diff --git a/src/calibre/ebooks/oeb/polish/check/css.py b/src/calibre/ebooks/oeb/polish/check/css.py
index ceeac61324..df0d621479 100644
--- a/src/calibre/ebooks/oeb/polish/check/css.py
+++ b/src/calibre/ebooks/oeb/polish/check/css.py
@@ -48,6 +48,9 @@ def message_to_error(message, name, line_offset, rule_metadata):
         cls = CSSParseError
     else:
         cls = CSSError if message.get('severity') == 'error' else CSSWarning
+    if message.get('rule') == 'property-no-unknown' and 'panose-1' in message.get('text', ''):
+        # suppress panose-1 it is allowed in CSSS 2.1 and generated by calibre conversation
+        return
     title = message.get('text') or _('Unknown error')
     title = title.rpartition('(')[0].strip()
     line = as_int_or_none(message.get('line'))
diff --git a/src/calibre/ebooks/oeb/polish/check/parsing.py b/src/calibre/ebooks/oeb/polish/check/parsing.py
index b955f9c222..029ca25ff8 100644
--- a/src/calibre/ebooks/oeb/polish/check/parsing.py
+++ b/src/calibre/ebooks/oeb/polish/check/parsing.py
@@ -5,22 +5,22 @@ __license__ = 'GPL v3'
 __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
 
 import re
-
 from lxml.etree import XMLSyntaxError
 
-from calibre import force_unicode, human_readable, prepare_string_for_xml
-from calibre.ebooks.chardet import replace_encoding_declarations, find_declared_encoding
-from calibre.utils.xml_parse import safe_xml_fromstring
+from calibre import human_readable, prepare_string_for_xml
+from calibre.ebooks.chardet import find_declared_encoding, replace_encoding_declarations
 from calibre.ebooks.html_entities import html5_entities
+from calibre.ebooks.oeb.base import OEB_DOCS, URL_SAFE, XHTML, XHTML_NS, urlquote
+from calibre.ebooks.oeb.polish.check.base import INFO, WARN, BaseError
 from calibre.ebooks.oeb.polish.pretty import pretty_script_or_style as fix_style_tag
 from calibre.ebooks.oeb.polish.utils import PositionFinder, guess_type
-from calibre.ebooks.oeb.polish.check.base import BaseError, WARN, ERROR, INFO
-from calibre.ebooks.oeb.base import OEB_DOCS, XHTML_NS, urlquote, URL_SAFE, XHTML
-from polyglot.builtins import iteritems, error_message
+from calibre.utils.xml_parse import safe_xml_fromstring
+from polyglot.builtins import error_message, iteritems
 
 HTML_ENTITTIES = frozenset(html5_entities)
 XML_ENTITIES = {'lt', 'gt', 'amp', 'apos', 'quot'}
 ALL_ENTITIES = HTML_ENTITTIES | XML_ENTITIES
+fix_style_tag
 
 replace_pat = re.compile('&(%s);' % '|'.join(re.escape(x) for x in sorted(HTML_ENTITTIES - XML_ENTITIES)))
 mismatch_pat = re.compile(r'tag mismatch:.+?line (\d+).+?line \d+')
@@ -305,40 +305,6 @@ def check_xml_parsing(name, mt, raw):
     return errors
 
 
-class CSSError(BaseError):
-
-    is_parsing_error = True
-
-    def __init__(self, level, msg, name, line, col):
-        self.level = level
-        prefix = 'CSS: '
-        BaseError.__init__(self, prefix + msg, name, line, col)
-        if level == WARN:
-            self.HELP = _('This CSS construct is not recognized. That means that it'
-                          ' most likely will not work on reader devices. Consider'
-                          ' replacing it with something else.')
-        else:
-            self.HELP = _('Some reader programs are very'
-                          ' finicky about CSS stylesheets and will ignore the whole'
-                          ' sheet if there is an error. These errors can often'
-                          ' be fixed automatically, however, automatic fixing will'
-                          ' typically remove unrecognized items, instead of correcting them.')
-            self.INDIVIDUAL_FIX = _('Try to fix parsing errors in this stylesheet automatically')
-
-    def __call__(self, container):
-        root = container.parsed(self.name)
-        container.dirty(self.name)
-        if container.mime_map[self.name] in OEB_DOCS:
-            for style in root.xpath('//*[local-name()="style"]'):
-                if style.get('type', 'text/css') == 'text/css' and style.text and style.text.strip():
-                    fix_style_tag(container, style)
-            for elem in root.xpath('//*[@style]'):
-                raw = elem.get('style')
-                if raw:
-                    elem.set('style', force_unicode(container.parse_css(raw, is_declaration=True).cssText, 'utf-8').replace('\n', ' '))
-        return True
-
-
 pos_pats = (re.compile(r'\[(\d+):(\d+)'), re.compile(r'(\d+), (\d+)\)'))
 
 
@@ -428,41 +394,6 @@ class BareTextInBody(BaseError):
         return True
 
 
-class ErrorHandler:
-
-    ' Replacement logger to get useful error/warning info out of css_parser during parsing '
-
-    def __init__(self, name):
-        # may be disabled during setting of known valid items
-        self.name = name
-        self.errors = []
-
-    def __noop(self, *args, **kwargs):
-        pass
-    info = debug = setLevel = getEffectiveLevel = addHandler = removeHandler = __noop
-
-    def __handle(self, level, *args):
-        msg = ' '.join(map(str, args))
-        line = col = None
-        for pat in pos_pats:
-            m = pat.search(msg)
-            if m is not None:
-                line, col = int(m.group(1)), int(m.group(2))
-        if msg and line is not None:
-            # Ignore error messages with no line numbers as these are usually
-            # summary messages for an underlying error with a line number
-            if 'panose-1' in msg and 'unknown property name' in msg.lower():
-                return  # panose-1 is allowed in CSS 2.1 and is generated by calibre
-            self.errors.append(CSSError(level, msg, self.name, line, col))
-
-    def error(self, *args):
-        self.__handle(ERROR, *args)
-
-    def warn(self, *args):
-        self.__handle(WARN, *args)
-    warning = warn
-
-
 def check_filenames(container):
     errors = []
     all_names = set(container.name_path_map) - container.names_that_must_not_be_changed