View source
compressed_raw
  • Contributors
    1. Loading...
Author Commit Message Date Builds
167 commits behind unified.
Alex
mlv_lite: better fix for invalid block size metadata (see commit 9d5a239) (introduced a new state for frame slots - capturing in progress)
Alex
Dummy merge of raw_video_10bit_12bit into compressed_raw
Alex
Merged unified into raw_video_10bit_12bit
Alex
Merged rec-trigger into raw_video_10bit_12bit
Alex
lossless.c: disable debug messages
Alex
mlv_lite: fix false assertion when resizing frames
Alex
mlv_lite: fix overflow at low resolutions (too many frame slots)
Alex
Merged unified into compressed_raw
Alex
mlv_lite: fix invalid block size metadata (some blocks incorrectly had the uncompressed block size in metadata) (writing to card started too early, before knowing the compressed size)
Alex
mlv_lite: workaround for a very mysterious lock-up
Alex
mlv_lite: count the number of reserved and regular (valid) slots (fixes pre-buffering allocation, among others)
Alex
mlv_lite: compute average compression rate (used in menu and for recording time estimations)
Alex
Merged rec-trigger into compressed_raw
Alex
mlv_dump: workaround to decode compressed MLV files (processing unsupported; todo - fix and port mlv_rec_lj92 branch)
Alex
chdk-dng: fixes for mlv_dump
Alex
mlv_lite: properly tag MLV files as lossless JPEG
Alex
mlv_lite: more detailed buffer status display (including buffer sizes)
Alex
mlv_lite: resize frame slots on the fly, after compression (todo: fix indicators, stats etc)
Alex
mlv_lite: moved VIDF initialization to init_mlv_chunk_headers
Alex
mlv_lite: prepare to support variable frame size (refactored to avoid hardcoding frame_size everywhere)
Alex
mlv_lite: move all EDMAC activity to compress_task (compression requires separate task, so use it for uncompressed cropping as well, to simplify code) (to be tested for corrupted frames)
Alex
Merged raw_video_10bit_12bit into compressed_raw
Alex
mlv_lite: lossless compression works! (proof of concept) To decode: mlv_dump --no-stripes --no-fixcp TODO: fix MLV metadata, reduce frame sizes, integrate with 10/12-bit recording
Alex
lossless.c: allow compressing a cropped area of the source image (similar to edmac_copy_rectangle)
Alex
lossless.c: no need for vsync-friendly API (it doesn't even work from vsync)
Alex
Lossless compression: support for 5D3 1.2.3
Alex
Merged unified into compressed_raw
Alex
lossless.c: nicer handling of debug messages
Alex
mlv_lite: in "pre only" trigger mode, use all available RAM for pre-recording
Alex
mlv_lite: fix max pre-recording frame count in low memory conditions
Alex
mlv_lite: fix roundoff error in pre-recorded buffer size and indicators
Alex
mlv_lite: display fractions of seconds when the pre-recording buffer becomes full; change rounding seconds to show fully completed seconds
Alex
mlv_lite: fix 1-frame trigger (was actually recording 2 frames); Renamed this option to "pre only" (as the saved frame(s) are always pre-recorded) and updated help
Alex
mlv_lite: fix some issues related to changing resolution (crash on submenu open while recording; message about max resolution not displayed)
Alex
raw.c: fix bad merge (duplicate DEFAULT_RAW_BUFFER entries)
Alex
mlv_lite: show pre-recording buffer size in menu (requires recording a test clip first)
Alex
mlv_lite: allow pre-record duration in 1-second increments
Alex
mlv_lite: allow frame skipping during pre-recording, between half-shutter triggers This will avoid recording stopping between (half-shutter) triggers, as these skipped frames would probably be dropped anyway. The pre-recording buffer is now maxed out in the 1-frame trigger mode.
Alex
mlv_lite: dynamic buffer allocation for pre-recording (reserve at least 50% of RAM for pre-recording, or more -- as long as it can record 500+ frames with 90% of measured write speed)
Alex
mlv_lite: fix pre-recording buffer size rounding (minor)
Alex
mlv-lite: experimental half-shutter trigger (compatible with pre-recording)
Alex
mlv_lite: prepare to allow multiple triggers with pre-recording
Alex
mlv_lite: refactor with INC_MOD (minor optimization in vsync hook)
Alex
imath.h: macro for incrementing a variable modulo N
Alex
Merged raw-h264-proxy into raw_video_10bit_12bit
Alex
Merged unified into raw_video_10bit_12bit
Alex
Merged raw_fixes into raw_video_10bit_12bit
Alex
mlv_lite: getting DCIM dir suffix from property doesn't seem to work (limitation of module properties); get it from ML core instead
Alex
fio-ml: added get_dcim_dir_suffix(); moved get_dcim_dir() from dryos.h
Alex
Merged unified (mlv-lite) into raw-h264-proxy
Alex
raw_rec: allow audio recording via H.264 proxy
Alex
raw_rec: prefer recording to CF whenever possible (when using H.264 proxy, select SD in Canon menu for best performance)
Alex
raw_rec: experimental H.264 proxy recording earlier attempts: http://www.magiclantern.fm/forum/index.php?topic=16608.msg177236#msg177236
Daniel Fort
conform raw_rec Bit depth menu to match mlv_rec
Alex
mlv_dump: fix dark/flat frame operation when the two clips have different spacing/padding offsets Test files: http://www.magiclantern.fm/forum/index.php?topic=16650.msg178809#msg178809
Alex
Moved bit depth setup from raw_rec/mlv_rec to raw backend; raw overlays should now refuse to run at bit depths other than 14
Alex
raw_rec, mlv_rec: minimize roundoff error when scaling black and white levels (using --black-fix in mlv_dump should be no longer needed)
Alex
mlv_dump: fix long exposure metadata in DNG
Alex
chdk-dng: drop APEX exposure metadata (redundant, conflicts with regular exposure tags)
Alex
mlv_dump: fix black level bias when converting from lower bit depths (on 5D3, please use --black-fix=2048 at all bit depths from now on) http://www.magiclantern.fm/forum/index.php?topic=5601.msg177802#msg177802
g3gg0
mlv_dump: changed default white level (if no parameter given) for --white-fix to 15k as recommended by a1ex
g3gg0
mlv_dump: added option to set white level manually, correct white level setting when black level was set
g3gg0
mlv_dump: modify black level code to properly scale or set black level as expected
Alex
Merged raw_fixes into raw_video_10bit_12bit
Alex
Merged raw_fixes into raw_video_10bit_12bit
Alex
Merged raw_fixes into raw_video_10bit_12bit
g3gg0
mlv_dump: put DNG 14bpp enforcement at the right position
g3gg0
mlv_dump: enforce 14bpp correction for DNG output mode. code seems still buggy for 12bpp.
g3gg0
mlv_dump: preallocate buffers on RAWI
g3gg0
mlv_dump: remove pointer arithmetics for xref table reading
g3gg0
mlv_dump: added patches from 5f7a9d850d71 and a22302855d05 regarding black level correction for other bit depths
Daniel Fort
mlv_dump: fix to extract only the first frame (numbered 0)
g3gg0
mlv_rec: fixed chroma smooth regression in mlv_dump caused by API change in 5167619. thanks dfort/danne/bouncyball.
g3gg0
mlv_dump: removed ancient bugfix options. if there are buggy files, they can get recovered using older versions.
eNnvi
Added 6D support and changed default 700D PREFERRED_RAW_TYPE
Alex
Silent pictures: in burst modes, reserve memory for lossless DNG compression
Alex
Silent pictures: experimental lossless DNG compression (5D3 1.1.3)
Alex
chdk-dng: allow DNG compression (Lossless JPEG)
Alex
Merged edmac into compressed_raw
Alex
Silent pics: only show the slit-scan mode option when you actually select slit-scan
g3gg0
mlv_rec: fix startup delay
g3gg0
mlv_rec: synchronize parts of code with raw_rec, more needed. use double buffering again. fix alternating frame errors.
Alex
Merged in daniel_fort/magic-lantern/raw_video_10bit_12bit_700D_slurp (pull request #781) 650D and 700D: enable CONFIG_EDMAC_RAW_SLURP
Daniel Fort
650D: enable CONFIG_EDMAC_RAW_SLURP
Daniel Fort
700D: enable CONFIG_EDMAC_RAW_SLURP
g3gg0
mlv_play: added direct raw_twk support along with zoom support (5D3 tested)
g3gg0
raw_twk: added experimental module that uses engines (PACK16, DEF, DSUNPACK) to accelerate raw->YUV conversion
Alex
Merged in daniel_fort/magic-lantern/raw_video_10bit_12bit_debug (pull request #780) Increase RAW_DEBUG_TYPE limit
Daniel Fort
use hex units for PREFERRED_RAW_TYPE debugging
Daniel Fort
finding usable RAW_DEBUG_TYPE above the 64 limit
Alex
Merge new_raw_format
Alex
raw_rec: fix some resolution inconsistencies
Alex
raw_rec: minor resolution fine-tuning fix
g3gg0
mlv_play: restore bpp and black level in raw_info on exit
Alex
raw_rec: remove unused odd bit depths from menu
Alex
engio_write wrapper
g3gg0
mlv_rec/mlv_play: add variable bit depth support. mlv_play requires experimental raw_twk if you want to playback other than 14bpp
Alex
raw_rec: allow max width to be multiple of 2 bytes and 8 pixels; all others are multiple of 16px
Alex
raw_rec: fix odd resolutions bug (make sure width is multiple of 8 pixels)
Alex
60D: use PREFERRED_RAW_TYPE (fixes digital ISO behavior and bad pixels in preview)
Alex
Raw backend: fix image freeze on 60D, maybe others (typo in 193dc59)
Alex
Raw backend: RAW_TYPE_ADDRESS no longer needed
Alex
raw_rec: allow 16-pixel increments for X resolution (experiment)
Alex
edmac_memcpy: looks like we can use widths that are not modulo 8 bytes
Alex
edmac-memcpy.c: refactored to avoid static struct edmac_info
Alex
raw_set_geometry: allow bit depths other than 14
Alex
Merge new_raw_format
David Milligan
EOSM: enable CONFIG_EDMAC_RAW_SLURP
David Milligan
raw bit depth: skip some frames after setting a new bit depth
David Milligan
raw bit depth: override bit depth related raw_info fields when changing bit depth
Alex
mlv_dump: fix DNG output for lower bit depths You need to disable all processing, e.g. mlv_dump M00-1234.MLV --dng --no-fixcp --no-stripes
Alex
Raw backend: fix image corruption when using other bit depths, on cameras using CONFIG_EDMAC_RAW_SLURP Other models will either have to enable this method, or patch Canon code to adjust image size.
David Milligan
10bit and 12bit raw video implementation for MLV Lite (confirmed working on 60D)