| Age | Commit message (Collapse) | Author |
|---|
|
Due to the layout of the pitch_frequencies lookup table, note values
computed in extraction were reflected around middle C (midi note
number 60). This didn't matter for matching, values would successfully
roundtrip. However when using samples in external programs or
converting soundfonts to standard formats the note values would lead
to incorrect playback of sounds. This change corrects the note values
so that external programs correctly infer the pitch of the sound when
played at a particular MIDI key.
|
|
lines for building on Macos (#2718)
The present changes are a direct copy-paste from https://github.com/zeldaret/mm/pull/1850
This fixes building on macos due to a kinda specific issue with Apple clang.
When trying to build on Macos (specifically MacOS 12, Monteray with Apple clang 13.0.0 (clang-1300.0.29.30), idk if other versions have this issue too) `make` stops with the following error from `atblgen`:
```
Failed to match line 1: ""
regexec error: "regexec() failed to match"
Error: Malformed build/n64-us/assets/audio/sequence_order.in?
```
`atblgen` makes the assumption that the `sequence_order.in` file has no extra data, spaces, empty lines, etc. but the file somehow ends up having empty lines between each line on macos.
This file is created by using the C preprocessor to process `include/tables/sequence_table.h`.
In normal circumstances this file should look like this snip,
```
(Sequence_0,NA_BGM_GENERAL_SFX)
(Sequence_1,NA_BGM_AMBIENCE)
(Sequence_2,NA_BGM_TERMINA_FIELD)
(Sequence_3,NA_BGM_CHASE)
(Sequence_4,NA_BGM_MAJORAS_THEME)
```
but it ends up looking like this instead
```
(Sequence_0,NA_BGM_GENERAL_SFX)
(Sequence_1,NA_BGM_AMBIENCE)
(Sequence_2,NA_BGM_TERMINA_FIELD)
(Sequence_3,NA_BGM_CHASE)
(Sequence_4,NA_BGM_MAJORAS_THEME)
```
which `atblgen` doesn't like.
I believe this happens because there are lines with comments between each macro in [`sequence_table.h`](https://github.com/zeldaret/mm/blob/0877ce4adf28a8e73e05c3c58682273b8bf28749/include/tables/sequence_table.h) and for some reason this Apple clang version decided to preserve those empty lines
The fix just makes `atblgen` skip empty lines.
I threw `atblgen` to valgrind to check the fix was working as intended and noted a bunch of memory that was being free before exit, so I fixed them.
I also noted the tools/audio makefile was not using `OPTFLAGS` when building those tools, so I fixed that too.
|
|
(#2689)
* Read WAV SMPL chunks for note info when INST chunks are not available
* Populate the SMPL chunk on conversion to wav
* Properly handle unaligned (odd) WAV chunk sizes
|
|
(#2690)
|
|
|
|
|
|
* Introduce afile_sizes, generate headers of sizes for soundfonts and sequences
* Initial tools/audio README
* Versioning for samplebank extraction
* Clean up the disassemble_sequence.py runnable interface
* Add static assertions for maximum bank sizes
* Boost optimization for audio tools
* Samplebank XML doc
* Soundfont XML doc
* More docs in sampleconv for vadpcm
* Various tools fixes/cleanup
* VADPCM doc
* Try to fix md formatting
* VADPCM doc can come later
* Fix merge with PR 9
* Fix blobs from MM
* Try to fix bss
* Try fix bss round 2
* Fix sampleconv memset bug
* Suggested documentation tweaks
|
|
1.1 audio extraction (#2291)
* [Audio 9/?] Multiversion samplebank and soundfont extraction xmls, 1.0 and 1.1 audio extraction
* Rework multiversion samplebanks to reduce duplicates
|
|
|
|
|
|
various sfx arrays (#2137)
* [Audio 8/?] Check-in handwritten sequences, build sequences, automate various sfx arrays
* Fix whitespace in aseq.h
* Fix sequence 0 sfx id generator
* Suggested changes, adjust some MML syntax and add more instruction descriptions
* Correct some formatting in aseq.h
* Add the dir of the input .seq file to the list of includes to sequence assembling so that assembler-level includes like .include or .incbin work intuitively
* aseq.h tweaks
* MM review suggestions, aseq.h adjustments
|
|
* [Audio 7/?] Extract sequences to assembly
Co-authored-by: MNGoldenEagle <17274702+MNGoldenEagle@users.noreply.github.com>
Co-authored-by: zelda2774 <69368340+zelda2774@users.noreply.github.com>
* Suggested changes, some extra sequence disassembler cleanup
* Remove unused multiprocessing import and regen assets/xml/audio/sequences
---------
Co-authored-by: MNGoldenEagle <17274702+MNGoldenEagle@users.noreply.github.com>
Co-authored-by: zelda2774 <69368340+zelda2774@users.noreply.github.com>
|
|
* [Audio 6/?] Build Soundfonts and the Soundfont Table
* Improve lots of error messages
* First suggested changes
* Make audio build debugging more friendly
Co-authored-by: Dragorn421 <Dragorn421@users.noreply.github.com>
* Some fixes from MM review
* Make soundfont_table.h generation depend on the samplebank xmls since they are read, report from which soundfont the invalid pointer indirect warning originates from
---------
Co-authored-by: Dragorn421 <Dragorn421@users.noreply.github.com>
|
|
|
|
* [Audio 4/?] Build Samplebanks, match Audiotable
* Fix some makefile formatting
* Add missing scope in MARK chunk handling
* Add comment to generate asm file when buffer bug data is emitted, remove duplicate CC4 definition
* Adjust comment
* SBCFLAGS
* Remove unnecessary comments on notes_lut
* Split build directories creation command into several to avoid it becoming too long
* objcopy -j -> --only-section
* Fix mkdir warning when extracted/VERSION/assets doesn't exist
|
|
* [Audio 2/?] Extract audio samples to wav
Co-authored-by: zelda2774 <69368340+zelda2774@users.noreply.github.com>
* How
* Hopefully fix warning I don't get locally
* Pad default sample filenames, comment on the vadpcm frame encoder functions, other suggested changes
* Small tweaks to above
* Remove some obsolete code
---------
Co-authored-by: zelda2774 <69368340+zelda2774@users.noreply.github.com>
|
|
* [Audio 1/?] Extract Samplebanks and Soundfonts to XML
* Remove config.py and use the version yamls for addresses, other suggested changes
* Adjust setup-audio
* Remove some commented out dead code (MM review)
|
