Extract esds box from wave box (#96)

* Extract esds from wave box

* Allow empty, multi-byte, and arbitrary NUL terminated strings

* Skip unsupported avc1 sub-boxes

* Fixed non-integer framerates

* Fixed bitrate calculation

* Fixed format issue

* Public read sample offset

* Fix lint warning.

---------

Co-authored-by: Alfred Gutierrez <alfg@users.noreply.github.com>

src/mp4box/avc1.rs

@@ -101,30 +101,37 @@ impl<R: Read + Seek> ReadBox<&mut R> for Avc1Box {
         let depth = reader.read_u16::<BigEndian>()?;
         reader.read_i16::<BigEndian>()?; // pre-defined
 
-        let header = BoxHeader::read(reader)?;
-        let BoxHeader { name, size: s } = header;
-        if s > size {
-            return Err(Error::InvalidData(
-                "avc1 box contains a box with a larger size than it",
-            ));
-        }
-        if name == BoxType::AvcCBox {
-            let avcc = AvcCBox::read_box(reader, s)?;
+        let end = start + size;
+        loop {
+            let current = reader.stream_position()?;
+            if current >= end {
+                return Err(Error::InvalidData("avcc not found"));
+            }
+            let header = BoxHeader::read(reader)?;
+            let BoxHeader { name, size: s } = header;
+            if s > size {
+                return Err(Error::InvalidData(
+                    "avc1 box contains a box with a larger size than it",
+                ));
+            }
+            if name == BoxType::AvcCBox {
+                let avcc = AvcCBox::read_box(reader, s)?;
 
-            skip_bytes_to(reader, start + size)?;
+                skip_bytes_to(reader, start + size)?;
 
-            Ok(Avc1Box {
-                data_reference_index,
-                width,
-                height,
-                horizresolution,
-                vertresolution,
-                frame_count,
-                depth,
-                avcc,
-            })
-        } else {
-            Err(Error::InvalidData("avcc not found"))
+                return Ok(Avc1Box {
+                    data_reference_index,
+                    width,
+                    height,
+                    horizresolution,
+                    vertresolution,
+                    frame_count,
+                    depth,
+                    avcc,
+                });
+            } else {
+                skip_bytes_to(reader, current + s)?;
+            }
         }
     }
 }

src/mp4box/dinf.rs

@@ -264,22 +264,16 @@ impl<R: Read + Seek> ReadBox<&mut R> for UrlBox {
 
         let (version, flags) = read_box_header_ext(reader)?;
 
-        let location = if size.saturating_sub(HEADER_SIZE + HEADER_EXT_SIZE) > 0 {
-            let buf_size = size - HEADER_SIZE - HEADER_EXT_SIZE - 1;
-            let mut buf = vec![0u8; buf_size as usize];
-            reader.read_exact(&mut buf)?;
-            match String::from_utf8(buf) {
-                Ok(t) => {
-                    if t.len() != buf_size as usize {
-                        return Err(Error::InvalidData("string too small"));
-                    }
-                    t
-                }
-                _ => String::default(),
-            }
-        } else {
-            String::default()
-        };
+        let buf_size = size
+            .checked_sub(HEADER_SIZE + HEADER_EXT_SIZE)
+            .ok_or(Error::InvalidData("url size too small"))?;
+
+        let mut buf = vec![0u8; buf_size as usize];
+        reader.read_exact(&mut buf)?;
+        if let Some(end) = buf.iter().position(|&b| b == b'\0') {
+            buf.truncate(end);
+        }
+        let location = String::from_utf8(buf).unwrap_or_default();
 
         skip_bytes_to(reader, start + size)?;
 

src/mp4box/hdlr.rs

@@ -53,20 +53,15 @@ impl<R: Read + Seek> ReadBox<&mut R> for HdlrBox {
         skip_bytes(reader, 12)?; // reserved
 
         let buf_size = size
-            .checked_sub(HEADER_SIZE + HEADER_EXT_SIZE + 20 + 1)
+            .checked_sub(HEADER_SIZE + HEADER_EXT_SIZE + 20)
             .ok_or(Error::InvalidData("hdlr size too small"))?;
+
         let mut buf = vec![0u8; buf_size as usize];
         reader.read_exact(&mut buf)?;
-
-        let handler_string = match String::from_utf8(buf) {
-            Ok(t) => {
-                if t.len() != buf_size as usize {
-                    return Err(Error::InvalidData("string too small"));
-                }
-                t
-            }
-            _ => String::from("null"),
-        };
+        if let Some(end) = buf.iter().position(|&b| b == b'\0') {
+            buf.truncate(end);
+        }
+        let handler_string = String::from_utf8(buf).unwrap_or_default();
 
         skip_bytes_to(reader, start + size)?;
 
@@ -127,4 +122,52 @@ mod tests {
         let dst_box = HdlrBox::read_box(&mut reader, header.size).unwrap();
         assert_eq!(src_box, dst_box);
     }
+
+    #[test]
+    fn test_hdlr_empty() {
+        let src_box = HdlrBox {
+            version: 0,
+            flags: 0,
+            handler_type: str::parse::<FourCC>("vide").unwrap(),
+            name: String::new(),
+        };
+        let mut buf = Vec::new();
+        src_box.write_box(&mut buf).unwrap();
+        assert_eq!(buf.len(), src_box.box_size() as usize);
+
+        let mut reader = Cursor::new(&buf);
+        let header = BoxHeader::read(&mut reader).unwrap();
+        assert_eq!(header.name, BoxType::HdlrBox);
+        assert_eq!(src_box.box_size(), header.size);
+
+        let dst_box = HdlrBox::read_box(&mut reader, header.size).unwrap();
+        assert_eq!(src_box, dst_box);
+    }
+
+    #[test]
+    fn test_hdlr_extra() {
+        let real_src_box = HdlrBox {
+            version: 0,
+            flags: 0,
+            handler_type: str::parse::<FourCC>("vide").unwrap(),
+            name: String::from("Good"),
+        };
+        let src_box = HdlrBox {
+            version: 0,
+            flags: 0,
+            handler_type: str::parse::<FourCC>("vide").unwrap(),
+            name: String::from_utf8(b"Good\0Bad".to_vec()).unwrap(),
+        };
+        let mut buf = Vec::new();
+        src_box.write_box(&mut buf).unwrap();
+        assert_eq!(buf.len(), src_box.box_size() as usize);
+
+        let mut reader = Cursor::new(&buf);
+        let header = BoxHeader::read(&mut reader).unwrap();
+        assert_eq!(header.name, BoxType::HdlrBox);
+        assert_eq!(src_box.box_size(), header.size);
+
+        let dst_box = HdlrBox::read_box(&mut reader, header.size).unwrap();
+        assert_eq!(real_src_box, dst_box);
+    }
 }

src/mp4box/mod.rs

@@ -237,7 +237,8 @@ boxtype! {
     DayBox => 0xa9646179,
     CovrBox => 0x636f7672,
     DescBox => 0x64657363,
-    WideBox => 0x77696465
+    WideBox => 0x77696465,
+    WaveBox => 0x77617665
 }
 
 pub trait Mp4Box: Sized {

src/mp4box/mp4a.rs

@@ -82,16 +82,28 @@ impl<R: Read + Seek> ReadBox<&mut R> for Mp4aBox {
         reader.read_u32::<BigEndian>()?; // reserved
         reader.read_u16::<BigEndian>()?; // reserved
         let data_reference_index = reader.read_u16::<BigEndian>()?;
-
-        reader.read_u64::<BigEndian>()?; // reserved
+        let version = reader.read_u16::<BigEndian>()?;
+        reader.read_u16::<BigEndian>()?; // reserved
+        reader.read_u32::<BigEndian>()?; // reserved
         let channelcount = reader.read_u16::<BigEndian>()?;
         let samplesize = reader.read_u16::<BigEndian>()?;
         reader.read_u32::<BigEndian>()?; // pre-defined, reserved
         let samplerate = FixedPointU16::new_raw(reader.read_u32::<BigEndian>()?);
 
+        if version == 1 {
+            // Skip QTFF
+            reader.read_u64::<BigEndian>()?;
+            reader.read_u64::<BigEndian>()?;
+        }
+
+        // Find esds in mp4a or wave
         let mut esds = None;
-        let current = reader.stream_position()?;
-        if current < start + size {
+        let end = start + size;
+        loop {
+            let current = reader.stream_position()?;
+            if current >= end {
+                break;
+            }
             let header = BoxHeader::read(reader)?;
             let BoxHeader { name, size: s } = header;
             if s > size {
@@ -99,13 +111,20 @@ impl<R: Read + Seek> ReadBox<&mut R> for Mp4aBox {
                     "mp4a box contains a box with a larger size than it",
                 ));
             }
-
             if name == BoxType::EsdsBox {
                 esds = Some(EsdsBox::read_box(reader, s)?);
+                break;
+            } else if name == BoxType::WaveBox {
+                // Typically contains frma, mp4a, esds, and a terminator atom
+            } else {
+                // Skip boxes
+                let skip_to = current + s;
+                skip_bytes_to(reader, skip_to)?;
             }
-            skip_bytes_to(reader, start + size)?;
         }
 
+        skip_bytes_to(reader, end)?;
+
         Ok(Mp4aBox {
             data_reference_index,
             channelcount,

src/reader.rs

@@ -262,6 +262,14 @@ impl<R: Read + Seek> Mp4Reader<R> {
             Err(Error::TrakNotFound(track_id))
         }
     }
+
+    pub fn sample_offset(&mut self, track_id: u32, sample_id: u32) -> Result<u64> {
+        if let Some(track) = self.tracks.get(&track_id) {
+            track.sample_offset(sample_id)
+        } else {
+            Err(Error::TrakNotFound(track_id))
+        }
+    }
 }
 
 impl<R> Mp4Reader<R> {

src/track.rs

@@ -167,11 +167,11 @@ impl Mp4Track {
     }
 
     pub fn frame_rate(&self) -> f64 {
-        let dur_msec = self.duration().as_millis() as u64;
-        if dur_msec > 0 {
-            ((self.sample_count() as u64 * 1000) / dur_msec) as f64
-        } else {
+        let dur = self.duration();
+        if dur.is_zero() {
             0.0
+        } else {
+            self.sample_count() as f64 / dur.as_secs_f64()
         }
     }
 
@@ -222,12 +222,12 @@ impl Mp4Track {
             }
             // mp4a.esds.es_desc.dec_config.avg_bitrate
         } else {
-            let dur_sec = self.duration().as_secs();
-            if dur_sec > 0 {
-                let bitrate = self.total_sample_size() * 8 / dur_sec;
-                bitrate as u32
-            } else {
+            let dur = self.duration();
+            if dur.is_zero() {
                 0
+            } else {
+                let bitrate = self.total_sample_size() as f64 * 8.0 / dur.as_secs_f64();
+                bitrate as u32
             }
         }
     }
@@ -437,7 +437,7 @@ impl Mp4Track {
         }
     }
 
-    fn sample_offset(&self, sample_id: u32) -> Result<u64> {
+    pub fn sample_offset(&self, sample_id: u32) -> Result<u64> {
         if !self.trafs.is_empty() {
             if let Some((traf_idx, sample_idx)) = self.find_traf_idx_and_sample_idx(sample_id) {
                 let mut sample_offset = self.trafs[traf_idx]

tests/lib.rs

@@ -99,8 +99,8 @@ fn test_read_mp4() {
     assert_eq!(track1.video_profile().unwrap(), AvcProfile::AvcHigh);
     assert_eq!(track1.width(), 320);
     assert_eq!(track1.height(), 240);
-    assert_eq!(track1.bitrate(), 0); // XXX
-    assert_eq!(track1.frame_rate(), 25.00); // XXX
+    assert_eq!(track1.bitrate(), 150200);
+    assert_eq!(track1.frame_rate(), 25.00);
 
     // track #2
     let track2 = mp4.tracks().get(&2).unwrap();