos]=next_char(inputstr,pos);
             if cc == '}'
                 break;
@@ -442,41 +442,6 @@
 
 %%-------------------------------------------------------------------------
 
-function str = valid_field(str,varargin)
-% From MATLAB doc: field names must begin with a letter, which may be
-% followed by any combination of letters, digits, and underscores.
-% Invalid characters will be converted to underscores, and the prefix
-% "x0x[Hex code]_" will be added if the first character is not a letter.
-    if(~isempty(regexp(str,'^[^A-Za-z]','once')))
-        if(~isoctavemesh && str(1)+0 > 255)
-            str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once');
-        else
-            str=sprintf('x0x%X_%s',char(str(1))+0,str(2:end));
-        end
-    end
-    if(isvarname(str))
-        return;
-    end
-    if(~isoctavemesh)
-        str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_');
-    else
-        cpos=regexp(str,'[^0-9A-Za-z_]');
-        if(isempty(cpos))
-            return;
-        end
-        str0=str;
-        pos0=[0 cpos(:)' length(str)];
-        str='';
-        for i=1:length(cpos)
-            str=[str str0(pos0(i)+1:cpos(i)-1) sprintf('_0x%X_',str0(cpos(i))+0)];
-        end
-        if(cpos(end)~=length(str))
-            str=[str str0(pos0(end-1)+1:pos0(end))];
-        end
-    end
-end
-%%-------------------------------------------------------------------------
-
 function newpos=skip_whitespace(pos, inputstr)
     newpos=pos;
     while newpos <= length(inputstr) && isspace(inputstr(newpos))
diff --git a/loadmsgpack.m b/loadmsgpack.m
index a718b45..42036c7 100644
--- a/loadmsgpack.m
+++ b/loadmsgpack.m
@@ -227,42 +227,6 @@
     out = struct();
     for n=1:len
         [key, idx] = parse(bytes, idx);
-        [out.(valid_field(char(key))), idx] = parse(bytes, idx);
-    end
-end
-
-function str = valid_field(str,varargin)
-% From MATLAB doc: field names must begin with a letter, which may be
-% followed by any combination of letters, digits, and underscores.
-% Invalid characters will be converted to underscores, and the prefix
-% "x0x[Hex code]_" will be added if the first character is not a letter.
-    isoct=exist('OCTAVE_VERSION','builtin');
-    cpos=regexp(str,'^[^A-Za-z]','once');
-    if(~isempty(cpos))
-        if(~isoct)
-            str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once');
-        else
-            str=sprintf('x0x%X_%s',char(str(1)),str(2:end));
-        end
-    end
-    if(isempty(regexp(str,'[^0-9A-Za-z_]', 'once' )))
-        return;
-    end
-    if(~isoct)
-        str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_');
-    else
-        cpos=regexp(str,'[^0-9A-Za-z_]');
-        if(isempty(cpos))
-            return;
-        end
-        str0=str;
-        pos0=[0 cpos(:)' length(str)];
-        str='';
-        for i=1:length(cpos)
-            str=[str str0(pos0(i)+1:cpos(i)-1) sprintf('_0x%X_',str0(cpos(i)))];
-        end
-        if(cpos(end)~=length(str))
-            str=[str str0(pos0(end-1)+1:pos0(end))];
-        end
+        [out.(encodevarname(char(key))), idx] = parse(bytes, idx);
     end
 end
diff --git a/loadubjson.m b/loadubjson.m
index 389d686..6ba7f7c 100644
--- a/loadubjson.m
+++ b/loadubjson.m
@@ -330,7 +330,7 @@
             end
             [val, pos] = parse_value(inputstr, pos, varargin{:});
             num=num+1;
-            object.(valid_field(str,varargin{:}))=val;
+            object.(encodevarname(str,varargin{:}))=val;
             [cc, pos]=next_char(inputstr,pos);
             if cc == '}' || (count>=0 && num>=count)
                 break;
@@ -357,38 +357,3 @@
         error_pos('unsupported type at position %d',inputstr, pos);
     end
 end
-%%-------------------------------------------------------------------------
-
-function str = valid_field(str,varargin)
-% From MATLAB doc: field names must begin with a letter, which may be
-% followed by any combination of letters, digits, and underscores.
-% Invalid characters will be converted to underscores, and the prefix
-% "x0x[Hex code]_" will be added if the first character is not a letter.
-    if(~isempty(regexp(str,'^[^A-Za-z]','once')))
-        if(~isoctavemesh && str(1)+0 > 255)
-            str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once');
-        else
-            str=sprintf('x0x%X_%s',char(str(1)),str(2:end));
-        end
-    end
-    if(isvarname(str))
-        return;
-    end
-    if(~isoctavemesh)
-        str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_');
-    else
-        cpos=regexp(str,'[^0-9A-Za-z_]');
-        if(isempty(cpos))
-            return;
-        end
-        str0=str;
-        pos0=[0 cpos(:)' length(str)];
-        str='';
-        for i=1:length(cpos)
-            str=[str str0(pos0(i)+1:cpos(i)-1) sprintf('_0x%X_',str0(cpos(i)))];
-        end
-        if(cpos(end)~=length(str))
-            str=[str str0(pos0(end-1)+1:pos0(end))];
-        end
-    end
-end
diff --git a/savejson.m b/savejson.m
index 1bd044b..3b21c0e 100644
--- a/savejson.m
+++ b/savejson.m
@@ -268,13 +268,13 @@
 bracketlevel=~jsonopt('singletcell',1,varargin{:});
 if(len>bracketlevel)
     if(~isempty(name))
-        txt={padding0, '"', checkname(name,varargin{:}),'": [', nl}; name=''; 
+        txt={padding0, '"', decodevarname(name,varargin{:}),'": [', nl}; name=''; 
     else
         txt={padding0, '[', nl};
     end
 elseif(len==0)
     if(~isempty(name))
-        txt={padding0, '"' checkname(name,varargin{:}) '": []'}; name=''; 
+        txt={padding0, '"' decodevarname(name,varargin{:}) '": []'}; name=''; 
     else
         txt={padding0, '[]'};
     end
@@ -324,7 +324,7 @@
 
 if(isempty(item)) 
     if(~isempty(name)) 
-        txt={padding0, '"', checkname(name,varargin{:}),'": []'};
+        txt={padding0, '"', decodevarname(name,varargin{:}),'": []'};
     else
         txt={padding0, '[]'};
     end
@@ -333,7 +333,7 @@
 end
 if(~isempty(name))
     if(forcearray)
-        txt={padding0, '"', checkname(name,varargin{:}),'": [', nl};
+        txt={padding0, '"', decodevarname(name,varargin{:}),'": [', nl};
     end
 else
     if(forcearray)
@@ -347,7 +347,7 @@
   for i=1:dim(1)
     names = fieldnames(item(i,j));
     if(~isempty(name) && len==1 && ~forcearray)
-        txt(end+1:end+5)={padding1, '"', checkname(name,varargin{:}),'": {', nl};
+        txt(end+1:end+5)={padding1, '"', decodevarname(name,varargin{:}),'": {', nl};
     else
         txt(end+1:end+3)={padding1, '{', nl};
     end
@@ -412,7 +412,7 @@
 
 if(isempty(item)) 
     if(~isempty(name)) 
-        txt={padding0, '"', checkname(name,varargin{:}),'": []'};
+        txt={padding0, '"', decodevarname(name,varargin{:}),'": []'};
     else
         txt={padding0, '[]'};
     end
@@ -421,7 +421,7 @@
 end
 if(~isempty(name)) 
     if(forcearray)
-        txt={padding0, '"', checkname(name,varargin{:}),'": {', nl};
+        txt={padding0, '"', decodevarname(name,varargin{:}),'": {', nl};
     end
 else
     if(forcearray)
@@ -463,7 +463,7 @@
 
 if(~isempty(name)) 
     if(len>1)
-        txt={padding1, '"', checkname(name,varargin{:}),'": [', nl};
+        txt={padding1, '"', decodevarname(name,varargin{:}),'": [', nl};
     end
 else
     if(len>1)
@@ -473,7 +473,7 @@
 for e=1:len
     val=escapejsonstring(item(e,:));
     if(len==1)
-        obj=['"' checkname(name,varargin{:}) '": ' '"',val,'"'];
+        obj=['"' decodevarname(name,varargin{:}) '": ' '"',val,'"'];
         if(isempty(name))
             obj=['"',val,'"'];
         end
@@ -514,7 +514,7 @@
               padding1,nl,padding0,class(item),nl,padding0,regexprep(mat2str(size(item)),'\s+',','),nl);
     else
     	txt=sprintf('%s"%s": {%s%s"_ArrayType_": "%s",%s%s"_ArraySize_": %s,%s',...
-              padding1,checkname(name,varargin{:}),nl,padding0,class(item),nl,padding0,regexprep(mat2str(size(item)),'\s+',','),nl);
+              padding1,decodevarname(name,varargin{:}),nl,padding0,class(item),nl,padding0,regexprep(mat2str(size(item)),'\s+',','),nl);
     end
 else
     if(numel(item)==1 && jsonopt('SingletArray',0,varargin{:})==0 && level>0)
@@ -526,9 +526,9 @@
     	txt=sprintf('%s%s',padding1,numtxt);
     else
         if(numel(item)==1 && jsonopt('SingletArray',0,varargin{:})==0)
-           	txt=sprintf('%s"%s": %s',padding1,checkname(name,varargin{:}),numtxt);
+           	txt=sprintf('%s"%s": %s',padding1,decodevarname(name,varargin{:}),numtxt);
         else
-    	    txt=sprintf('%s"%s": %s',padding1,checkname(name,varargin{:}),numtxt);
+    	    txt=sprintf('%s"%s": %s',padding1,decodevarname(name,varargin{:}),numtxt);
         end
     end
     return;
@@ -705,35 +705,6 @@
     txt=regexprep(txt,'NaN',jsonopt('NaN','"_NaN_"',varargin{:}));
 end
 
-%%-------------------------------------------------------------------------
-function newname=checkname(name,varargin)
-isunpack=jsonopt('UnpackHex',1,varargin{:});
-newname=name;
-if(isempty(regexp(name,'0x([0-9a-fA-F]+)_','once')))
-    return
-end
-if(isunpack)
-    isoct=jsonopt('IsOctave',0,varargin{:});
-    if(~isoct)
-        newname=regexprep(name,'(^x|_){1}0x([0-9a-fA-F]+)_','${native2unicode(hex2dec($2))}');
-    else
-        pos=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','start');
-        pend=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','end');
-        if(isempty(pos))
-            return;
-        end
-        str0=name;
-        pos0=[0 pend(:)' length(name)];
-        newname='';
-        for i=1:length(pos)
-            newname=[newname str0(pos0(i)+1:pos(i)-1) char(hex2dec(str0(pos(i)+3:pend(i)-1)))];
-        end
-        if(pos(end)~=length(name))
-            newname=[newname str0(pos0(end-1)+1:pos0(end))];
-        end
-    end
-end
-
 %%-------------------------------------------------------------------------
 function newstr=escapejsonstring(str)
 newstr=str;
diff --git a/saveubjson.m b/saveubjson.m
index 2cff233..302df37 100644
--- a/saveubjson.m
+++ b/saveubjson.m
@@ -265,13 +265,13 @@
 len=numel(item); % let's handle 1D cell first
 if(len>bracketlevel) 
     if(~isempty(name))
-        txt=[N_(checkname(name,varargin{:})) am0]; name=''; 
+        txt=[N_(decodevarname(name,varargin{:})) am0]; name=''; 
     else
         txt=am0; 
     end
 elseif(len==0)
     if(~isempty(name))
-        txt=[N_(checkname(name,varargin{:})) Zmarker]; name=''; 
+        txt=[N_(decodevarname(name,varargin{:})) Zmarker]; name=''; 
     else
         txt=Zmarker; 
     end
@@ -319,7 +319,7 @@
 
 if(~isempty(name)) 
     if(forcearray)
-        txt=[N_(checkname(name,varargin{:})) am0];
+        txt=[N_(decodevarname(name,varargin{:})) am0];
     end
 else
     if(forcearray)
@@ -341,7 +341,7 @@
         om0=Omarker{1};
      end
      if(~isempty(name) && len==1 && ~forcearray)
-        txt=[txt N_(checkname(name,varargin{:})) om0]; 
+        txt=[txt N_(decodevarname(name,varargin{:})) om0]; 
      else
         txt=[txt om0]; 
      end
@@ -382,7 +382,7 @@
 
 if(~isempty(name)) 
     if(forcearray)
-        txt=[N_(checkname(name,varargin{:})) om0];
+        txt=[N_(decodevarname(name,varargin{:})) om0];
     end
 else
     if(forcearray)
@@ -417,7 +417,7 @@
 end
 if(~isempty(name)) 
     if(len>1)
-        txt=[N_(checkname(name,varargin{:})) am0];
+        txt=[N_(decodevarname(name,varargin{:})) am0];
     end
 else
     if(len>1)
@@ -427,7 +427,7 @@
 for e=1:len
     val=item(e,:);
     if(len==1)
-        obj=[N_(checkname(name,varargin{:})) '' '',S_(val),''];
+        obj=[N_(decodevarname(name,varargin{:})) '' '',S_(val),''];
         if(isempty(name))
             obj=['',S_(val),''];
         end
@@ -466,10 +466,10 @@
     	txt=[Omarker{1} N_('_ArrayType_'),S_(class(item)),N_('_ArraySize_'),I_a(size(item),cid(1),Imarker,varargin{:}) ];
       else
           if(isempty(item))
-              txt=[N_(checkname(name,varargin{:})),Zmarker];
+              txt=[N_(decodevarname(name,varargin{:})),Zmarker];
               return;
           else
-    	      txt=[N_(checkname(name,varargin{:})),Omarker{1},N_('_ArrayType_'),S_(class(item)),N_('_ArraySize_'),I_a(size(item),cid(1),Imarker,varargin{:})];
+    	      txt=[N_(decodevarname(name,varargin{:})),Omarker{1},N_('_ArrayType_'),S_(class(item)),N_('_ArraySize_'),I_a(size(item),cid(1),Imarker,varargin{:})];
           end
       end
       childcount=2;
@@ -479,9 +479,9 @@
     else
         if(numel(item)==1 && jsonopt('SingletArray',0,varargin{:})==0)
             numtxt=regexprep(regexprep(matdata2ubjson(item,level+1,varargin{:}),'^\[',''),']$','');
-           	txt=[N_(checkname(name,varargin{:})) numtxt];
+           	txt=[N_(decodevarname(name,varargin{:})) numtxt];
         else
-    	    txt=[N_(checkname(name,varargin{:})),matdata2ubjson(item,level+1,varargin{:})];
+    	    txt=[N_(decodevarname(name,varargin{:})),matdata2ubjson(item,level+1,varargin{:})];
         end
     end
     return;
@@ -697,34 +697,6 @@
     end
 end
 
-%%-------------------------------------------------------------------------
-function newname=checkname(name,varargin)
-isunpack=jsonopt('UnpackHex',1,varargin{:});
-newname=name;
-if(isempty(regexp(name,'0x([0-9a-fA-F]+)_','once')))
-    return
-end
-if(isunpack)
-    isoct=jsonopt('IsOctave',0,varargin{:});
-    if(~isoct)
-        newname=regexprep(name,'(^x|_){1}0x([0-9a-fA-F]+)_','${native2unicode(hex2dec($2))}');
-    else
-        pos=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','start');
-        pend=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','end');
-        if(isempty(pos))
-            return;
-        end
-        str0=name;
-        pos0=[0 pend(:)' length(name)];
-        newname='';
-        for i=1:length(pos)
-            newname=[newname str0(pos0(i)+1:pos(i)-1) char(hex2dec(str0(pos(i)+3:pend(i)-1)))];
-        end
-        if(pos(end)~=length(name))
-            newname=[newname str0(pos0(end-1)+1:pos0(end))];
-        end
-    end
-end
 %%-------------------------------------------------------------------------
 function val=N_(str)
 global ismsgpack