Release Notes: With the 1.14 series, in an attempt to better handle out-of-disk situations, the WAV writing code was changed, and that broke writing of WAVs to standard output (raw data with a single WAV header in front). This should work again with the current release.
Release Notes: This version features a performance boost all across the board, with improved Huffman decoding and additional optimization for SSE and AVX. Several outputs have fixes regarding playback of the last bits in buffer. There are LFS fixes, this time for BSD systems with proper 64-bit off_t in 32 bit, providing aliases for builds that insist on having in complicated. Helper programs for dumping ID3 data and stripping anything non-MPEG were added officially.
Release Notes: Writing CDDA data via --cdr works again; it had been broken by attempts to catch failing writes because of a full disk, and the actual failure will properly trigger an early exit now. A bug in mpg123_chomp_string() has been fixed which made empty lines be badly treated (relevant for HTTP). There is now more resync strictness to really catch (all) cases with a changing decoder structure. This fixes possibly fatal behavior in the parser when readahead checking is explicitly disabled (e.g., via mpg123 --no-seekbuffer on http streams). By default, readahead checking is always active.
Release Notes: Writing of audio data with RIFF/WAVE headers to standard output in various configurations (redirect to file, pipe, "-", or /dev/stdout) has been fixed, hopefully for the last time. Also, this version improves portability of the build script.
Release Notes: This release fixes some bugs and build issues (mpg123_scan() and proper seek-back, better Frankenstein stream handling, HTTP header parsing, OS/2, BSD and Windows compatibility, and WAV-to-stdout, among others). It brings mainly enhancements in the form of more convenience with long tracks (audio books). It also adds application/x-scpls as another accepted playlist MIME type.
Release Notes: This is another regression-fixing release, which happens when skimping on verifying ground-breaking reworking of existing code. The resync logic no longer signals premature end of stream on bad headers, short seeks don't get bit reservoir unnecessarily cleared, and sample-accurate seek for files without gapless metadata is accurate again.