diff --git a/pypdf/_cmap.py b/pypdf/_cmap.py
index 54c54436e..e6c4bce88 100644
--- a/pypdf/_cmap.py
+++ b/pypdf/_cmap.py
@@ -396,7 +396,7 @@ def build_font_width_map(
     st: int = 0
     en: int = 0
     try:
-        default_font_width = _default_fonts_space_width[cast(str, ft["/BaseFont"].get_object)] * 2.0
+        default_font_width = _default_fonts_space_width[cast(str, ft["/BaseFont"].get_object())] * 2.0
     except KeyError:
         pass
     if "/DescendantFonts" in ft:  # ft["/Subtype"].startswith("/CIDFontType"):
@@ -404,9 +404,9 @@ def build_font_width_map(
         # Widths for a CIDFont are defined using the DW and W entries.
         # DW2 and W2 are for vertical use. Vertical type is not implemented.
         ft1 = ft["/DescendantFonts"][0].get_object()  # type: ignore
-        try:
-            font_width_map["default"] = cast(float, ft1["/DW"])
-        except Exception:
+        if "/DW" in ft1:
+            font_width_map["default"] = cast(float, ft1["/DW"].get_object())
+        else:
             font_width_map["default"] = default_font_width
         if "/W" in ft1:
             w = ft1["/W"].get_object()
@@ -418,13 +418,15 @@ def build_font_width_map(
             if isinstance(second, int):
                 # C_first C_last same_W
                 en = second
+                width = w[2].get_object()
                 for c_code in range(st, en + 1):
-                    font_width_map[chr(c_code)] = w[2]
+                    font_width_map[chr(c_code)] = width
                 w = w[3:]
             elif isinstance(second, list):
                 # Starting_C [W1 W2 ... Wn]
                 c_code = st
-                for width in second:
+                for ww in second:
+                    width = ww.get_object()
                     font_width_map[chr(c_code)] = width
                     c_code += 1
                 w = w[2:]
diff --git a/pypdf/_page.py b/pypdf/_page.py
index d5de76f77..cbae35817 100644
--- a/pypdf/_page.py
+++ b/pypdf/_page.py
@@ -1750,7 +1750,7 @@ def _debug_for_extract(self) -> str:  # pragma: no cover
             out += "No Font\n"
         return out
 
-    def _get_acutual_font_widths(
+    def _get_actual_font_widths(
         self,
         cmap: Tuple[
             Union[str, Dict[int, str]], Dict[str, str], str, Optional[DictionaryObject]
@@ -1817,7 +1817,7 @@ def _handle_tj(
                 rtl_dir,
                 visitor_text)
         font_widths, actual_str_size["space_width"], actual_str_size["str_height"] = (
-            self._get_acutual_font_widths(cmap, text_operands, font_size, space_width))
+            self._get_actual_font_widths(cmap, text_operands, font_size, space_width))
         actual_str_size["str_widths"] += font_widths
 
         return text, rtl_dir, actual_str_size
diff --git a/tests/test_cmap.py b/tests/test_cmap.py
index 07a778520..55df3f1f2 100644
--- a/tests/test_cmap.py
+++ b/tests/test_cmap.py
@@ -269,3 +269,15 @@ def test_iss2925():
     name = "iss2925.pdf"
     reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
     assert "slicing on the PDG to extract the relevant contextual" in reader.pages[3].extract_text()
+
+
+@pytest.mark.enable_socket
+def test_iss2966():
+    """Regression test for issue #2966: indirect objects in fonts"""
+    url = (
+        "https://github.com/user-attachments/files/17904233/repro_out.pdf"
+    )
+    name = "iss2966.pdf"
+    reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+    assert "Lorem ipsum dolor sit amet" in reader.pages[0].extract_text()
+