diff --git a/crates/audio/README.md b/crates/audio/README.md index 7853f11b1..c1f7dcef8 100644 --- a/crates/audio/README.md +++ b/crates/audio/README.md @@ -264,7 +264,7 @@ match craft_audio::decode(&bytes, None) { `WAVEFORMATEXTENSIBLE_CHANNEL_MASK` comment. This includes the usual layouts for 5 and 6 channels: without the comment, back surrounds are read as side surrounds. It includes the usual 7-channel layout too: without the comment, its back left, back right and back centre are read as back - centre, side left and side right. craft-audio reads the comment in FLAC files, not in FLAC in Ogg. + centre, side left and side right. - FLAC at 32 bits codes stereo channels independently, because a side channel would need 33 bits. - Outside RFC 9639's streamable subset (§7), which some decoders require: FLAC at 32 bits, at a sample rate the frame header cannot name, or with a layout that needs the mask comment. diff --git a/crates/audio/src/read/flac.rs b/crates/audio/src/read/flac.rs index ad4f4d88b..86090e457 100644 --- a/crates/audio/src/read/flac.rs +++ b/crates/audio/src/read/flac.rs @@ -85,10 +85,25 @@ pub(crate) fn layout(b: &[u8], at: usize) -> Option { /// `channels` channels described by the channel mask comment of the stream at `at`, or else in /// the layout decoders assume for the count (RFC 9639 §8.6.2, §9.1.3). fn layout_of(b: &[u8], at: usize, channels: u16) -> ChannelLayout { - let mask = channel_mask(b, at).unwrap_or_else(|| crate::layout::flac_assumed_mask(usize::from(channels))); + layout_from(channel_mask(b, at), channels) +} + +/// `channels` channels described by a channel mask comment's value, or else in the layout decoders +/// assume for the count. +pub(crate) fn layout_from(mask: Option, channels: u16) -> ChannelLayout { + let mask = mask.unwrap_or_else(|| crate::layout::flac_assumed_mask(usize::from(channels))); ChannelLayout::from_mask_and_channels(mask, channels) } +/// The mask in a comment, when its key is the channel mask's and its value a `0x` hexadecimal number. +pub(crate) fn comment_mask(key: &[u8], value: &[u8]) -> Option { + if !key.eq_ignore_ascii_case(b"WAVEFORMATEXTENSIBLE_CHANNEL_MASK") { + return None; + } + let hex = value.strip_prefix(b"0x").or_else(|| value.strip_prefix(b"0X"))?; + u32::from_str_radix(std::str::from_utf8(hex).ok()?, 16).ok() +} + /// The value of the first `WAVEFORMATEXTENSIBLE_CHANNEL_MASK` comment that holds a `0x` hexadecimal /// number, in the stream at `at`. fn channel_mask(b: &[u8], at: usize) -> Option { @@ -118,11 +133,7 @@ fn comments_mask(block: &[u8]) -> Option { let comment = block.get(at + 4..(at + 4).checked_add(len)?)?; at += 4 + len; let Some(eq) = comment.iter().position(|&c| c == b'=') else { continue }; - if !comment[..eq].eq_ignore_ascii_case(b"WAVEFORMATEXTENSIBLE_CHANNEL_MASK") { - continue; - } - let hex = comment[eq + 1..].strip_prefix(b"0x").or_else(|| comment[eq + 1..].strip_prefix(b"0X")); - if let Some(mask) = hex.and_then(|h| std::str::from_utf8(h).ok()).and_then(|h| u32::from_str_radix(h, 16).ok()) { + if let Some(mask) = comment_mask(&comment[..eq], &comment[eq + 1..]) { return Some(mask); } } diff --git a/crates/audio/src/read/symph.rs b/crates/audio/src/read/symph.rs index a27d23081..8a6f35313 100644 --- a/crates/audio/src/read/symph.rs +++ b/crates/audio/src/read/symph.rs @@ -16,7 +16,7 @@ use symphonia::core::errors::Error as SymError; use symphonia::core::formats::probe::{Hint, ProbeOptions, ProbeableFormat, Score, Scoreable}; use symphonia::core::formats::{FormatOptions, FormatReader, SeekMode, SeekTo, Track}; use symphonia::core::io::{MediaSourceStream, MediaSourceStreamOptions, ScopedStream}; -use symphonia::core::meta::MetadataOptions; +use symphonia::core::meta::{MetadataOptions, RawValue}; use symphonia::core::units::{TimeBase, Timestamp}; use super::id3v2_len; @@ -446,6 +446,21 @@ fn layout_of(channels: Option<&Channels>) -> ChannelLayout { } } +/// The channel mask comment among the tags the reader has read. +fn metadata_channel_mask(reader: &mut dyn FormatReader) -> Option { + let mut metadata = reader.metadata(); + loop { + let mask = metadata.current()?.media.tags.iter().find_map(|t| match &t.raw.value { + RawValue::String(value) => super::flac::comment_mask(t.raw.key.as_bytes(), value.as_bytes()), + _ => None, + }); + if mask.is_some() { + return mask; + } + metadata.pop()?; + } +} + /// The stored sample format, using the codec and its declared bit depth. Lossy and unusual /// codecs decode to `Compressed`. fn sample_format_of(p: &AudioCodecParameters) -> SampleFormat { @@ -547,7 +562,7 @@ impl Symph { /// Open audio track `track` (the first when `None`) of the file in `bytes`. pub(crate) fn open(bytes: Cow<'_, [u8]>, detected: Option, track: Option) -> Result { guarded(move || { - let (reader, flac_layout) = match open(bytes, detected)? { + let (mut reader, flac_layout) = match open(bytes, detected)? { Opened::Reader(reader, layout) => (reader, layout), Opened::Empty(info) => { // a stream with no audio still has one track, as WAV and AIFF do @@ -573,6 +588,8 @@ impl Symph { let (track_id, mut info) = info_of(reader.as_ref(), detected, track)?; if let Some(layout) = flac_layout.filter(|l| l.channels() == info.channels) { info.layout = layout; + } else if info.codec == CodecId::Flac && info.container == Container::Ogg { + info.layout = super::flac::layout_from(metadata_channel_mask(reader.as_mut()), info.channels); } let t = reader.tracks().iter().find(|t| t.id == track_id).ok_or_else(|| AudioError::Malformed("track vanished".into()))?; let params = audio_params(t).cloned().ok_or_else(|| AudioError::Malformed("track vanished".into()))?; diff --git a/crates/audio/src/write/flac/oracle_tests.rs b/crates/audio/src/write/flac/oracle_tests.rs index 2ac0a853e..6dd19f177 100644 --- a/crates/audio/src/write/flac/oracle_tests.rs +++ b/crates/audio/src/write/flac/oracle_tests.rs @@ -236,6 +236,34 @@ fn layouts_decoders_would_not_assume_are_recorded() { } } +/// FLAC in Ogg reads its layout as a FLAC file does: from the mask comment, or else the one +/// decoders assume for the count. +#[cfg(feature = "symphonia")] +#[test] +fn ogg_flac_reads_the_mask_comment() { + use crate::ChannelLayout; + let Some(ff) = ffmpeg() else { return }; + let cases: [(&str, &str, &[&str], u16, u32); 3] = [ + ("ogg-mask", "aevalsrc=0|0|0|0:c=4c:d=0.1:s=48000", &["-metadata", "WAVEFORMATEXTENSIBLE_CHANNEL_MASK=0x0107"], 4, 0x107), + ("ogg-5.1", "aevalsrc=0|0|0|0|0|0:c=5.1:d=0.1:s=48000", &[], 6, 0x60F), + ("ogg-stereo", "aevalsrc=0|0:c=stereo:d=0.1:s=48000", &["-metadata", "TITLE=x"], 2, 0x3), + ]; + for (name, source, extra, channels, mask) in cases { + let path = common::fixture_subdir("flac").join(format!("{name}.oga")); + let status = Command::new(&ff) + .args(["-v", "error", "-y", "-f", "lavfi", "-i", source, "-c:a", "flac", "-f", "ogg"]) + .args(extra) + .arg(&path) + .status() + .unwrap(); + assert!(status.success(), "{name}"); + let bytes = std::fs::read(&path).unwrap(); + let expected = ChannelLayout::from_mask_and_channels(mask, channels); + assert_eq!(crate::probe(&bytes, None).unwrap().layout, expected, "probe {name}"); + assert_eq!(crate::decode(&bytes, None).unwrap().0.layout, expected, "decode {name}"); + } +} + #[test] fn the_mask_comment_is_written_only_when_decoders_would_assume_otherwise() { use super::mask_comment;