Commit Graph

76 Commits (7cea254f4ce3298ff0ef63c63853acde4dfad414)

Author SHA1 Message Date
Christopher Snowhill 728c44242c Do not reset output sample rate automatically
This was buggy as hell, and resulted in errors. Now the user should
restart playback if they change output device formats.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-07 22:02:17 -08:00
Christopher Snowhill 477feaab1d Now properly supports sample format changing
Sample format can now change dynamically at play time, and the player
will resample it as necessary, extrapolating edges between changes to
reduce the potential for gaps.

Currently supported formats for this:

- FLAC
- Ogg Vorbis
- Any format supported by FFmpeg, such as MP3 or AAC

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-07 19:18:45 -08:00
Christopher Snowhill acb1dd75d3 Cog Audio: Fix memory leaks with new buffering
By applying copious amounts of autorelease pools, memory is freed in a
timely manner. Prior to this, buffer objects were freed, but not being
released, and thus accumulating in memory indefinitely, as the original
threads and functions had autorelease pools that scoped the entire
thread, rather than individual function blocks that utilized the new
buffering system. This fixes memory growth caused by playback.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-07 04:06:36 -08:00
Christopher Snowhill 1ef8df675f Cog Audio: Implement support for channel config
This implements the basic output and mixing support for channel config
bits, optionally set by the input plugin.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-07 01:10:05 -08:00
Christopher Snowhill 85c7073649 Reformat my own source code with clang-format
Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-06 21:49:27 -08:00
Christopher Snowhill 62edb39761 Cog Audio: Major rewrite of audio buffering
Rewrite attempt number two. Now using array lists of audio chunks, with
each chunk having its format and optionally losslessness stashed along
with it. This replaces the old virtual ring buffer method. As a result
of this, the HRIR toggle now works instantaneously.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-06 03:08:34 -08:00
Christopher Snowhill 0131f7c925 Revert "Core Audio output: Rewrote major portions"
This reverts commit 637ea4efe1.
2022-02-05 04:14:03 -08:00
Christopher Snowhill 4cdca2f5f8 Converter: Fix DSD gaplessness
Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-05 04:03:40 -08:00
Christopher Snowhill 637ea4efe1 Core Audio output: Rewrote major portions
After all this rewriting, down or upmixing the audio is now handled with
the lowest latency possible, meaning that toggling the HRIR option now
takes effect immediately.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-05 03:45:02 -08:00
Christopher Snowhill 39cc33cac4 Converter: Remove no longer necessary includes
These were only needed when I wasn't using Accelerate framework.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-04 19:53:17 -08:00
Christopher Snowhill 074e4115dd sox resampler: Perform post file flush
Flush the resampler when the source file terminates, so that it outputs
delayed samples properly. This fixes gapless decoding of resampled
files.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-01 22:47:11 -08:00
Christopher Snowhill d4990de7f3 Adopt the sox resampler instead of RetroArch
Removing RetroArch code from my project.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-01 18:55:39 -08:00
Christopher Snowhill 61a30c959c Bundled resources: Use NSBundle interface
These methods should use NSBundle, rather than CF* C functions

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-02-01 14:40:02 -08:00
Christopher Snowhill 708c7dc721 Headphone Virtualization: Implement customization
Implement the ability to configure and select an HRIR preset to use with
the HRIR filter, or remove the preset. It will validate the file's
usefulness before setting it for the player to use.

Also, fixed back center channel filtering for 7.0 format audio.

Signed-off-by: Christopher Snowhill <kode54@gmail.com>
2022-01-25 21:30:33 -08:00
Christopher Snowhill e7b78085ca New feature: Implemented headphone virtualization
This new virtualizer uses the Accelerate framework to process samples.
I've bundled a HeSuVi impulse for now, and will add an option to select
an impulse in the future. It will validate the selection before sending
it to the actual filter, which outright fails if it receives invalid
input. Impulses will be supported in any arbitrary format that Cog
supports, but let's not go too hog wild, it requires HeSuVi 14 channel
presets.
2022-01-25 16:50:42 -08:00
Christopher Snowhill 9d1fd08574 HDCD Decoder: Only process lossless tracks 2022-01-21 22:47:11 -08:00
Christopher Snowhill 6f0a737123 Cog Audio: Implement HDCD decoding 2022-01-19 02:08:57 -08:00
Christopher Snowhill de193b70e2 Converter: Improve extrapolation for resampler, and also pad decimated DSD, even if not resampling it 2022-01-19 00:40:40 -08:00
Christopher Snowhill c4c29be35a Output converter: Limit extrapolation to only be trained on twice as many samples as the extrapolation order 2022-01-18 16:43:10 -08:00
Christopher Snowhill 71b2f7a4f2 Implement graphic equalizer 2022-01-16 07:32:47 -08:00
Christopher Snowhill c640481c25 Add SIMD optimized volume scaling 2022-01-16 00:46:21 -08:00
Christopher Snowhill 72453ba05a Cog Audio: Converter now rounds up sample counts affected by resamlping ratio 2022-01-15 23:41:31 -08:00
Christopher Snowhill fbef034903 Cog Audio: Further overhaul output buffering and track queue code 2022-01-15 02:09:26 -08:00
Christopher Snowhill 6315377eaf Cog Audio: Fix extrapolator to use signed type for sample delta 2022-01-14 11:37:52 -08:00
Christopher Snowhill af6677cf34 Cog Audio: Keep track of last resampler quality used so we don't reinitialize when not necessary, and also keep track of DSD decimated sample rate separately, so the input format remains unmodified 2022-01-14 11:18:50 -08:00
Christopher Snowhill 3ce3edde88 Cog Audio: Fix track end extrapolator used by resampler 2022-01-14 07:39:55 -08:00
Christopher Snowhill 748891f285 Cog Audio / WavPack input: Add DSD decimation to converter, and change WavPack input to emit only raw DSD 2022-01-14 06:26:09 -08:00
Christopher Snowhill 903b457a28 Cog Audio: Make it possible to play obscene sample rate files without major failures 2022-01-14 02:00:32 -08:00
Christopher Snowhill c8d2864862 Cog Audio: Enhance playback queue handler, so it always halts buffering when there are at least 30 seconds worth of buffers filled, possibly spanning multiple files. Also improve the chain reset function so that playlist changes and playback order control reset the queue properly when the queue refill function is currently entered in another thread. 2022-01-13 23:05:32 -08:00
Christopher Snowhill 7cc89c9f92 Cog Audio: Enhance track end extrapolator so it will always have a reasonable amount of buffered data for extrapolation 2022-01-13 23:03:53 -08:00
Christopher Snowhill 92d29e7acf Cog Audio: Now preserves already resampled output when switching output formats 2022-01-13 19:43:18 -08:00
Christopher Snowhill cc134ce293 Cog Audio: Converter node actually pauses when it is being reconfigured, and resets its buffer when the output format is changed 2022-01-12 23:13:00 -08:00
Christopher Snowhill 9feaffc92d Cog Audio: Made RetroArch resampler safer 2022-01-12 18:45:43 -08:00
Christopher Snowhill f1d70aaa54 Cog Audio: Add explanatory comment to source code 2022-01-12 03:36:27 -08:00
Christopher Snowhill 84d445cffe Cog Audio: Fix converter so that seeking on startup playback position resume works properly 2022-01-12 03:01:55 -08:00
Christopher Snowhill 91c3feac2e Cog Audio converter: Fix handling signed versus unsigned for 8 bit samples 2022-01-11 19:37:47 -08:00
Christopher Snowhill f44e4e793b Cog Audio converter: Fix end of track flush and extrapolation 2022-01-11 19:37:17 -08:00
Christopher Snowhill ed882e25cb Cog Audio: Resampler now extrapolates into latency padding for better gapless playback 2022-01-11 18:19:30 -08:00
Christopher Snowhill ee05fe9e44 Core Audio output: Add sanity checking for supported formats, and support 64 bit float input data, in case anything actually uses that 2022-01-11 17:09:06 -08:00
Christopher Snowhill 72210c67e4 Replaced AudioQueue output with AUAudioUnit 2022-01-11 07:06:40 -08:00
Christopher Snowhill c4c9a741ef Replaced AudioToolbox converter process with homebrew solution, using the RetroArch sinc resampler 2022-01-11 04:09:19 -08:00
Christopher Snowhill dfe92ffedc Audio Player: Fix so ReplayGain isn't reset on seek 2021-12-28 21:05:25 -08:00
Christopher Snowhill 38c530043d Output converter: Add more upmixing modes 2021-12-28 15:39:23 -08:00
Christopher Snowhill b13235ec5f Fix 6.1 stereo downmix mapping 2021-12-28 00:10:44 -08:00
Christopher Snowhill 258a56a276 Audio Output: Implement more upmixing modes, and correct some upmixing behavior 2021-12-28 00:09:57 -08:00
Christopher Snowhill 25a43a9083 Audio Output: Fix converter to always emit trailing samples on end of track 2021-12-27 17:04:16 -08:00
Christopher Snowhill 1cc34ea7d4 Audio Output: Made converter cleanup atomic, to wait on any threads that are still running 2021-12-26 04:34:33 -08:00
Christopher Snowhill d22ee14a36 Audio Output: Take a more nuclear approach to output resetting when seeking 2021-12-25 23:41:45 -08:00
Christopher Snowhill 1278b64afd Audio Output: Fixed converter so it won't loop endlessly when it either reaches the end of a stream or is otherwise supposed to terminate 2021-12-25 22:47:17 -08:00
Christopher Snowhill 5246731189 Core Audio: Overhauled audio output to use AudioQueue instead of AudioUnit, fixed downmixing to mono, and implemented upmixing mono or other formats to more channels if the device requests them 2021-12-25 22:32:43 -08:00